The ability to distinguish whether an image is generated by artificial intelligence (AI) is a crucial ingredient in human intelligence, usually accompanied by a complex and dialectical forensic and reasoning process. However, current fake image detection models and databases focus on binary classification without understandable explanations for the general populace. This weakens the credibility of authenticity judgment and may conceal potential model biases. Meanwhile, large multimodal models (LMMs) have exhibited immense vision-language capabilities on various tasks, bringing the potential for explainable fake image detection. Therefore, we pioneer the probe of LMMs for explainable fake image detection by presenting a multimodal database encompassing descriptions of textual authenticity, the FakeBench. For construction, we first introduce a fine-grained taxonomy of generative visual forgery concerning human perception, based on which we collect forgery descriptions in human natural language with a human-in-the-loop strategy. FakeBench examines LMMs with four evaluation criteria: detection, reasoning, explanation and fine-grained forgery analysis, to obtain deeper insights into image authenticity-relevant capabilities. Experiments on various LMMs confirm their merits and demerits in different aspects of fake image detection tasks. This research presents a paradigm shift towards transparency for the fake image detection area and reveals the need for greater emphasis on forensic elements in visual-language research and AI risk control. FakeBench will be available athttps://github.com/Yixuan423/FakeBench

In the field of image quality assessment (IQA), researchers have been studying the mean opinion score (MOS) of image quality for decades. They focus on developing IQA methods with the help of MOS without using the potential of the distribution of opinion scores (DOS). We find that the Gaussian mixture distribution (GMD) can more accurately describe the DOS of image quality on SJTU IQSD and KonIQ-10K databases compared to some traditional distributions. Therefore, this paper proposes a blind IQA method that predicts the MOS of image quality by learning the GMD-based image quality. The proposed method consists of a visual feature learning module and a GMD learning module. The visual feature learning module uses a multi-stage Swin Transformer model and a CLIP feature extractor to extract visual features from an image. The GMD learning module then maps the extracted visual features to the GMD-based image quality using a mixture density network, where the mean of the GMD represents the MOS of image quality. We not only use the MOS of image quality to train the proposed method, but also employ the DOS of image quality for auxiliary training to improve the prediction performance of the proposed method. To address the lack of DOS in some existing IQA databases, we introduce a pseudo DOS generation strategy to generate the DOS of image quality for training, which significantly improves the applicability of the proposed method. Numerous analyses show that the proposed method is superior to most state-of-the-art IQA methods in predicting both the MOS and the DOS, thus facilitating a deeper investigation into the DOS of image quality in IQA.

In this paper, we study an online cost optimization problem for distributed data access. The goal of this problem is to dynamically create and delete data copies in a multi-server distributed system as time goes, in order to minimize the total storage and network cost of serving access requests. We propose an online algorithm with randomized storage periods of data copies in the servers, and derive an optimal probability density function of storage periods, which makes the algorithm achieve a competitive ratio of $1 + \frac{{\sqrt 2 }}{2}$. An example is presented to show that the competitive analysis of our algorithm is tight. Experimental evaluations using real data access traces demonstrate that our algorithm outperforms the best known deterministic algorithm.

This paper studies an online replication problem for distributed data access. The goal is to dynamically create and delete data copies in a multi-server system as time passes to minimize the total storage and network cost of serving access requests. We study the problem in the emergent learning-augmented setting, assuming simple binary predictions about inter-request times at individual servers. We develop an online algorithm and prove that it is (5+α/3)-consistent (competitiveness under perfect predictions) and (1+1/α)-robust (competitiveness under terrible predictions), where α◰(0, 1] is a hyper-parameter representing the level of distrust in the predictions. We also study the impact of mispredictions on the competitive ratio of the proposed algorithm and adapt it to achieve a bounded robustness while retaining its consistency. We further establish a lower bound of 3/2 on the consistency of any deterministic learning-augmented algorithm. Experimental evaluations are carried out to evaluate our algorithms using real data access traces.

In this technical report, we focus on reproducing the results reported in the paper “MemXCT: Memory-Centric X-ray CT Reconstruction with Massive Parallelization” [1]. MemXCT is a scalable approach to X-ray Computed Tomography reconstruction which removes redundant computation. We reproduced the single CPU/GPU performance as well as strong scaling experiments. We set up our configurations on Microsoft Azure CycleCloud and have two clusters. One cluster has 4 nodes with 60 CPUs on each node and the other cluster has 4 nodes with 4 NVIDIA V100 GPUs on each node. Both clusters come with InfiniBand. The original author conducted his experiments on Theta and Blue Waters supercomputers. We were able to reproduce part of the results in the original paper, however, failed to produce similar performance on other experiments. This report was submitted as part of the reproducibility challenge in SC20 Student Cluster Competition. Digital artifacts from these experiments are available at: 10.5281/zenodo.5598108.

Wearable Ambulatory Monitor (WAM) is a portable electronic device that monitor the body's function. WAM is used daily and has high potential for home rehabilitation. However, it is limited by its high cost. The arrival of smartphones provides an alternative for motion tracking as they have a number of motion sensors. However, most smartphone applications only use them to detect general movement and not precise motion. A Complementary Feedback Filter (CFF) is proposed and developed to fuse readings from the smartphone's noisy motion sensors in order to get accurate orientation of body segments. The experimental results have shown that movement orientation accuracy obtained from applying CFF on smartphone sensors is comparable to XSENS Awinda, a commercial motion tracking system. The combination of the proposed CFF and the smartphone's motion sensors are then applied to RehabPartner, a motion tracking application for the elderly to do rehabilitation exercises at home.

With the rapid expansion of cities around the world, large number of movements are made daily as people commute from their homes to their destinations, including workplaces. From these movements, trends and patterns can be derived which in turn, can provide valuable insights for urban planning. This is particularly relevant in the 'smart cities' context. However, such movement data are often difficult to gather and analyse without infringing on privacy rights, especially with the increasing concerns on privacy issues. This paper reports on the use of aggregated mobile phone tracking data together with train network data to analyse movement patterns in, out, and within La Défense (Paris' business district). The findings can assist city planners by providing a better understanding of people's travel patterns.

Image outlier detection has been an important research issue for many computer vision tasks . However, most existing outlier detection methods fail in the high-dimensional image datasets. In order to address this problem, we propose a novel image outlier detection method by combining autoencoder with Adaboost (ADAE). By ensembling many weak autoencoders, our method can better capture the statistical correlations among the features of normal data than the single autoencoder . Therefore, the proposed ADAE is able to determine the outliers efficiently. In order to reduce the many parameters in ADAE, we introduce the Sparse Group Lasso (SGL) constraint into the learning objective of ADAE. We combine Adagrad with Proximal Gradient Descent to optimize this additional learning objective. We also propose the multi-objective evolutionary algorithm to determine the best penalty factors of SGL. By evaluating on several famous image datasets, the detection results testify to the outstanding outlier detection performance of ADAE. The evaluation results also show SGL can make the detection model more compact while maintaining the similar detection performance.

Reproducing previous work plays an important role in validating its effectiveness under different experiment settings. In this paper, we reproduce the results from the paper “The vectorization of the Tersoff multi-body potential: an exercise in performance portability” [1]. In particular, we reproduce the accuracy of the reduced precision solvers, the speed-up of the vectorized implementation, and the strong scaling test using Intel Xeon E5-2699 v4 Broadwell CPUs. We show that our reproduced results match the claims made in the original paper.

Availability and all-in-one functionality of smartphones have become a multipurpose personal tool to improve our daily life. Recent advancements in hardware and accessibility of smartphones have spawn huge potential for assistive healthcare, in particular telerehabilitation. However, using smartphone sensors face certain challenges, in particular, accurate orientation estimation, which is usually less of a problem in specialized motion tracking sensor devices. Drift is one of the challenges. We first propose a simple feedback loop complementary filter (CFF) to reduce the error caused by the integration of the gyroscope's data in the orientation estimation. Next, we propose a new and better orientation estimation algorithm which combines quaternion-based kalman filter with corrector estimates using gradient descent (KFGD). We then evaluate CFF's and KFGD's performance on two early-stage rehabilitation exercises. The results show that CFF is capable of fast motion tracking and confirm that the feedback loop can correct the error caused by the integration of gyroscope data. The KFGD orientation estimation is comparable to XSENS Awinda and has shown itself to be stable than and outperforms CFF. KFGD also outperforms the prominent Madgwick algorithm using mobile data. Thus, KFGD is suitable for low-cost motion sensors or mobile inertial sensors, especially during early recovery stage of sport injuries and exercise for the elderly.

Many current computing systems such as clouds and supercomputers charge users for their resource usages. A user's demand is often changing over time, indicating that it is difficult to keep the high resource utilization all the time for cost efficiency. Resource sharing is a classical and effective approach for high resource utilization. In view of the heterogeneous resource demands of users' workloads, multi-resource allocation fairness is a must for resource sharing in such pay-as-you-use computing systems. However, we find that, existing multi-resource fair policies such as Dominant Resource Fairness (DRF), implemented in currently popular resource management systems such as Apache YARN [4] and Mesos [23], are not suitable for the pay-as-you-use computing systems. We show that this is because of their memoryless characteristic that can cause the following problems in the pay-as-you-use computing systems: 1). users can get resource benefits by cheating; 2). users might not be able to get the total amount of resources that they are entitled to in terms of their resource contributions. In this paper, we propose a new policy called H-MRF, which generalizes DRF and Asset Fairness with the long-term notion. We show that it can address these problems and is suitable for pay-as-you-use computing systems. We have implemented it into YARN by developing a prototype called MRYARN. Finally, we evaluate H-MRF using both testbed and simulated experiments. The experimental results show that there are about 1.1 ~1.5 sharing benefit degrees and 1.2× ~ 1.8× performance improvement for users with H-MRF, better than existing fair schedulers.

To address the computing challenge of `big data', a number of data-intensive computing frameworks (e.g., MapReduce, Dryad, Storm and Spark) have emerged and become popular. YARN is a de facto resource management platform that enables these frameworks running together in a shared system. However, we observe that, in cloud computing environment, the fair resource allocation policy implemented in YARN is not suitable because of its memoryless resource allocation fashion leading to violations of a number of good properties in shared computing systems. This paper attempts to address these problems for YARN. Both single-level and hierarchical resource allocations are considered. For single-level resource allocation, we propose a novel fair resource allocation mechanism called Long-Term Resource Fairness (LTRF)for such computing. For hierarchical resource allocation, we propose Hierarchical Long-Term Resource Fairness (H-LTRF) by extending LTRF. We show that both LTRF and H-LTRF can address these fairness problems of current resource allocation policy and are thus suitable for cloud computing. Finally, we have developed LTYARN by implementing LTRF and H-LTRF in YARN, and our experiments show that it leads to a better resource fairness than existing fair schedulers of YARN.

The virtualization technology enables multi-tier web application to be hosted in the Cloud. But the Service Level Agreement (SLA) is likely to be negatively affected when the network traffic is high as they may quickly overload the data center network and increase the response time of the system. Additionally, data centers experience high operational costs due to their considerable amount of energy consumption. The Virtual Machine (VM) technology and VM migration have been widely utilized to deal with such challenges. In this paper, we present design and implementation of an adaptive network-aware VM migration algorithm. The VM and target selection policies are based on the steady state network traffic in the system to minimize the negative effect of migration on other flows on the network. Moreover, we address the high energy consumption of the data center by employing an energy-aware VM placement algorithm. The effectiveness of our proposed VM migration algorithm is evaluated by extensive simulations in CloudSim using real workload traces. We compared two overloading detection policies in our experiments. The results show that our algorithm is able to improve the SLA violation (SLAV) and energy consumption up to 73% and 81%, respectively.

In today's big data era, there is a tremendously huge amount of data available. Layman users lack not only the knowledge and experience in data analytics to make sense of these data but also the computational resources for executing the analytics. In this paper, we propose and develop a layman analytics system (LAS), which provides the layman users with a scalable and ready-to-use analytics tool to automatically generate analytics workflows for classification tasks. The LAS is designed to benefit from existing open-source data analytics tools using generic ontological modeling of analytics operators from these tools as well as adaptive constraint refinement for metadata learning. Moreover, the LAS can be deployed on both public and private clouds to cater to the need of scalable computing and easy maintenance. To demonstrate the performance of the LAS, we conducted experiments with 114 data sets obtained from the University of California Irvine Machine Learning Repository. The workflows generated by the LAS were benchmarked against the OpenML whereby each data set has a range of classification accuracy obtained using classifiers designed and fine-tuned by data experts. The comparisons showed that 87 out of 114 data sets have exceeded the 50th percentile of the benchmark data. Among these 87 data sets, the LAS outperforms the 90th percentile of the benchmarks on 49 data sets.

A series of online multi-task learning (OMTL) algorithms have been proposed to avoid the expensive training cost and poor adaptability of traditional batch multi-task learning (MTL) algorithms in recent years. However, these OMTL algorithms usually assume that all tasks are closely related, which may not hold in practical scenarios. More importantly, their theoretical reliability is weakened due to the lack of proof on the cumulative regrets. To overcome these limitations, we present a robust online multi-task classification framework (ROM) and its two optimization algorithms (ROM-PGD, ROM-RDA). The proposed algorithms can not only automatically capture the common features among all tasks and individual features for each task, but also identify the potential existence of outlier task. Theoretically, we prove that the regret bounds of these two algorithms are sub-linear compared with the best separating algorithm in hindsight. Empirical studies on both synthetic and real-world datasets also demonstrate the effectiveness of our proposed algorithms when compared with the state-of-the-art OMTL algorithms.

For a subset of physically challenged people, assistive technologies, such as alternative form of text entry, can be of tremendous benefit. However, speed of text entry with current methods limits their more widespread adoption. Eye gaze technologies have potential for text entry, but still tend to be relatively slow. Recently dwell-free eye-typing systems have been proposed, but can be vulnerable to common text entry problems, such as selection of the wrong letters. In this article, a recognition approach for inferring the words which the user intends to type is proposed. The method is robust to missing letters and even when a neighboring letter on the keyboard is incorrectly selected. Simulation and experiment results suggest that our proposed approach has better accuracy and more resilience to common text entry errors than other currently proposed dwell-free systems.

Cloud computing is emerging as a promising alternative to supercomputers for some high-performance computing (HPC) applications. With cloud as an additional deployment option, HPC users and providers are faced with the challenges of dealing with highly heterogeneous resources, where the variability spans across a wide range of processor configurations, interconnects, virtualization environments, and pricing models. In this paper, we take a holistic viewpoint to answer the question-why and whoshould choose cloud for HPC, for what applications, and how should cloud be used for HPC? To this end, we perform comprehensive performance and cost evaluation and analysis of running a set of HPC applications on a range of platforms, varying from supercomputers to clouds. Further, we improve performance of HPC applications in cloud by optimizing HPC applications' characteristics for cloud and cloud virtualization mechanisms for HPC. Finally, we present novel heuristics for online application-aware job scheduling in multi-platform environments. Experimental results and simulations using CloudSim show that current clouds cannot substitute supercomputers but can effectively complement them. Significant improvement in average turnaround time (up to 2X)and throughput (up to 6X) can be attained using our intelligent application-aware dynamic scheduling heuristics compared tosingle-platform or application-agnostic scheduling.

MapReduce is a popular parallel computing paradigm for large-scale data processing in clusters and data centers. A MapReduce workload generally contains a set of jobs, each of which consists of multiple map tasks followed by multiple reduce tasks. Due to 1) that map tasks can only run in map slots and reduce tasks can only run in reduce slots, and 2) the general execution constraints that map tasks are executed before reduce tasks, different job execution orders and map/reduce slot configurations for a MapReduce workload have significantly different performance and system utilization. This paper proposes two classes of algorithms to minimize the makespan and the total completion time for an offline MapReduce workload. Our first class of algorithms focuses on the job ordering optimization for a MapReduce workload under a given map/reduce slot configuration. In contrast, our second class of algorithms considers the scenario that we can perform optimization for map/reduce slot configuration for a MapReduce workload. We perform simulations as well as experiments on Amazon EC2 and show that our proposed algorithms produce results that are up to 15 ~ 80 percent better than currently unoptimized Hadoop, leading to significant reductions in running time in practice.

The Internet has become the most vulnerable part of critical civil infrastructures. Proactive measures such as early warnings are required to reduce the risk of disasters that can be created using it. With the continuous growth in scale, complexity and variety of networked systems the quality of data is continuously decreasing. This paper investigates the ability to employ Bayesian inference for network scenario analysis with low quality data to produce early warnings. Theoretical account of the approach and experimental results using a real world attack scenario and a real network traffic capture is presented.

Infrastructure-as-a-Service (IaaS) cloud providers offer many elasticities and flexibilities for users to run their systems in the cloud. The monetary cost issues of running those systems in the cloud are hardly ignored and there is less work discussing improving the monetary efficiency of running large scale systems in dynamic cloud environments. In this paper, we focus on improving the monetary efficiency of running Hadoop systems in the dynamic public cloud. In particular, we carry out detailed study on improving the monetary efficiency by leveraging spot instances. From a cloud broker's perspective, we propose a price-aware virtual machine auto-scaling with migration algorithm to improve the monetary efficiency of running Hadoop in the cloud using spot instances. We evaluate our proposed algorithm through simulation using Amazon EC2 spot price traces and real world workload traces. Compared with other baseline algorithms, our approach can improve the monetary efficiency by up to 9.3x.

This paper investigates network traffic before and after a vulnerability called Heart bleed becomes a public issue around March to May, 2014. To detect anomalies and potential threats due to the vulnerability, a wavelet entropy-based change-point detection method is proposed and compared with three other methods: prediction-based, clustering-based and Fourier transform-based. We show that the proposed wavelet entropy-based method outperforms the others in terms of ease of parameter setting, false alarm and detection accuracy. Using the proposed method and a visualization tool, we have studied Heart bleed vulnerability and successfully captured changes in packet volume and flow.

Recent trends indicate that the pay-as-you-go Infrastructure-as-a-Service (IaaS) cloud computing has become a popular platform for big data processing applications, due to its merits of accessibility, elasticity and flexibility. However, the resource demands of processing workloads are often varying over time for individual users, implying that it is hard for a user to keep the high resource utilization for cost efficiency all the time. Resource sharing is a classic and effective approach to improve the resource utilization via consolidating multiple users' workloads. However, we show that, current existing fair policies such as max-min fairness, widely adopted and implemented in many popular big data processing systems including YARN, Spark, Mesos, and Dryad, are not suitable for pay-as-you-go cloud computing. We show that it is because of their memory less allocation feature which can arise a series of problems in the pay-as-you-go cloud environment, namely, cost-inefficient workload submission, untruthfulness and resource-as-you-pay unfairness. This paper presents these problems and outlines our plans to address them for pay-as-you-go cloud computing. We introduce our preliminary work done on the single-resource fairness and our ongoing work for multi-resource fairness, and outline our future work.

Approaching a comprehensive performance benchmark for on-line transaction processing (OLTP) applications in a cloud environment is a challenging task. Fundamental features of clouds, such as the pay-as-you-go pricing model and unknown underlying configuration of the system, are contrary to the basic assumptions of available benchmarks such as TPC-W or RUBiS. In this paper, we introduce a systematic performance benchmark approach for OLTP applications on public clouds that use virtual machines(VMs). We propose WPress benchmark, which is based on the widespread blogging software, WordPress, as a representative OLTP application and implement an open source workload generator. Furthermore, we utilize a CPU micro-benchmark to investigate CPU performance of cloud-based VMs in greater detail. Average response time and total VM cost are the performance metrics measured by WPress. We evaluate small and large instance types of three real-life cloud providers, Amazon EC2, Microsoft Azure and Rackspace cloud. Results imply that Rackspace cloud has better average response times and total VM cost on small instances. However, Microsoft Azure is preferable for large instance type.

Recent studies have suggested significant differences in motor performances of Parkinson's Disease (PD) patients who have L-dopa induced dyskinesias (LIDs), even when off of L-dopa medication. The pathophysiology of LIDs remains obscure, so applying data-mining techniques to the patients' motor performance may provide some heuristic insight. This paper investigated visually-guided tracking performance of PD patients using data mining techniques to reveal the differences between dyskinesia and non-dyskinesia patients. We found that K-means clustering of the root mean square (RMS) tracking error at faster tracking speeds and with ambiguous visual stimuli was able to effectively discriminate between the two groups with 77.8% accuracy. Decision tree classification was less accurate (68.4%) and determined that years since diagnosis was the best feature to distinguish between groups. Our results suggest that data mining methodologies may provide novel insights into features of the neurovegetative disease.

Fair resource allocation is a key building block of any shared computing system. However, MemoryLess Resource Fairness (MLRF), widely used in many existing frameworks such as YARN, Mesos and Dryad, is not suitable for pay-as-you-use computing. To address this problem, this paper proposes Long-Term Resource Fairness (LTRF), a novel fair resource allocation mechanism. We show that LTRF satisfies several highly desirable properties. First, LTRF incentivizes clients to share resources via group-buying by ensuring that no client is better off in a computing system that she buys and uses individually. Second, LTRF incentivizes clients to submit non-trivial workloads and be willing to yield unneeded resources to others. Third, LTRF has a resource-as-you-pay fairness property, which ensures the amount of resources that each client should get according to her monetary cost, despite that her resource demand varies over time. Finally, LTRF is strategy-proof, since it can make sure that a client cannot get more resources by lying about her demand. We have implemented LTRF in YARN by developing LTYARN, a long-term YARN fair scheduler, and shown that it leads to a better resource fairness than other state-of-the-art fair schedulers.

Trust and security are the top concerns of cloud computing users and key barriers to widespread uptake of cloud computing services across industries. The resistance towards cloud computing is especially experienced in industries handling sensitive data, such as healthcare, government, banking, and so on. While cloud computing brings about several conveniences for end-users, many new issues surfaced from cloud computing's promise of elasticity and availability. With these new issues, traditional security and trust techniques may not be able to fully resolve cloud computing's trust and security problems. This special issue focuses on a broad range of research challenges and issues in trust and security in cloud computing. By addressing trust management, cloud attack vectors, data sharing, dynamic user privileges and high throughput encryption implementations, we addressed some of the top concerns in trust and security in cloud computing. Cloud providers describe their promised behaviour by way of service level agreements (SLAs). However, providers offering similar functionality may have SLAs that are often inconsistent with the aspects considered important by customers. Customers face problems identifying a trustworthy cloud provider solely on the basis of its SLA. To enable reliable customer identification of trustworthy cloud providers, Habib et al. proposed a multi-faceted trust management system architecture for cloud computing marketplaces and related approaches. Their approach provides the means for identifying trustworthy cloud providers in terms of different attributes, such as compliance, data governance and information security. Their proposed trust management system also utilized real data from the Cloud Security Alliance's Consensus Assessment Initiative Questionnaire as one of the sources of trust information. Cloud services are vulnerable to Internet Protocol (IP) prefix hijacking due to their dependence on routing infrastructure. It is important to understand what impact a prefix hijacking attack can cause and how the number and locations of participants can affect the attacking results. Liu et al. modelled this problem as an attack planning task and solved it by applying a form of genetic algorithm. By analyzing the best solution to the problem, they found that the type of victims plays a more important role in IP prefix hijacking than that of attackers. Attackers can gain great impact even when the prefixes of a small number of victims are hijacked. For attack planning, the degree of an autonomous system is a major criterion to be considered. These findings are critical, as they secure cloud networks by preventing and eliminating IP prefix hijacking attacks. Secure data sharing on Cloud storage is an important emerging area in cloud security research. However in this area, dynamic privileges among user groups are usually not considered. In many circumstances, some users may have unnecessarily higher privileges than others. There may also be cases where the data owner may want to dynamically control the privileges in data sharing. Zhao and Li presented an efficient framework for data sharing systems to achieve dynamic privileges, based on the chameleon hash function and one-way function. With this framework, any data sharing and access control scheme can be turned into a dynamic privileged scheme, in which the data owner can change the group of each user dynamically and change the structure of privileges flexibly when it is needed. The proposed framework also requires much less storage than previous schemes in handling dynamic privileges among users. In cloud computing, the techniques for data protection need to be efficient, that is, high performance and throughput. Rahimunnisa et al. proposed a high throughput architecture for the hardware implementation of Advanced Encryption Standard algorithm. Their work, implemented in a Field Programmable Gate Array (FPGA), mainly targets low-cost embedded applications and introduces parallel operation in folded architectures to obtain better throughput—giving a high 37.1 Gb/s throughput with a maximum frequency of 505.5 MHz—which is 20% higher than the maximum throughput reported in the literature. We would like to thank the Editor-in-Chief, Professor Hsiao-Hwa Chen, and Co-Editor-in-Chief, Professor Hamid R. Sharif, for the opportunity to host this special issue in Wiley's Security and Communication Networks. We also thank all the authors who contributed to this Special Issue for publication consideration. Last but not least, we thank the contributions of expert reviewers who provided invaluable advice and recommendations through the revisions. Dr Ryan Ko is a Senior Lecturer at the University of Waikato, New Zealand, and a Research Advisor for Cloud Security Alliance (CSA)'s Asia Pacific region. He established New Zealand's first Master's Degree in Cyber Security and dedicated Cyber Security Lab at the University of Waikato and is the principal investigator for the STRATUS research project. His main research areas are cyber security, cloud data provenance, and cloud computing security and trust. Prior to joining the faculty, he was a lead computer scientist with Hewlett–Packard (HP) Labs' Cloud and Security Lab. Recipient of the CSA Ron Knode Service Award, he is active as co-chair and board member of several cyber security industry consortia and chapters. He is also active as subject matter expert/item writer in the development of the (ISC)2 Certified Cloud Security Professional certification. He holds a BEng (Hons) (Computer Engineering) and PhD from Nanyang Technological University, Singapore, and is a member of IEEE and ACM. Markus Kirchberg is an Adjunct Associate Professor at the National University of Singapore, Singapore. Markus is the Head of Visa Research, Asia Pacific. He joined VISA Inc. in May 2012 as Lead Research Scientist, and his responsibility is for technology innovation inside Visa Labs. Markus's main technical focus areas include data management, cloud computing, and large-scale data analytics. Markus also holds an Adjunct Associate Professor position with the Department of Computer Science at the National University of Singapore (NUS), where he teaches Master's Degree courses focusing on applied data analytics. Prior to joining VISA Inc., Markus worked as Expert at the Cloud and Security Lab, HP Labs Singapore (2010–2012); Research Fellow and Principal Investigator at the Institute for Infocomm Research, A*STAR in Singapore (2007–2010); and Lecturer at Massey University in New Zealand (2000–2007). He holds a PhD in Information Systems from Massey University, New Zealand (2007, part-time) and a Master of Sciences in Computer Science Degree from the Clausthal University of Technology, Germany (2000). Bu Sung Lee received his BSc (Hons) and PhD from the Electrical and Electronics Department, Loughborough University of Technology, U.K., in 1982 and 1987, respectively. He is currently an Associate Professor in the Nanyang Technological University, Singapore. Bu Sung Lee holds a joint appointment as Director, Service Platform Lab, HP Labs. Singapore from July 2010 till end-June 2012. Bu Sung Lee has been very active in the area of establishing a Research and Education Network locally as well as globally. He is the founding president and current President of Singapore Advanced Research and Education Network (SingAREN) and chair of the TransEurasia Information Network Cooperation Center (TEIN*CC) governors, which manages the TEIN regional network. Bu Sung Lee has published over 200 peer-reviewed papers. His research covers cloud computing, data analytics, and network. His particular interest areas are in data replication, scheduling, and more recently in data fusion as applied to cyber security.

High availability of software components has long been studied. For a software system, when unavailability of a component has caused a suspension of the system, the system has to be recovered or resumed as soon as possible. To substitute an unavailable software component with a backup copy is therefore unavoidable in achieving high availability of software systems. In this paper, in comparison with using redundancies, we take an alternative approach that steps away from the physical code equivalence of the software but focuses more on the equivalence in using the function unit without concerning about the implementation itself. We investigate the problem of Web service availability in service-oriented software systems and then report a framework for Web service availability in such systems using automated and rule-based Web service substitution. The framework takes a novel approach to manage the runtime replacement of services, combining (i) an approach that classifies services using co-occurrence of terms in various tags of the service descriptions, (ii) an approach to establish the compatibility and substitution of service operation interfaces and (iii) a middleware for handling service replacements. Our approach is designed to address the problem of Web service availability from the client side and assumes that the client has no control of the Web service providers. This is a completely distributed approach in comparison with other related work and presents a valuable benefit of client orientation. As two additional distinguishing characteristics, our framework also meets the challenges of (i) semantic heterogeneity of Web services in identifying substitute service and (ii) transparency and independence in handling unavailability at the level of Web services. We show in our experiments that the service substitute identification based on the proposed framework achieves a best precision of 85%. We demonstrate our implementation of the middleware for service unavailability handling in the framework. We also present experiments on service substitution within a demo business application in the presence of unavailability. Copyright © 2012 John Wiley & Sons, Ltd.

MapReduce is a popular computing paradigm for large-scale data processing in cloud computing. However, the slot-based MapReduce system (e.g., Hadoop MRv1) can suffer from poor performance due to its unoptimized resource allocation. To address it, this paper identifies and optimizes the resource allocation from three key aspects. First, due to the pre-configuration of distinct map slots and reduce slots which are not fungible, slots can be severely under-utilized. Because map slots might be fully utilized while reduce slots are empty, and vice-versa. We propose an alternative technique called Dynamic Hadoop SlotAllocation by keeping the slot-based model. It relaxes the slot allocation constraint to allow slots to be reallocated to either map or reduce tasks depending on their needs. Second, the speculative execution can tackle the straggler problem, which has shown to improve the performance for a single job but at the expense of the cluster efficiency. In view of this, we propose Speculative Execution Performance Balancing to balance the performance tradeoff between a single job and a batch of jobs. Third, delay scheduling has shown to improve the data locality but at the cost of fairness. Alternatively, we propose a technique called Slot PreSchedulingthat can improve the data locality but with no impact on fairness. Finally, by combining these techniques together, we form a step-by-step slot allocation system called DynamicMR that can improve the performance of MapReduce workloads substantially. The experimental results show that our DynamicMR can improve the performance of Hadoop MRv1 significantly while maintaining the fairness, by up to 46~115 percent for single jobs and 49~112 percent for multiple jobs. Moreover, we make a comparison with YARN experimentally, showing that DynamicMR outperforms YARN by about 2~9 percent for multiple jobs due to its ratio control mechanism of running map/reduce tasks.

Generally, H.264/AVC video coding standard with hierarchical bipredictive picture (HBP) structure outperforms the classical prediction structures such as “IPPP...” and “IBBP...” through better exploitation of data correlation using reference frames and unequal quantization setting among frames. However, multiple reference frames (MRFs) techniques are not fully exploited in the HBP scheme because of the computational requirement for B-frames, unavailability of adjacent reference frames, and with no explicit sorting of the reference frames for foreground or background being used. To exploit MRFs fully and explicitly in background referencing, we observe that not a single frame of a video is appropriate to be the reference frame as no one covers adequate background of a video. To overcome the problems, we propose a new coding scheme with the HBP, which uses the most common frame in scene (McFIS), generated by background modeling, as a long-term reference (LTR) frame for the third unipredictive reference frame, so that foreground and background areas are expected to be referenced from the two frames in the HBP structure and the McFIS, respectively. There are two approaches to generate McFIS under the proposed methodology. In the first approach, we generate a McFIS using a number of original frames of a scene in a video and then encode it as an I-frame with a higher quality. For the rest of the scene, this generated I-frame is used as an LTR frame. In the second approach, we generate an McFIS from the decoded frames and then use it as an LTR frame, without the need to encode the McFIS. The first and the second approaches are suitable for a video with static background and dynamic background, respectively. In general, the second approach requires more computational time than that of the the first approach. The experiments confirm that the proposed scheme outperforms three state-of-the-art algorithms by improving the image quality significantly with reduced computational time.

We focus on the problems of cooperative virtual machine management of cloud users in a smart grid environment. In such an environment, the cloud users can cooperate to share the available computing resources in private cloud and public cloud to reduce the total cost. To achieve an optimal and fair solution, we develop the framework composed of the virtual machine allocation, cost management, and cooperation formation models. The problem is challenging due to the uncertainties (e.g., uncertain power price and unpredictable users' demand). Therefore, for the virtual machine allocation, we develop the stochastic programming model to obtain the optimal solutions of virtual machines to be hosted in the local data center, to be hosted on the public cloud servers, or to be migrated to the data centers of other cooperative cloud users. Then, among cooperative cloud users, the cost management is formulated as the coalitional game whose fair share of the total cost is obtained as the Shapley value. Next, given that the cloud users are rational, we formulate the cooperation formation as the network formation game to analyze the stability of the cooperation. In the experiment, we evaluate our proposed framework with real trace data. The results clearly show that the cooperative virtual machine management can achieve the minimum total cost of cloud users compared with expected value and worst case formulations.

This paper proposes a Hybrid Protocol for Application Level Multicast with Dual Metric (HPAM-D) for live video streaming. HPAM-D constructs data distribution trees based on two performance metrics, namely, latency to source and loss rate experienced by the clients. The protocol is evaluated against the latency based single-metric HPAM and a modified HPAM-L which aims only to optimize loss rate regardless of the latency to source. HPAM-L acts as a control experiment in the evaluation. Another novel feature introduced in HPAM-D is the detection of likely local congestion via a heuristic based on the computation of a client's relative loss rate with respect to the data received by its parent as opposed to the commonly used absolute loss rate experienced by a client. Simulation results show that HPAM-D not only helps a client to reduce unnecessary parent switches thereby reducing protocol overheads but also maintains a low loss rate for its clients without compromising the RDP performance compared to the latency based, single-metric HPAM.

Cloud computing is emerging as an alternative to supercomputers for some of the high-performance computing (HPC) applications that do not require a fully dedicated machine. With cloud as an additional deployment option, HPC users are faced with the challenges of dealing with highly heterogeneous resources, where the variability spans across a wide range of processor configurations, interconnections, virtualization environments, and pricing rates and models. In this paper, we take a holistic viewpoint to answer the question - why and who should choose cloud for HPC, for what applications, and how should cloud be used for HPC? To this end, we perform a comprehensive performance evaluation and analysis of a set of benchmarks and complex HPC applications on a range of platforms, varying from supercomputers to clouds. Further, we demonstrate HPC performance improvements in cloud using alternative lightweight virtualization mechanisms - thin VMs and OS-level containers, and hyper visor- and application-level CPU affinity. Next, we analyze the economic aspects and business models for HPC in clouds. We believe that is an important area that has not been sufficiently addressed by past research. Overall results indicate that current public clouds are cost-effective only at small scale for the chosen HPC applications, when considered in isolation, but can complement supercomputers using business models such as cloud burst and application-aware mapping.

MapReduce is a popular parallel computing paradigm for large-scale data processing in clusters and data centers. However, the slot utilization can be low, especially when Hadoop Fair Scheduler is used, due to the pre-allocation of slots among map and reduce tasks, and the order that map tasks followed by reduce tasks in a typical MapReduce environment. To address this problem, we propose to allow slots to be dynamically (re)allocated to either map or reduce tasks depending on their actual requirement. Specifically, we have proposed two types of Dynamic Hadoop Fair Scheduler (DHFS), for two different levels of fairness (i.e., cluster and pool level). The experimental results show that the proposed DHFS can improve the system performance significantly (by 32% ~ 55% for a single job and 44% ~ 68% for multiple jobs) while guaranteeing the fairness.

General Purpose Graphics Processing Units (GPGPU) have been used in high performance computing platforms to accelerate the performance of scientific applications such as simulations. With the increased computing resources required for large-scale network simulation, one GPU device may not have enough memory and computation capacities. It is therefore necessary to enhance the system scalability by introducing multiple GPU devices. It is also attractive to investigate the performance scalability of Multi-GPU simulations. This paper describes the simulation of information propagation on multiple GPU devices, including the optimized network simulation algorithms, the network partitioning and replication strategy, and the data synchronization scheme. The experimental results for scalable random networks show that the number of simulation steps, computation time, synchronization time, and data transfer time all affect the overall simulation performance. In order to compare with random networks, we also conduct simulations of scale-free networks. We can observe that the node replication ratio in scale-free networks is smaller than that in random networks and therefore the cost of data transfer and synchronization is significantly reduced. This indicates that the network structure is also an important factor that influences the simulation performance in a Multi-GPU system.

Recommending learning materials for e-learning systems often encounters two issues: how to classify and organize learning materials and how to make effective recommendations. In this paper, we propose a new algorithm to handle these two problems. Specifically, we compile each learning material to concepts according to their relevance which is modeled as the length of a term-weight vector. Then recommendations are generated by taking into account the document’s similarity with some good learning material, the personalized time-aware usefulness of the learning material, the concepts of the learning material as well as their difficulty levels. Experimental results based on a small sample demonstrate the effectiveness of our method in terms of knowledge gain obtained.

This paper studies event recognition in a building based on the patterns of power consumption. It is a big challenge to identify what kinds of events happened in a building without additional devices such as camera and motion sensors, etc. Instead, we learn when and how the events happened from the historical record of power consumption and apply the lesson into the design of an event recognition system (ERS). The ERS will find out abnormal power usage to avoid wasting power, which leads to the energy savings in a building. The ERS involves big data analytics with a large size of dataset collected in a real time. Such a data intensive system is usually viewed as a workflow. A workflow management is a significant task of the system requiring data analysis in terms of the system scalability to maintain high throughput or fast speed analysis. We propose a workflow framework that allows users to perform remote and parallel workflow execution, whose tasks are efficiently scheduled and distributed in cloud computing environment. We run the ERS as a target system for the proposed framework with power consumption data (whose size is approximately 20GB or more) collected from each of over 240 rooms in a building at Dept. of Engineering, Tokyo University in 2011. We show that the proposed framework accelerates the speed of data analysis by providing scaling infrastructure and parallel processing feature utilizing cloud computing technologies. We also share our experience and results on the big data analytics and discuss how the studies contribute to achieve Green Campus.

Formulation of appropriate data analytics workflows requires intricate knowledge and rich experiences of data analytics experts. This problem is further compounded by continuous advancement and improvement in analytical algorithms. In this paper, a generic non-domain specific solution for the creation of appropriate workflows targeted at supervised learning problems is proposed. Our adaptive workflow recommendation engine based on collaborative analytics matches analytics needs with relevant workflows in repository. It is capable of picking workflows with better performance as compared to randomly selected workflows. The recommendation engine is now augmented by a workflow optimizer that applies genetic programming to further improve the recommended workflows through iterative evolution, leading to better alternative workflows. This unique Collaborative Analytics Recommender System is tested on seven UCI benchmark datasets. It is shown that the final workflows produced by the system could closely approximate, in terms of accuracy, the best workflows that analytics experts could possibly design.

SUMMARY Hadoop, based on the popular MapReduce framework, is an open‐source distributed computing framework that has been gaining much popularity and usage. It aims to allow programmers to focus on building applications that deals with processing large amount of data, without having to handle other issues when performing parallel computations. However, tuning the performance of Hadoop applications is not an easy task due to the level of abstraction of the framework. In this paper, we present three case studies and some of the challenges and issues that are to be considered in performance tuning when running applications in Hadoop. The focus is mainly on the impact of input data on Hadoop's performance and how they can be tuned. Copyright © 2011 John Wiley & Sons, Ltd.

There are many ways to build a predictive model from data. Besides the numerous classification or regression algorithms to choose from, there are countless possibilities of useful data transformation prior to modeling. To assist in discovering good predictive analytics workflows, we introduced recently a collaborative analytics system that allows workflow sharing and reuse. We designed a recommendation engine for the system to enable matching of analytics needs with relevant workflows stored in repository. The engine relies on meta-predictive modeling of traffic-analysis workflow-characteristics. In this paper, we present a feasibility study of applying this collaborative analytics system to predict traffic congestion. Different ways to build predictive models from traffic dataset are pooled as shared workflows. We demonstrate that through dynamic recommendation of workflows that are suitable for the real-time varying traffic data, a reliable congestion prediction can be achieved. The promising results showcase that systematic collaboration among data scientists made possible by our system can be a powerful tool to produce very accurate prediction from data.

The fast growth of cloud service offerings has attracted more enterprises to migrate their IT applications into cloud. Nonetheless, complex enterprise user requirements, especially interdependent relations across them, raise new challenges of cloud service selection. In addition, a major concern for these enterprises is ensuring compliance with their policies on the use of cloud services. In this paper, we present a systematic framework, based on formal verification and constraint solving techniques, to help enterprises tackle problems when adopting cloud computing. Our framework enables automatic detection of conflicts covering violation of enterprise policies and inconsistency of user requirements, and explanation generation which identifies problematic user requirements. The framework next select automatically cloud services which satisfy all enterprise policies and user requirements (with interdependent relations). We have prototyped and successfully applied our approach to projects which manage heterogeneous cloud infrastructure services for large enterprises.

Version control systems are an indispensable part of the software engineering process. Without them, large software development projects would soon incur in loss of productivity and severe maintenance issues. The revision history is invaluable during reviewing process as well as bug hunting. With the advent of Web 2.0, many tasks have been migrated to a cloud environment, and among them is software development. This spikes the need for a fast, always-on version control system residing in the cloud. Moreover, the cloud computational power can be leveraged to provide a higher degree of automation. In this paper, we describe Protego, a cloud-based version control system targeting flexibility and automation. Flexibility in the way projects and histories are recorded and updated, automation in the way users interact with the system, with a reduced need for everyday tedious operations.

In this paper, we present our design of a Processing Element (PE) Aware MapReduce base framework, Pamar. Pamar is designed for supporting distributed computing on clusters where node PE configurations are asymmetric on different nodes. Pamar's main goal is to allow users to seamlessly utilize different kinds of processing elements (e.g., CPUs or GPUs) collaboratively for large scale data processing. To show proof of concept, we have incorporated our designs into the Hadoop framework and tested it on cluster environments having asymmetric node PE configurations. We demonstrate Pamar's ability to identify PEs available on each node and match-make user jobs with nodes, base on job PE requirements. Pamar allows users to easily parallelize applications across large datasets and at the same time utilizes different PEs for processing different classes of functions efficiently. The experiments show improvement in job queue completion time with Pamar over clusters with asymmetric nodes as compared to clusters with symmetric nodes.

This paper reports on the development of the Cloud Oriented Data Analytics (CODA) framework which has functions for composing, managing, and processing workflows for data analytics in cloud computing. The framework provides a number of reusable software components for data analytics to users which can be composed as workflows through well-known workflow composers, e.g., RapidMiner, Taverna, and JOpera. In particular, workflow scheduling, workflow recommendation, resource provisioning, resource monitoring, data locality, and security for the workflow computation are addressed by the framework. By using the framework, we demonstrate that workflows can be easily composed and processed in cloud computing. By coordinating the submitted workflows, we can obtain a significant improvement in performance.

This paper presents a scheme to optimize the mapping of HPC applications to a set of hybrid dedicated and cloud resources. First, we characterize application performance on dedicated clusters and cloud to obtain application signatures. Then, we propose an algorithm to match these signatures to resources such that performance is maximized and cost is minimized. Finally, we show simulation results revealing that in a concrete scenario our proposed scheme reduces the cost by 60% at only 10-15% performance penalty vs. a non optimized configuration. We also find that the execution overhead in cloud can be minimized to a negligible level using thin hypervisors or OS-level containers.

In this paper, we address the energy-efficient contact discovery issue in Pocket Switched Networks (PSNs), in which the nodes' mobility pattern shows strong social property. In a PSN, although the end-to-end connection may be disconnected most of the time for a given source-destination pair, several nodes periodically gather at certain hot spots and form well connected clusters. Based on such mobility pattern, cooperation among nodes is utilized in our contact discovery design. The nodes that have already joined a cluster collaboratively wake up to discover new contacts. With local synchronization, the nodes can be operated in sleep mode more frequently, leading to high energy efficiency. Both the theoretical analysis and simulation results show that our cooperative duty cycling (CDC) greatly reduces the energy consumption while achieving comparable data delivery performance.

In this paper we propose MaxCD - a joint multi-flow scheduling and cooperative downloading protocol for drive-thru networks, with the goal of maximizing the amount of data packets that can be downloaded per drive-thru. Based on the macro-level opportunistic scheduling and node cooperation, the best wireless link(s) (with the highest data rate) between the roadside unit (RSU) and vehicular users are fully utilized. In addition, a multichannel collision-free relay mechanism is designed to address the reliable and fast data exchange issue when the vehicular users are outside the service area of the RSU. Our theoretical analysis vindicates the performance gain of the cooperation and extensive simulations demonstrate the efficiency of MaxCD.

This paper proposes a multi-tenant workflow framework that allows users to create data analytic workflows whose tasks are efficiently scheduled and distributed in cloud computing environment. We provide a demo of an event room assignment (ERA) as a test application of the framework. The ERA dynamically and automatically assigns registered events (e.g., meetings, classes, conferences, etc.) to available rooms meeting the user requirements such as the event size, purpose, reservation period, etc. The assignment will lead to the energy efficiency with respect to the power usage (e.g., lighting, ventilation, devices, etc.), and the energy savings can be achieved without affecting people's comfort. We run the ERA with power consumption data (whose size is approximately 50GB) collected from each of over 200 rooms in a building at Dept. of Engineering, Tokyo University. Through the demonstration, we will show that the proposed framework accelerates the speed of data analysis by providing user-friendly workflow composition and parallel processing features utilizing cloud computing technologies.

As REST (Representational State Transfer)-ful services are closely coupled to the HTTP (Hypertext Transfer Protocol), which eventually sits above the connection-based TCP (Transmission Control Protocol), it is common for RESTful services to experience latency and transfer inefficiencies especially in situations requiring the services to transfer large-scale data (i.e. above gigabytes of data) in RESTful workflows. Such inefficiencies are undesirable and impractical, and are compounded for RESTful service orchestrations in data-intensive industries such as Big Data analytics, cloud computing and life sciences. In this paper, we propose a non-invasive novel technique, Fast-Optimised-REST (FOREST), which enables RESTful services to overcome the traditional bottlenecks experienced during transfer of large sets of data. The initial experimental results show promise and demonstrated very significant reductions of up to 80% from original REST-ful data transfer times for extremely large data sets.

We propose an adaptive power management (APM) algorithm for a data center with an objective to minimize the total cost of power bought from an electrical grid. This APM algorithm is developed for a smart grid environment which is envisioned to be a cooperative, responsive, and economical power system. In particular, APM algorithm takes the spot power price from an electrical grid, the power supply from a renewable power source, and users' demand in terms of application workload processing into account when managing the power consumption. Therefore, an APM algorithm is considered to be the demand side management in a smart grid. To obtain an optimal decision of the APM algorithm, an optimization model based on stochastic programming with multi-stage recourse is developed. This optimization model considers various uncertainties and is able to determine the optimal solution for the APM algorithm. The APM algorithm is evaluated by numerical studies. The numerical results clearly show that the APM algorithm can minimize the power cost of a data center.

Many private and/or public organizations have been reported to create and monitor targeted Twitter streams to collect and understand users' opinions about the organizations. Targeted Twitter stream is usually constructed by filtering tweets with user-defined selection criteria e.g. tweets published by users from a selected region, or tweets that match one or more predefined keywords. Targeted Twitter stream is then monitored to collect and understand users' opinions about the organizations. There is an emerging need for early crisis detection and response with such target stream. Such applications require a good named entity recognition (NER) system for Twitter, which is able to automatically discover emerging named entities that is potentially linked to the crisis. In this paper, we present a novel 2-step unsupervised NER system for targeted Twitter stream, called TwiNER. In the first step, it leverages on the global context obtained from Wikipedia and Web N-Gram corpus to partition tweets into valid segments (phrases) using a dynamic programming algorithm. Each such tweet segment is a candidate named entity. It is observed that the named entities in the targeted stream usually exhibit a gregarious property, due to the way the targeted stream is constructed. In the second step, TwiNER constructs a random walk model to exploit the gregarious property in the local context derived from the Twitter stream. The highly-ranked segments have a higher chance of being true named entities. We evaluated TwiNER on two sets of real-life tweets simulating two targeted streams. Evaluated using labeled ground truth, TwiNER achieves comparable performance as with conventional approaches in both streams. Various settings of TwiNER have also been examined to verify our global context + local context combo idea.

Data leakages out of cloud computing environments are fundamental cloud security concerns for both the end-users and the cloud service providers. A literature survey of the existing technologies revealed the inadequacies of current technologies and the need for a new methodology. This position paper discusses the requirements and proposes a novel auditing methodology that enables tracking of data transferred out of Clouds. Initial results from our prototypes are reported. This research is aligned to our vision that by providing transparency, accountability and audit trails for all data events within and out of the Cloud, trust and confidence can be instilled into the industry as users will get to know what exactly is going on with their data in and out of the Cloud.

As cloud computing and virtualization technologies become mainstream, the need to be able to track data has grown in importance. Having the ability to track data from its creation to its current state or its end state will enable the full transparency and accountability in cloud computing environments. In this paper, we showcase a novel technique for tracking end-to-end data provenance, a meta-data describing the derivation history of data. This breakthrough is crucial as it enhances trust and security for complex computer systems and communication networks. By analyzing and utilizing provenance, it is possible to detect various data leakage threats and alert data administrators and owners; thereby addressing the increasing needs of trust and security for customers' data. We also present our rule-based data provenance tracing algorithms, which trace data provenance to detect actual operations that have been performed on files, especially those under the threat of leaking customers' data. We implemented the cloud data provenance algorithms into an existing software with a rule correlation engine, show the performance of the algorithms in detecting various data leakage threats, and discuss technically its capabilities and limitations.

It is highly desirable and challenging for a wireless ad hoc network to have self-organization properties in order to achieve wide network characteristics. Studies have shown that Small-World properties, primarily low average path length (APL) and high clustering coefficient, are desired properties for networks in general. However, due to the spatial nature of the wireless networks, achieving small-world properties remains highly challenging. Studies also show that, wireless ad hoc networks with small-world properties show a degree of distribution that lies between geometric and power law. In this paper, we show that in a wireless ad hoc network with non-uniform node density with only local information, we can significantly reduce the APL and retain the clustering coefficient. To achieve our goal, our algorithm first identifies logical regions using the Lateral Inhibition technique, then identifies the nodes that beamform and finally the beam properties using Flocking. We use Lateral Inhibition and Flocking because they enable us to use local state information as opposed to other techniques. We support our work with simulation results and analysis, which show that a reduction of up to 40% can be achieved for a high-density network. We also show the effect of hopcount used to create regions on APL, clustering coefficient and connectivity.

In this paper, we propose a low-complexity video coding scheme based upon 2-D singular value decomposition (2-D SVD), which exploits basic temporal correlation in visual signals without resorting to motion estimation (ME). By exploring the energy compaction property of 2-D SVD coefficient matrices, high coding efficiency is achieved. The proposed scheme is for the better compromise of computational complexity and temporal redundancy reduction, i.e., compared with the existing video coding methods. In addition, the problems caused by frame decoding dependence in hybrid video coding, such as unavailability of random access, are avoided. The comparison of the proposed 2-D SVD coding scheme with the existing relevant non-ME-based low-complexity codecs shows its advantages and potential in applications.

Discrete cosine transform (DCT) is the orthogonal transform that is most commonly used in image and video compression. The motion-compensation residual (MC-residual) is also compressed with the DCT in most video codecs. However, the MC-residual has different characteristics from a nature image. In this paper, we develop a new orthogonal transform-rotated orthogonal transform (ROT) that can perform better on the MC-residual than the DCT for coding purposes. We derive the proposed ROT based on orthogonal-constrained L1-Norm minimization problem for its sparse property. Using the DCT matrix as the starting point, a better orthogonal transform matrix is derived. In addition, by exploring inter-frame dependency and local motion activity, transmission of substantial side information is avoided. The experiment results confirm that, with small computation overhead, the ROT is adaptive to change of local spatial characteristic of MC-residual frame and provides higher compression efficiency for the MC-residual than DCT, especially for high- and complex-motion videos.

In H.264/advanced video coding (AVC), lossless coding and lossy coding share the same entropy coding module. However, the entropy coders in the H.264/AVC standard were original designed for lossy video coding and do not yield adequate performance for lossless video coding. In this paper, we analyze the problem with the current lossless coding scheme and propose a mode-dependent template (MD-template) based method for intra lossless coding. By exploring the statistical redundancy of the prediction residual in the H.264/AVC intra prediction modes, more zero coefficients are generated. By designing a new scan order for each MD-template, the scanned coefficients sequence fits the H.264/AVC entropy coders better. A fast implementation algorithm is also designed. With little computation increase, experimental results confirm that the proposed fast algorithm achieves about 7.2% bit saving compared with the current H.264/AVC fidelity range extensions high profile.

This paper addresses the problem of delivering data packets for highly dynamic mobile ad hoc networks in a reliable and timely manner. Most existing ad hoc routing protocols are susceptible to node mobility, especially for large-scale networks. Driven by this issue, we propose an efficient Position-based Opportunistic Routing (POR) protocol which takes advantage of the stateless property of geographic routing and the broadcast nature of wireless medium. When a data packet is sent out, some of the neighbor nodes that have overheard the transmission will serve as forwarding candidates, and take turn to forward the packet if it is not relayed by the specific best forwarder within a certain period of time. By utilizing such in-the-air backup, communication is maintained without being interrupted. The additional latency incurred by local route recovery is greatly reduced and the duplicate relaying caused by packet reroute is also decreased. In the case of communication hole, a Virtual Destination-based Void Handling (VDVH) scheme is further proposed to work together with POR. Both theoretical analysis and simulation results show that POR achieves excellent performance even under high node mobility with acceptable overhead and the new void handling scheme also works well.

Multiple description coding (MDC) is one of the widely used mechanisms to combat packet-loss in non-feedback systems. However, the number of descriptions in the existing MDC schemes is very small (typically 2). With the number of descriptions increasing, the coding complexity increases drastically and many decoders would be required. In this paper, the compressive sensing (CS) principles are studied and an alternative coding paradigm with a number of descriptions is proposed based upon CS for high packet loss transmission. Two-dimentional discrete wavelet transform (DWT) is applied for sparse representation. Unlike the typical wavelet coders (e.g., JPEG 2000), DWT coefficients here are not directly encoded, but re-sampled towards equal importance of information instead. At the decoder side, by fully exploiting the intra-scale and inter-scale correlation of multiscale DWT, two different CS recovery algorithms are developed for the low-frequency subband and high-frequency subbands, respectively. The recovery quality only depends on the number of received CS measurements (not on which of the measurements that are received). Experimental results show that the proposed CS-based codec is much more robust against lossy channels, while achieving higher rate-distortion (R-D) performance compared with conventional wavelet-based MDC methods and relevant existing CS-based coding schemes.

With the wide applications of saliency information in visual signal processing, many saliency detection methods have been proposed. However, some key characteristics of the human visual system (HVS) are still neglected in building these saliency detection models. In this paper, we propose a new saliency detection model based on the human visual sensitivity and the amplitude spectrum of quaternion Fourier transform (QFT). We use the amplitude spectrum of QFT to represent the color, intensity, and orientation distributions for image patches. The saliency value for each image patch is calculated by not only the differences between the QFT amplitude spectrum of this patch and other patches in the whole image, but also the visual impacts for these differences determined by the human visual sensitivity. The experiment results show that the proposed saliency detection model outperforms the state-of-the-art detection models. In addition, we apply our proposed model in the application of image retargeting and achieve better performance over the conventional algorithms.

Dynamic programming (DP) is a popular and efficient technique in many scientific applications such as computational biology. Nevertheless, its performance is limited due to the burgeoning volume of scientific data, and parallelism is necessary and crucial to keep the computation time at acceptable levels. The intrinsically strong data dependency of dynamic programming makes it difficult and error-prone for the programmer to write a correct and efficient parallel program. Therefore, this paper builds a runtime system named EasyPDP aiming at parallelizing dynamic programming algorithms on multicore and multiprocessor platforms. Under the concept of software reusability and complexity reduction of parallel programming, a DAG Data Driven Model is proposed, which supports those applications with a strong data interdependence relationship. Based on the model, EasyPDP runtime system is designed and implemented. It automatically handles thread creation, dynamic data task allocation and scheduling, data partitioning, and fault tolerance. Five frequently used DAG patterns from biological dynamic programming algorithms have been put into the DAG pattern library of EasyPDP, so that the programmer can choose to use any of them according to his/her specific application. Besides, an ideal computing distribution model is proposed to discuss the optimal values for the performance tuning arguments of EasyPDP. We evaluate the performance potential and fault tolerance feature of EasyPDP in multicore system. We also compare EasyPDP with other methods such as Block-Cycle Wavefront (BCW). The experimental results illustrate that EasyPDP system is fine and provides an efficient infrastructure for dynamic programming algorithms.

In cloud computing, cloud providers can offer cloud consumers two provisioning plans for computing resources, namely reservation and on-demand plans. In general, cost of utilizing computing resources provisioned by reservation plan is cheaper than that provisioned by on-demand plan, since cloud consumer has to pay to provider in advance. With the reservation plan, the consumer can reduce the total resource provisioning cost. However, the best advance reservation of resources is difficult to be achieved due to uncertainty of consumer's future demand and providers' resource prices. To address this problem, an optimal cloud resource provisioning (OCRP) algorithm is proposed by formulating a stochastic programming model. The OCRP algorithm can provision computing resources for being used in multiple provisioning stages as well as a long-term plan, e.g., four stages in a quarter plan and twelve stages in a yearly plan. The demand and price uncertainty is considered in OCRP. In this paper, different approaches to obtain the solution of the OCRP algorithm are considered including deterministic equivalent formulation, sample-average approximation, and Benders decomposition. Numerical studies are extensively performed in which the results clearly show that with the OCRP algorithm, cloud consumer can successfully minimize total cost of resource provisioning in cloud computing environments.

Platform-as-a-Service (PaaS) clouds allow faster and more effective application development than traditional non-PaaS ways. One issue in designing PaaSs is how to make the development process deliver applications resilient to potential changes of the constraints. This is because any successful applications today must be as resilient as possible to dynamic external or internal constraining factors. Along this line, the first type of dynamic constraints we need to consider is the compatibility between possible components of the application. PaaSs must only engage compatible components to collaborate with each other in the same instance of applications. Other constraints include the environment that the application is running as well as the preferences of the users (or devices) that interact with the application. We present a data-flow based approach, for PaaS clouds, to designing cloud-based applications that are resilient to failures due to dynamic constraints on resources and on component compatibility. The uniqueness of our approach is the following: The procedure of building cloud-based applications is time-stamped. In this way, the composition of the application is updated anytime in accordance to the constraints in order to maximize the resilience of the application at that time. We have designed a graph structure called Instance Dependency Graphs (IDGs), and have used time-based IDGs to capture, analysis and optimize the resilience of the application. We present a case study to validate our approach.

In a wireless domain where a mobile user accesses heterogeneous wireless technologies with multiple interfaces, a multi-path scheduling algorithm can benefit mobile users' experience by aggregating different network bandwidth together. However, existing literature actually shows that for TCP flows, it may not be the case. To better exploit multi-path scheduling for TCP connections, this paper presents a multi-path scheduling algorithm named Adaptive Load Balancing Algorithm (ALBAM) to split traffic across different network access for these multi-homed users. Unlike other multi-path scheduling algorithms, ALBAM takes full advantage of the infrastructure of wireless domain and it has the following advantages: (1) it does not involve any upgrade of protocol stack on user devices. (2) ALBAM does not introduce any protocol signaling cost into the bandwidth-constrained wireless networks. (3) ALBAM reduces the number of out-of-order packets. Thus, ALBAM improves the throughput of the TCP connections. To evaluate the performance of ALBAM, we conduct comparative simulations of ALBAM against a related technique, i.e. Opportunistic Multipath Scheduling. The results show that ALBAM can achieve good bandwidth aggregation and provide better performance to TCP connections in the wireless domain.

Cloud computing service providers offer cost-effective means to burst computational needs and utilise live migration of virtual machines (VMs) for effective and efficient work-load movements with short service downtimes. However, there is a lack of support for migrating VMs between different service providers as well as private and public cloud offerings, main challenges arise from the bandwidth and storage costs of data during migration potentially mitigating any cost benefit. In this paper, we propose and evaluate techniques for efficient and effective transfer and storage of VM images, which have high duplication, for both instance and volume-based cloud storage. Our main focus is on both the public and private cloud infrastructure and the movement of VMs between them.

Provenance, a meta-data describing the derivation history of data, is crucial for the uptake of cloud computing to enhance reliability, credibility, accountability, transparency, and confidentiality of digital objects in a cloud. In this paper, we survey current mechanisms that support provenance for cloud computing, we classify provenance according to its granularities encapsulating the various sets of provenance data for different use cases, and we summarize the challenges and requirements for collecting provenance in a cloud, based on which we show the gap between current approaches to requirements. Additionally, we propose our approach, Data PROVE, that aims to effectively and efficiently satisfy those challenges and requirements in cloud provenance, and to provide a provenance supplemented cloud for better integrity and safety of customers' data.

Selective attention in the human visual system is performed as the way that humans focus on the most important parts when observing a visual scene. Many bottom-up computational models of visual attention have been devised to get the saliency map for an image, which are data-driven or task-independent. However, studies show that the task-driven or top-down mechanism also plays an important role during the formation of visual attention, especially with the cases of object detection and location. In this paper, we proposed a new computational visual attention model by combining bottom-up and top-down mechanisms for man-made object detection in scenes. This model shows that the statistical characteristics of orientation features can be used as top-down clues to help for determining the location for salient objects in natural scenes. Experiments confirm the effectiveness of this visual attention model.

This paper addresses the downlink scheduling issue in drive-thru networks which is characterized by flow-level dynamics and user basis deadlines. Vehicular users requesting for data download service with variable file sizes regularly arrive at and depart from the limited coverage range of roadside access point. If the corresponding data queue at the access point cannot be serviced in a certain time period, it has to be cleared, resulting in degraded QoS. To minimize the number of uncompleted file download jobs in the face of multiple-user contention, a Dynamic Predictive Scheduling (DPS) algorithm is proposed. Based on the prediction of the remaining bandwidth of the different users, a scheduling tree is constructed to facilitate the selection of the data queue to serve at particular time slots. Through extensive simulation, it is shown that DPS consistently outperforms competitive scheduling schemes under varying workloads.

E-learning systems are now considered a core IT support in most Institute of Higher Learning. Each student is assigned a set of courses based on their preferences for each semester. Learning materials are posted and evaluated through the system. Mining access and usage log data of the e-learning system can give insights such as on how these materials are accessed by students in their order of preferences on each of the courses. A study was carried out, and the results show that there is good correlation between students’ performance and e-learning usage. In this paper, we describe our methodology on analyzing students’ usage of e-learning systems for course management with the help of cloud computing services. We present a student performance predictive model built from the retrieved logs that is confirmed to achieve sufficiently good accuracy.

H.264/AVC video coding standard with hierarchical bipredictive picture (HBP) generally outperforms the other prediction structures such as ‘IPPP…’ and ‘IBBP…’ through better exploitation of data correlation using the preceding and succeeding reference frames. However, due to the different coding order of frames, the HBP scheme could not fully exploit the data correlations using multiple reference frames for occluded background, repetitive motion, etc. In this paper, we propose a new HBP scheme which uses the most common reference frame in scene (McFIS) as a third reference frame with other two closest bipredictive reference frames assuming that foreground and background areas of the current frame are referenced from the two bipredicted frames and the McFIS respectively. The experimental results confirm that the proposed scheme outperforms two state-of-art algorithms by improving significant image quality with comparable computational time.

Amazon Elastic Compute Cloud (EC2) provides a cloud computing service by renting out computational resources to customers (i.e., cloud users). The customers can dynamically provision virtual servers (i.e., computing instances) in EC2, and then the customers are charged by Amazon on a pay-per-use basis. EC2 offers three options to provision virtual servers, i.e., on-demand, reservation, and spot options. Each option has different price and yields different benefit to the customers. Spot price (i.e., price of spot option) could be the cheapest, however, the spot price is fluctuated and even more expensive than the prices of on-demand and reservation options due to supply-and-demand of available resources in EC2. Although the reservation and on-demand options have stable prices, their costs are mostly more expensive than that of spot option. The challenge is how the customers efficiently purchase the provisioning options under uncertainty of price and demand. To address this issue, two virtual server provisioning algorithms are proposed to minimize the provisioning cost for long- and short-term planning. Stochastic programming, robust optimization, and sample-average approximation are applied to obtain the optimal solutions of the algorithms. To evaluate the performance of the algorithms, numerical studies are extensively performed. The results show that the algorithms can significantly reduce the total provisioning cost incurred to customers.

The key barrier to widespread uptake of cloud computing is the lack of trust in clouds by potential customers. While preventive controls for security and privacy are actively researched, there is still little focus on detective controls related to cloud accountability and audit ability. The complexity resulting from large-scale virtualization and data distribution carried out in current clouds has revealed an urgent research agenda for cloud accountability, as has the shift in focus of customer concerns from servers to data. This paper discusses key issues and challenges in achieving a trusted cloud through the use of detective controls, and presents the Trust Cloud framework, which addresses accountability in cloud computing via technical and policy-based approaches.

Trust is one of the main obstacles to widespread Cloud adoption. In order to increase trust in Cloud computing, we need to increase transparency and accountability of data in the Cloud for both enterprises and end-users. However, current system tools are unable to log file accesses and transfers effectively within a Cloud environment. In this paper, we present Flogger, a novel file-centric logger suitable for both private and public Cloud environments. Flogger records file- centric access and transfer information from within the kernel spaces of both virtual machines (VMs) and physical machines (PMs) in the Cloud, thus giving full transparency of the entire data landscape in the Cloud. With Flogger, services can be built above it to provide Cloud providers, end-users and regulators with the relevant provenance, e.g. a tool for an end- user to track whether his/ her file was 'touched' by an unauthorized user. We present the initial developments of Flogger, and interesting results from our experiments. We also present compelling future work that will shape the beginnings of a new logging paradigm: distributed VM/ PM file-centric logging.

We look at the problem of network wide broadcast using the multi-rate feature of a wireless ad hoc network. Existing research has primarily focused on achieving minimum latency by construction of minimum weight connected dominating sets (WCDS) based on neighbourhood information. In this paper, we are interested in stateless multi-rate broadcasting algorithms in which nodes determine their broadcasting behaviour based on neighbourhood transmissions. The primary contribution of this paper is that we show how broadcast effectiveness at different data rates are related and how this relationship can be used to optimize algorithm design. We propose three stateless broadcasting algorithms and demonstrate the performance improvements achievable. Our simulation results show that significant benefits can be obtained in terms of minimizing both the number of forwarding nodes as well as the broadcast latency.

Existing literature has looked to exploit wireless broadcast advantage (WBA) in order to optimize the performance of a wide variety of network operations. In this paper, we obtain a measure of WBA in a multihop scenario. We consider that all nodes in the network store and propagate implicitly received information from neighbourhood transmissions, resulting in the creation of a distributed cache, which we term broadcast cache. We obtain a lower bound on the growth of the broadcast cache in terms of the fewest set of transmissions in the network, which we define as the minimum set of non-altruistic transmissions. Subsequently, we use our results to obtain feasibility conditions that determine whether WBA can be effectively utilized depending on flow requirements.

In video coding, an intra (I)-frame is used as an anchor frame for referencing the subsequence frames, as well as error propagation prevention, indexing, and so on. To get better rate-distortion performance, a frame should have the following quality to be an ideal I-frame: the best similarity with the frames in a group of picture (GOP), so that when it is used as a reference frame for a frame in the GOP we need the least bits to achieve the desired image quality, minimize the temporal fluctuation of quality, and also maintain a more consistent bit count per frame. In this paper we use a most common frame of a scene in a video sequence with dynamic background modeling and then encode it to replace the conventional I-frame. The extensive experimental results confirm the superiority of our proposed scheme in comparison with the existing state-of-art methods by significant image quality improvement and computational time reduction.

The uniqueness of security vulnerabilities in ad hoc networks has given rise to the need for designing novel intrusion detection algorithms, different from those present in conventional networks. In this work, we propose an autonomous host-based intrusion detection system for detecting malicious sinking behavior. The proposed detection system maximizes the detection accuracy by using cross-layer features to define a routing behavior. For learning and adaptation to new attack scenarios and network environments, two machine learning techniques are utilized. Support Vector Machines (SVMs) and Fisher Discriminant Analysis (FDA) are used together to exploit the better accuracy of SVM and faster speed of FDA. Instead of using all cross-layer features, features from MAC layer are associated/correlated with features from other layers, thereby reducing the feature set without reducing the information content. Various experiments are conducted with varying network conditions and malicious node behavior. The effects of factors such as mobility, traffic density, and the packet drop ratios of the malicious nodes are analyzed. Experiments based on simulation show that the proposed cross-layer approach aided by a combination of SVM and FDA performs significantly better than other existing approaches.

The H.264 video coding standard exhibits higher performance compared to the other existing standards such as H.263, MPEG-X. This improved performance is achieved mainly due to the multiple-mode motion estimation and compensation. Recent research tried to reduce the computational time using the predictive motion estimation, early zero motion vector detection, fast motion estimation, and fast mode decision, etc. These approaches reduce the computational time substantially, at the expense of degrading image quality and/or increase bitrates to a certain extent. In this paper, we use phase correlation to capture the motion information between the current and reference blocks and then devise an algorithm for direct motion estimation mode prediction, without excessive motion estimation. A bigger amount of computational time is reduced by the direct mode decision and exploitation of available motion vector information from phase correlation. The experimental results show that the proposed scheme outperforms the existing relevant fast algorithms, in terms of both operating efficiency and video coding quality. To be more specific, 82 ~92% of encoding time is saved compared to the exhaustive mode selection (against 58 ~74% in the relevant state-of-the-art), and this is achieved without jeopardizing image quality (in fact, there is some improvement over the exhaustive mode selection at mid to high bit rates) and for a wide range of videos and bitrates (another advantages over the relevant state-of-the-art).

Cooperative communication techniques have earlier been applied to design of the IEEE 802.11 medium access control (MAC) and shown to perform better. High rate stations can help relay packets from low-rate stations resulting in better throughput for the entire network. However, this also involves additional energy costs on the part of the relay which can result in reducing the network lifetime. We propose a cooperative MAC protocol NetCoop with the objective of maximizing the network lifetime and achieving high throughput. Based on this design, we also propose a flexible strategy which allows cooperation to be achieved using more than one relay. We show that this can achieve at least as good throughput as that of single relay cooperation while maintaining a high network lifetime.

Existing ad-hoc network routing strategies base their operations on flooding route requests throughout the network and choosing the shortest path thereafter. However, this typically results in a large number of unnecessary transmissions, which could be expensive for resource-constrained nodes such as those in a sensor network. In this paper, we propose a new mechanism HopAlert which optimizes route establishment and packet routing by limiting the number of nodes taking part in the route discovery process while achieving a low number of hops establishment. Using analysis and simulations, we show that this results in more routes with shorter hop counts than a reactive flooding protocol such as AODV while achieving higher savings.

In recent years, a lot of terminal mobility support schemes have been proposed. These schemes mainly focus on providing seamless mobility by reducing the handover delay. They work well when the handover takes place between two overlapping wireless networks. However, in some worst scenario, it is possible that a mobile node (MN) becomes disconnected from the network for some time before it can join another wireless network. In such case, existing mobility protocols will suffer great data loss and the data connection could break. The inability to preserve data connection across long handover delay or connectivity disruption is an unavoidable obstacle before a convergent terminal mobility can be achieved. In this paper, we propose a disruption tolerant mobility architecture (DTMA) to address this issue. In DTMA, terminal mobility is supported via the use of local proxy server. User applications communicate with other hosts via the local proxy on the mobile node itself. To enable disruption tolerance, DTMA adapts delay tolerant network (DTN) architecture as a transport protocol. When the node is disconnected from the network, data sent to it will be cached by the network and delivered to the node later when it reconnects to the network. The traffic overhead of DTMA is analyzed and it is shown that DTMA provides disruption tolerance with acceptable traffic overhead.

As a local mobility management protocol, Proxy Mobile IPv6 (PMIPv6) is designed to enable mobile nodes to move in a domain without any involvement in the mobility signalling operation. However, a single domain cannot satisfy mobile nodes' movement requirement. Hence, we propose a solution to enable mobile nodes to move across multiple PMIPv6 domains. We introduce a new network entity called traffic distributor (TD) to manage the inter-domain handover. Through redirecting the traffic to the domain which currently visited by the mobile node, TD can ensure mobile node continue its session even when it moves across domains. We conduct a series of experiments to compare our proposal with Neumann proposal which is another proposal to handle inter-PMIPv6-domain issues. Experimental results show that our proposal is a viable alternative for inter-domain handover, it can outperform Neumann proposal in terms of binding cache entry numbers,transmission delay and handover delay.

Motion estimation (ME) and motion compensation (MC) using variable block size, fractional search, and multiple reference frames (MRFs) help the recent video coding standard H.264 to improve the coding performance significantly over the other contemporary coding standards. The concept of MRF achieves better coding performance in the cases of repetitive motion, uncovered background, non-integer pixel displacement, lighting change, etc. The requirement of index codes of the reference frames, computational time in ME&MC, and memory buffer for pre-coded frames limits the number of reference frames used in practical applications. In typical video sequence, the previous frame is used as a reference frame with 68~92% of cases. In this paper, we propose a new video coding method using a reference frame (i.e., the most common frame in scene (McFIS)) generated by the Gaussian mixture based dynamic background modelling. The McFIS is not only more effective in terms of rate-distortion and computational time performance compared to the MRFs but also error resilient transmission channel. The experimental results show that the proposed coding scheme outperforms the H.264 standard video coding with five reference frames by at least 0.5 dB and reduced 60% of computation time.

In this paper, we describe our goal of an effective course management system for assisting course managers to make informed decisions about what materials should be most appropriate to be presented to students (learners) and what learning strategies or methods should be used for the students. The system is supported by our design of a novel framework for user-driven data analytics in the cloud. Different modules of the framework will be illustrated in detail in the context of course management.

H.264/AVC FRExt (Fidelity Range Extensions) and Motion JPEG2000 are the latest inter-frame and intra-frame video coding standards, respectively. It is well known that an inter-frame method achieves higher coding efficiency compared with an intra-frame one, and the Motion JPEG2000 has been selected for digital cinema compression. In this paper, we attempt to compare these two different schemes with theoretical and experimental analysis for super-HD (high definition) visual signals. One additional contribution of the paper is that the impact of block partition, motion search range and skipped block size for inter-frame coding is discussed. Based on the analysis, we extend the standard H.264/AVC FRExt by using larger block size and search range. The experimental results show that this extension leads to higher coding efficiency and makes the H.264/AVC FRExt more suitable for super-HD video coding.

In this paper, we propose a low-complexity video codec based on two-dimensional Singular Value Decomposition (2D-SVD). We exploit the common temporal characteristics of video without resorting to motion estimation. It has been demonstrated that this codec has higher coding efficiency than the relevant existing low complexity codecs. Moreover, the proposed codec performs well to deal with packet loss that is unavoidable in error-prone transmission. Therefore it is with advantages and good potential for wireless video applications such as mobile video calls and wireless surveillance.

1The pattern-based video coding (PVC) outperforms the H.264 through better exploitation of block partitioning and partial block skipping. In the PVC scheme the best pattern is determined against the moving regions (MRs) in a macroblock (MB) of the current frame against the co-located MB in the reference frame; motion estimation (ME) and motion compensation (MC) are carried out using the pattern covered MRs, and the rest of the regions are treated as skipped areas. The MRs can be due to the object areas and the uncovered background (UCB) areas. Thus, the ME & MC by the pattern for the MRs of the UCB would not be accurate if there is no similar region in the reference frame. As a result no coding gain can be achieved for the UCB. Recently a dynamic background frame termed as the McFIS (the most common frame of a scene) has been generated using Gaussian mixture models for object detection. In this paper we propose a new PVC technique which will use the McFIS as a reference frame to determine the MRs where only object areas will be captured as the MRs. Thus, the proposed technique overcomes the mismatch problem of the UCB for ME&MC. The experimental results confirm the superiority of the proposed scheme in comparison with the existing PVC and McFIS-based methods by achieving significant image quality gain.

The existing image compression methods (e.g., JPEG2000, etc.) are vulnerable to bit-loss, and this is usually tackled by channel coding that follows. However, source coding and channel coding have conflicting requirement. In this paper, we address the problem with an alternative paradigm, and a novel compressive sensing (CS) based compression scheme is therefore proposed. Discrete wavelet transform (DWT) is applied for sparse representation, and based on the property of 2-D DWT, a fast CS measurements taking method is presented. Unlike the unequally important discrete wavelet coefficients, the resultant CS measurements carry nearly the same amount of information and have minimal effects for bit-loss. At the decoder side, one can simply reconstruct the image via l1minimization. Experimental results show that the proposed CS-based image codec without resorting to error protection is more robust compared with existing CS technique and relevant joint source channel coding (JSCC) schemes.

The conventional Intra (I-) frame is used for error propagation prevention, backward/forward play, random access, indexing, etc. This frame is also used as an anchor frame for referencing the subsequence frames. To get better rate-distortion performance a frame should have the following quality to be an ideal I-frame: the best similarity with the frames in a GOP, so that (i) when it is used as a reference frame for a frame in the GOP we need less bits to achieve the desired image quality; (ii) if any frame is missing at the decoding end we can retrieve the missing frame from it. In this paper we will generate a most common frame of a scene (McFIS) in a video sequence using dynamic background modelling and then encode it to replace the conventional I-frame. By using McFIS as an I-frame, we not only gain the above mentioned two benefits but also ensure adaptive GOP for better rate-distortion performance compared to the existing coding schemes. The experimental results confirm the superiority of our proposed scheme in comparison with the existing state-of-art methods by significant image quality and computation time.

Domain Name Systems (DNS) servers provide a critical service to users and application on the internet. At the top of the DNS hierarchy is the Root DNS servers providing pointers to Top Level Domain servers on the Internet. Over the past years the number of Root DNS servers has grown to cope with the exponential growth of the Internet. This paper reports on the status and performance of the Root DNS servers gathered by the Gulliver's project SEIL probes which are deployed globally. This is the first effort to carry out long term study, from 2007 till 2009, on the performance of Root DNS servers. The data gathered shows the performance of anycast Root DNS servers in terms of Loss Rate and Query Response Time (QRT). Anycast servers' performance on the whole has been improving with more deployment of its instances, while unicast servers' performance has been roughly the same. We have correlated changes in the QRT to events that has occurred, eg. Cable breaks and deployment of a new anycast instance.

Traditional MANET routing protocols are quite susceptible to nodes' mobility, especially for large-scale networks in which the end-to-end path length is usually large. In order to improve the routing performance in the face of fast changing network topology, we propose a novel Robust Geographic Routing (RGR) protocol which takes advantage of the broadcast nature of wireless medium by employing opportunistic routing like forwarding strategy. At the same time, a Virtual Destination based Void Handling (VDVH) scheme is also proposed to work together with RGR. Simulation results show that RGR achieves excellent performance even under high node mobility and the new void handling scheme also works well and further enhances the performance of RGR.

Cloud computing provides users an efficient way to dynamically allocate computing resources to meet demands. Cloud providers can offer users two payment plans, i.e., reservation and on-demand plans for resource provisioning. Price of resources in reservation plan is generally cheaper than that in on-demand plan. However, since the reservation plan has to be acquired in advance, it may not fully meet future demands in which the on-demand plan can be used to guarantee the availability to the user. In this paper, we propose an optimal virtual machine placement (OVMP) algorithm. This algorithm can minimize the cost spending in each plan for hosting virtual machines in a multiple cloud provider environment under future demand and price uncertainty. OVMP algorithm makes a decision based on the optimal solution of stochastic integer programming (SIP) to rent resources from cloud providers. The performance of OVMP algorithm is evaluated by numerical studies and simulation. The results clearly show that the proposed OVMP algorithm can minimize users' budgets. This algorithm can be applied to provision resources in emerging cloud computing environments.

Cloud computing has been emerged as the flexible, efficient, and economical distributed computing platform to meet the dynamic and random demand from the users. In this paper, we consider cloud computing environment with resource market between private clouds (i.e., buyers) and service providers (i.e., sellers) in public cloud. Economic analysis is proposed for different types of resource markets, i.e., monopoly (single service provider), competitive and cooperative oligopolies (few service providers). We study the optimal strategy for service provider in monopoly market, the Nash equilibria in competitive oligopoly market, and bargaining solution in cooperative oligopoly market. In addition, the decision and condition for service providers to to establish collusion in the oligopoly market are also investigated.

Green computing is a new paradigm of designing the computer system which considers not only the processing performance but also the energy efficiency. Power management is one of the approaches in green computing to reduce the power consumption in distributed computing system. In this paper, we first propose an optimal power management (OPM) used by a batch scheduler in a server farm. This OPM observes the state of a server farm and makes the decision to switch the operation mode (i.e., active or sleep) of the server to minimize the power consumption while the performance requirements are met. An optimization problem based on constrained Markov decision process (CMDP) is formulated and solved to obtain an optimal decision of OPM. Given that OPM is used in the server farm, then an assignment of users to the server farms by a job broker is considered. This assignment is to ensure that the cost due to power consumption and network transportation is minimized. The performance of the system is extensively evaluated. The result shows that with OPM the job waiting time can be maintained below the maximum threshold while the power consumption is much smaller than that without OPM.

Traditional MANET routing protocols are quite susceptible to link failure as well as vulnerable to malicious node attack. In this paper, we propose a novel protocol called Position based Opportunistic Routing (POR) which takes full advantage of the broadcast nature of wireless channel and opportunistic forwarding. The data packets are transmitted as a way of multicast (which is actually implemented by MAC interception) with multiple forwarders. A forwarder list determined by previous hop according to local position information is inserted into the IP header and the candidates take turn to forward the packet based on a predefined orders. This redundancy and randomness make it quite efficient and robust. In addition, inherited from position based routing, POR's control overhead is almost negligible which justifies its good scalability. Both theoretical analysis and simulation results show that POR not only achieves outstanding performances in normal situations but also yields excellent resilience in hostile environments.

Significant throughput degradation of multihop path communication in wireless mesh network is one of the major problems in wireless communication. The main reason for the lack of bandwidth is channel interference, which is caused by contention for the shared channel between wireless nodes. The natural approach to overcome this problem is exploiting the availability of multiple channels multiple interfaces (MCMI) networks. However, it is costly and may not be practical to dedicate one interface per channel for every node. Thus in this paper, we study the MCMI network, where the number of interfaces that every node has is less than the number of available channels. Simple and distributed channel scheduling algorithms for communication in multiple channels multiple interfaces networks are discussed. The objective of the proposed algorithms is to minimize the channel interference that causes the throughput degradation in multihop networks. The proposed algorithms are evaluated with extensive simulations. The simulation results show that the proposed algorithms well exploited the availability of multiple channels multiple interfaces to overcome the throughput degradation problem.

In disruption tolerant networks, aggressive packet forwarding scheme like flooding has a major drawback in terms of network congestion. In this paper, we proposed a new routing algorithm, called adaptive randomized epidemic routing (ARER). ARER dynamically adjusts the forwarding probability for each message according to a new metric, replications density. Meanwhile, ARER arranges the forwarding sequence and the dropping priority based on their assigned weight. The weight is determined by the replication density, the delivery predictability, and TTL. An extensive simulation of ARER using various scenarios was carried out and its performance was compared to well known DTN routing protocols: epidemic routing, randomized routing and spray-and-wait routing. Our results show that ARER outperforms them in all scenarios in terms of packet delay and delivery.

To improve the local mobility performance of a wireless network, we present a new framework, namely Efficient Framework for Local Mobility (EFLoM). In EFLoM, we introduce three entities: Local Anchor Router (LAR), Wireless Access Gateway (WAG) and Mobile Node (MN). LAR is a router that is in charge of MN's IP mobility management. It keeps track of MN's movement, and delivers the MN's packets to its current location. A WAG is an access router that is responsible for managing the traffic flows within the same local mobility domain. In EFLoM, when Corresponding Node (CN) and MN connect to the same local mobility domain, packets are sent directly between CN and MN by WAG, without suffering sub-optimal routing problems. A MN is a mobile host which updates its data packets' header with its current location address before packets are sent out. To evaluate our new framework, we compare EFLoM with Hierarchical Mobile IPv6 Mobility Management (HMIPv6) which is an existing protocol that has good performance for local mobility. Both analytical analysis and simulation result show that EFLoM can attain substantial improvements over HMIPv6 in terms of handover delay, end to end delay, traffic overhead, which therefore presents EFLoM as a new engineering alternative to existing MIPv6 based techniques.

Communications in single channel single interface ad hoc network suffer from channel access contention, which results in bandwidth scarcity. One way to overcome this problem is through the use of multiple channels multiple interfaces (MCMI). In this paper, we present a dual-interface multiple channels destination-sequenced distance-vector (DSDV-M) routing protocol, which is the extension of the DSDV routing protocol to MCMI version. The objective of the proposed algorithm is to reduce the channel interference so that multiple transmissions can occur concurrently, thus increasing network capacity. The proposed algorithm is evaluated against channel scheduling algorithms; the simulation results show that the proposed algorithm is able to exploit the availability of multiple channels multiple interfaces to improve network capacity. In addition, we are also able to gain greater than twice the throughput of DSDV with single channel single interface.

Abstract A virtual organization is established when physical organizations collaborate to share their computing resources with the aim of serving each other when there is a likelihood of insufficient local resources during peak resource usage periods at any organization. Contention becomes a potential problem when a large number of requests, which can overwhelm the aggregate capacity of shared resources, are submitted from the participating organizations coincidentally at the same period. In particular, when a small number of requests that require large amounts of computing resources are admitted in place of a large number of requests that require less computing resources, the overall system performance, in terms of admission ratio, can deteriorate significantly. Hence, admission control is necessary to reduce resource oversubscription. Because domain‐shared computing resources are likely to be combined to form a large‐scale system, it is not possible to define a fixed admission policy solely based on the request's CPU and execution time requirements. In this paper, we introduce an admission control framework, based on a pricing model, for a multi‐domain‐shared computing infrastructure. The performance of the admission control framework is evaluated under different scenarios that contribute to the overall degree of competition for shared resources. The results are presented and analyzed in this paper. Copyright © 2008 John Wiley & Sons, Ltd.

Mobile IP enables IP mobility support for mobile node (MN), but it suffers from triangular routing, packet redirecting, increase in IP header size, and the need for new infrastructure support. This paper details an alternative to enable terminal mobility support for MN. This scheme does not suffer from triangular routing effect and does not require dedicated infrastructure support such as home agent. It also does not increase the size of the IP header and does not require redirection of packets. These benefits are enabled with a tradeoff, which requires modifications on MN and its correspondent node. It uses an innovative IP-to-IP address mapping method to provide IP address transparency for applications and taps on the pervasiveness of SIP as a location service. From our analysis, we show that TMSP is much more efficient than mobile IP in terms of the number of hops as well as overhead. Our prototype implementation also shows that TMSP provides seamless communication for both TCP and UDP connections and the computational overhead for TMSP has minimal impact on packet transmission.

In this paper, we present a comparative performance evaluation of our proposed neighbor discovery scheme versus other basic neighbor discovery schemes based on omnidirectional broadcast. The scheme is able to discover nodes beyond the omnidirectional transmission range with the help of minimal number of neighbors. It does not require high-power broadcast and works in a distributed manner. Simulation results show that the proposed scheme is able to discover neighbors faster than other basic neighbor discovery schemes.

In recent years, wireless communication technologies have been pervasive. We have many wireless options to connect to the Internet: IEEE 802.11 WLANs, GPRS, UMTS and many other competing technologies. These wireless standards were designed independently of each other. Hence, these standards do not specify a common solution to perform connection handover between two different standards. Given that each of these standards has its own advantages and disadvantages, users may want to use the best available connection for their needs. In this paper, we propose a scheme that enables seamless transitions between different wireless standards through SIP protocol. With our scheme, IP mobility across both vertical and horizontal handovers is possible without any modification to the existing applications. We define a guideline for handovers between networks, and then perform experiments with our scheme implementation to confirm the feasibility of our proposed scheme. Our results show that with our scheme, connection disruption is minimized.

A critical problem for P2P streaming applications is to construct and maintain the overlay such that it continues to efficiently distribute data stream even in dynamic network environment. A common solution to this problem is to constantly adapt the overlay structure to the changing network conditions. However, in this paper, we propose an algorithm to schedule the sending order of queued data at each peer by taking into account both the data content as well as the overlay conditions. The data packets which are important to most users on the entire P2P system are sent out earlier. Simulation results show that the scheduling algorithm improves the overall streaming quality of the P2P system with little overhead added to the network traffic. Moreover, the improvement in overall streaming quality is also achieved regardless of the video streaming formats.

In this paper, we propose a mobility management scheme with quality-of-service(QoS) support for heterogeneous multihomed mobile node. With this scheme, the multihomed mobile node can move between networks seamlessly and at the same time, utilize all available links based on the QoS requirement. Session initiation protocol(SIP)[l] is used in session management and each multihomed node is identified by a unique SIP uniform resource identifier(URI). This mobility scheme supports both horizontal and vertical handover and no triangular routing is involved. Depending on the QoS requirement, each data flow can either be allocated to one of the links or distributed across multiple links. The ability to distribute data at packet level across multiple links is unique compared to many approaches using flow level distribution. The proposed scheme not only effectively increases the throughput and reduces the delay, but also alleviates out-of-sequence packet problem where certain applications are sensitive to.

The IEEE 802.11e standard has introduced specifications for service differentiation among different classes of data by specifying four service classes and a new contention resolution mechanism called EDCA. However, while the protocol shows a better performance for higher priority data such as voice and video, the performance is seen to drop drastically at high loads. In this paper, we explore the effectiveness of a multi-stage contention scheme for providing QoS differentiation among four different service classes, as specified by EDCA. From our analysis, we observe that the multi-stage with prioritization that we propose gives a much better performance than EDCA for higher priority data. Moreover, its good performance even at high network loads shows that this design is much more scalable.

In ad hoc networks, the vulnerability of nodes to routing attacks is a serious concern. In this work we propose a cross-layer based routing attack detection system for ad hoc networks. Previous work that uses mostly audit trails collected from the routing protocol suffers from inadequacy of features to construct a reliable model for detecting anomalous routing behavior. On the other hand, use of linear detectors lead to very high false positives and false negatives because of the inherent non-linear nature of the feature space. In this work, we address these issues by collating features from multiple protocols at different layers and using a non-linear detector based on support vector machine (SVM). The consequent problem of computational expense of the detection process is addressed by a combination of novel data reduction techniques. Simulation results show that the performance of the proposed CRADS is far superior than conventional protocol-specific detection systems.

The design and engineering of complex materials and products often requires intricate interactions between domain experts in science, material and engineering as well as the utilization of diverse software systems for discovery and optimization. If left as it is, design engineers would most likely be at a loss on how to engage the entire entourage of the multi- disciplinary processes as well as the compute-intensive and data-intensive nature of the activities involved. This paper describes a possible solution through the development of a Grid-based Problem Solving Environment for Engineering of Materials (GPEM). The GPEM aims to provide a one-stop platform where engineers will perform material discovery, design optimization and material characterization, with grid computing as the enabling technology. Upon describing the details of the process workflow and the adopted architecture design, the paper will present the current implementation of GPEM, in the design optimization of fractal structures.

Ad Hoc Qs is a MAC independent traffic differentiation algorithm in Ad Hoc WLANs which are based on IEEE 802.11. Ad Hoc Qs uses Host Qs which is another Mac independent traffic differentiation algorithm. Host Qs is able to differentiate different traffic priorities only in one station and by adding a Prioritizer to it (Ad hoc Qs) it can provide traffic differentiation in wireless media (between multiple wireless nodes). The proposed slotted-delay scheduler (Prioritizer) is not adaptive and can not provide persistent traffic differentiation when the number of active stations in the network varies. We first propose an analytical model for Host Qs and use it for choosing proper values for virtual contention windows for desired traffic ratios in each station. Then we propose an analytical model for Ad Hoc Qs. Using this model we can adaptively choose the proper slot size according to the number of active stations and priority of traffic they send in order to get desired traffic differentiation. The simulation results verify the accuracy of the proposed adaptive slot size.

This paper presents a semi-Markov model that evaluates the performance of TCP connections with asymmetric Additive Increase and Multiplicative Decrease (AIMD) congestion control settings involved in sharing of a common drop-tail router. We study the fairness of the connections and their individual bandwidth utilizations as well as packet loss rates. We confirm that certain asymmetric AIMD settings may achieve fairness in bandwidth sharing. We also found that while connections with asymmetric AIMD settings operate at different bandwidth utilizations, they generally experience similar packet loss rate.

An IP mobility support protocol that enables personal and terminal mobility for IP-based applications is put forward. This protocol does not require new network entities or support from network service providers. It comprises an innovative IP-to-IP address mapping module at the network layer and an user agent to interact with a directory service server and correspondent nodes. It does not require a permanent IP address and a home server. It does not use tunnelling on mobile nodes nor alter route path of IP packets. In this paper, we describe our implementation and present our experimental results. Experiments show that this protocol works for UDP and TCP connections without affecting the throughput of the mobile node on a wireless LAN. Related works are also discussed and quantitatively compared. As an example, this protocol provides seamless execution for applications like VoIP and video conferencing on mobile nodes that roam across wireless networks.

A hybrid protocol for application level multicast (HPAM) for live video streaming without native IP multicast support is proposed. HPAM exploits the simplicity and optimality of a lightweight, centralized server with the robustness and scaleability of distributed clients. HPAM self-organizes clients on the fly to form efficient source-based overlay trees while the server facilitates peer discovery and also serves as a reliable backup should the distributed algorithms fails. Tree construction, refinement and recovery from partitions are carried out independently by the clients. Simulation results show that HPAM can build and maintain reasonably latency-efficient overlay trees with a lower overhead than a fully centralized system (host based multicast) and yet more responsive to group dynamics and network environment than a fully distributed system (host multicast).

Peer-to-peer (P2P) structure is widely applied in multimedia streaming applications to support large number of clients spread all over the Internet. For real-time streaming applications, data arrival time is critical. In this paper, we design a scheduling scheme to manipulate the order of data transmission between peers in order to improve the transmission efficiency in P2P streaming. Simulation results show that our scheme improves average data arrival time and increases the rate of data arriving on time in dynamic fluctuating network.

In network mobility support for wireless ad hoc networks, mobile routers (MRs) are connected to access router (AR) by ad hoc routing. Mobile network nodes (MNNs) connect to an MR to communicate to distant correspondent nodes (CNs). Previous works on network mobility in mobile IPv6 involve packet redirection through a home agent; this redirection is costly in both bandwidth and delay. In this paper, we propose a terminal-assisted network mobility management scheme, which requires neither permanent node's IP address nor additional network servers/infrastructure support. With this scheme, IP packets redirection is not necessary; IP packets are routed directly between MNNs and CNs. Our scheme is implemented in three components: an IP-to-IP address mapping scheme on the MNN, an IPv6 header extension on each IP packet, and an IP address redirection scheme on the MRs. An MNN is located by a uniform resource identifier (URI) provided by a directory service that maps URI to IP address. Through numerical analysis, we show that our mobility management scheme exhibits better efficiency compared with that of mobile IPv6.

Latency-rate (LR) schedulers have shown their ability in providing fair and weighted sharing of bandwidth with an upper bound on delivery latency of packets while earliest departure first (EDF) schedulers have shown their ability in providing LR-decoupled service whereby the delivery latency of packets is not bounded by the reserved rate. However, EDF schedulers require traffic shapers to ensure flow protection. We propose quantum-based earliest deadline first scheduling (QEDF), a quantum-based scheduler that provides flow protection, throughput guarantee and delay bound guarantee for flows that require LR-coupled and LR-decoupled types of reservations. It classifies flows into time-critical (TC), jitter-sensitive (JS), and rate-based (RB) classes and uses a quality-of-service forwarding rule to determine the next packet to be serviced by the scheduler. It provides nonpreemptive priority service to TC queues. This allows LR-decoupled reservation for flows that have a low rate and intolerable delay. Packets from JS queues can be delayed by other packets if forwarding the latter will not result in the former missing its deadline. As a quantum-based scheduler, the QEDF scheduler provides throughput guarantees for RB queues. We present both analytical and simulation results of QEDF, whereby we evaluated QEDF in its deployment as a single-class as well as a multiservice scheduler

We investigate the causal order of events in distributed virtual environments (DVEs). We first define the critical causal order relation among the events. Then, we propose some mechanisms to enhance the prevalent RO (receive order delivery) mechanism in DVEs so that the real-time property of DVEs is preserved while the critical causal order violations are reduced. These mechanisms are implemented as a middleware. Experimental results show that the middleware performs well in reducing the critical causality violations in simulation and incurs little processing overhead.

Integrated network management frameworks for self-managing systems in a grid environment consisting of disparate applications, devices and subsystems require the use of a common definition of these managed resources. The common information model (CIM) provides such a standard for their description. However, the CIM specification lacks formalism which limits its use in knowledge aggregation and reasoning. This paper discusses the design of a repository adapter for resource information modeled in CIM. The adapter translates CIM constructs to an ontology-based language, the Data Centre Markup Language (DCML), thereby formalizing the model. Issues encountered during this process are identified and areas for future work are discussed.

Resource contention on shared resources occurs when workload demands exceed the aggregate capacity of shared resources in the community. The token-exchange incentive scheme is traditionally employed to motivate organizations to contribute sufficiently to the community, as a means to minimize free riding. The same incentive scheme can concurrently be used to serve as a mechanism for performing admission control on jobs submitted by users. However, due to the likelihood of fluctuations in demand for computing resources, the initial assignment of tokens on the basis of each organization's resource contribution may have a significant impact on the performance trade-off between fairness and the system admission ratio. To address this problem, we extend the token-exchange scheme by designing trading policies that are responsive to the instantaneous degree of contention, so that, the trade-off between fairness and the admission ratio is less sensitive to the actual quantity of tokens assigned to each organization.

We present an event-driven video adaptation system in this paper. Events are detected by audio/video analysis and annotated by the description schemes (DSs) provided by MPEG-7 multimedia description schemes (MDSs). And then, adaptation take account of users' preference of events and network characteristic to adapt video by event selection and frame dropping as following three steps: 1) the event information is parsed from MPEG-7 annotation XML file together with bitstream to generate generic bitstream syntax description (gBSD), 2) users' preference, network characteristic and adaptation QoS (AQoS) are considered for making adaptation decision, 3) adaptation engine automatically parses adaptation decisions and gBSD to achieve adaptation. Different from most existing adaptation work, the system adapts video by interesting events according to users' preference. To achieve a generic adaptation solution, the system is developed following MPEG-7 and MPEG-21 standards. gBSD based adaptation avoids complex video computation. 30 students from various departments test the system with satisfaction. Although, the system is tested on basketball video adaptation so far, it is easy to extend to other video domains

In this paper, we present an event-on-demand (EoD)video adaptation system. The proposed system supports users in deciding their events of interest and considers network conditions to adapt video source by event selection and frame dropping.Firstly, events are detected by audio/video analysis and annotated by the description schemes (DSs)provided by MPEG-7 Multimedia Description Schemes (MDSs). And then, to achieve a generic adaptation solution, the adaptation is developed following MPEG-21 Digital Item Adaptation (DIA)framework. We look at early release of the MPEG-21 Reference Software on XML generation and develop our own system for EoD video adaptation in three steps:1) the event information is parsed from MPEG-7 annotation XML file together with bitstream to generate generic Bitstream Syntax Description (gBSD). 2) Users' preference, Network Characteristic and Adaptation QoS (AQoS) are considered for making adaptation decision. 3) adaptation engine automatically parses adaptation decisions and gBSD to achieve adaptation.Unlike most existing adaptation work, the system adapts video of events with interest according to users' preference. Implementation following MPEG-7 and MPEG-21 standards provides a generic video adaptation solution. gBSD based adaptation avoids complex video computation. 30 students from various departments were invited to test the system and their responses has been positive.

This paper presents a rate control algorithm for the dominant discrete cosine transform (DCT)-based video coding. It is developed based on a more accurate rate-distortion (R-D) model, specifically, a new distortion-quantization (D-Q) model. Different from previous work that employs a uniform D-Q model or an empirical distortion model, our work proposes an accurate distortion model, which can quantitatively describe the relationship of distortion with respect to video source information and the selected quantization resolution. Based on understanding the distribution of source frequency coefficients and the quantization theory, our distortion model is proposed. This distortion model is combined with the classical R-D theory to generate a new rate model. Finally, the proposed model is implemented on an MPEG-4 encoder to perform rate control for a low-delay visual communication system. We also compare the proposed rate control with the VM18 rate control and it has been shown to be more efficient

A cooperative virtual organization (VO) is formed when distinct organizations pool their computing and data resources together. In a typical VO, there is no central authority that governs the amount of resources that each organization should contribute to the community. To prevent free-riding on shared resources, we introduce policies to curb excessive usage in an equitable manner. While centralized schemes are simple, they are generally inefficient due to the presence of irregular workload traffic. This paper theoretically demonstrates how under specific conditions, an economy-based framework can be designed to achieve fairness. From our formalization, we conceptually show that agent homogeneity and load-based pricing schemes on shared resources can help achieve this requirement.

We apply the mobility support of IPv6 to a wireless network where flat ad hoc routing protocol is used for interconnecting mobile routers. The use of flat routing protocol removes the need for more than one level of recursive nesting to form a wireless network backbone. We introduce a redirect header for mobile IPv6, which reduces the overhead of tunnelling by half in wireless network where bandwidth is limited. Through numerical evaluations, we show that this header aids in reducing the tunnelling overhead. We also discuss the current route optimisation proposals and their applicability to our network scenarios.

In hybrid wireless/wired network, TCP suffers severe performance degradation due to the lack of inference of packet loss. TCP Veno is a novel refinement that is able to diagnose the different causes of packet loss. Based on the diagnosis, TCP Veno takes the necessary control actions, rather than "blind" reduction of throughput without any discrimination. It has been demonstrated that, in this way, TCP Veno achieves better performance than TCP Reno in wireless/wired environment. However, various uncertainties of network may impact the accuracy of TCP Veno's packet loss identification. In this paper, we consider the problem of bursty congestion, as the cause of packet loss is misdiagnosed. To solve this problem in TCP Veno, we proposed a refined packet loss distinguishing scheme. The new TCP Veno is called TCP Veno+. The experimental results show that in environment with bursty congestion, TCP Veno+ can achieve much more accurate congestion loss identification as well as better friendlessness to TCP Reno

This paper presents the experiences gained from ongoing research collaboration between the School of Computer Engineering at Nanyang Technological University and the Southampton e-Science centre at the University of Southampton using a service-oriented approach for complex engineering design optimisation. The service-oriented approach enables programmatic collaboration to be realized while maintaining the autonomy of individual codes at the different institutes and organizations. In the current work, a genetic algorithm optimisation logic implemented as a grid service at Southampton is used to drive the design search process, while the aerodynamic analysis code located in Singapore is used to evaluate the objective function of the design points. Experience gathered from the current study on airfoil shape optimisation is valuable for establishing effective, efficient and customised programmatic links between institutions to solve complicated engineering design problems.

In this paper, a closed form network connectivity formula for a one-dimensional mobile ad hoc network (MANET) is developed. Precisely, we derive the probability that a MANET is fully connected given a certain number of nodes randomly and uniformly placed along a path between a source and destination pair. This formula is particularly useful in the process of design and deployment of a MANET. The formula also provides a critical constraint function to the problem of network optimization. It formulates the relationship between the number of mobile nodes required for a particular MANET given a desired network connectivity probability. An approximation is employed to achieve the final closed form expression. The approximation is then tested by simulation to show the accuracy of our formula under practical network conditions.

This paper proposes a new bit allocation scheme for fine-granular-scalability (FGS) video coding, through which we can achieve better video quality. Different from traditional rate-distortion (R-D) optimization schemes, we consider the characteristics of the bit-plane (BF) coding method. To be specific, we first find the approximate linear relationship between the bit rate of the FGS-layer and the percentage of nonzero binary-scaled coefficients (NZBC) in each BF; second, with mathematical justification, we derive an optimal strategy by analyzing the overall distortion with respect to NZBC. Finally, we perform our optimum bit allocation (OBA) on a FGS coder. Experimental results prove that our scheme can achieve smooth video quality with a higher average PSNR gain compared with uniform bit allocation (UBA). And for certain frames with lower PSNR, it has a gain of up to 3 dB. It is highly source-independent and more robust compared with previous bit allocation schemes.

As more and more Web services are deployed, Web service's discovery mechanisms become essential. Similar services can have quite different QoS levels. For service selection and management purpose, it is necessary to explicitly, precisely, and unambiguously specify various constraints and QoS metrics for Web services descriptions. This paper provides a novel DAML-QoS ontology as a complement for DAML-S ontology to provide a better QoS metrics model. Three layers are defined together with clear role descriptions for developments. Cardinality constraints are utilized to describe the QoS property constraints. Basic profile is presented for general Web service's description and the speed startup of ontology definition. Matchmaking algorithm for QoS property constraints is presented and different matching degrees are described. When incorporated with DAML-S, multiple service levels can be described through attaching multiple QoS profiles to one service profile. Well-defined Metrics can be further utilized by measurement organizations to guarantee the promised service level.

In wireless networks, random loss due to bit error leads to significant performance degradation to conventional multicast congestion control schemes. In this paper, we propose and study an enhanced multicast congestion control scheme, called EPGMCC (enhanced PGMCC), which achieves significant performance improvement over PGMCC over loss-prone wireless links. The key idea of EPGMCC is to discriminate random loss from congestion loss according to the measured network congestion level, and perform different actions accordingly. Specifically: 1) at the sender, it adapts the AIMD algorithms in TCP Veno (Cheng Peng Fu et al., 2003) to avoid the unnecessary window-halving induced by random loss and 2) at the receiver, a new scheme is designed to better measure the loss rate and thus help the sender to select the right representative of the group. Our extensive simulation results demonstrate that EPGMCC achieves significant throughput improvement over PGMCC in hybrid wired/wireless networks, furthermore, the improved throughput achieved by EPGMCC is not grabbed from the bandwidth of coexisting connections but from the utilization of available bandwidth that left unused. Throughput improvement of up to 40% can be demonstrated over typical wireless access link with 1% random loss rate.

Diverse links (i.e., wireless links, satellite links and ADSL links) are being widely deployed in current Internet, unlike wired links, these heterogeneous links are causing significant performance degradation of TCP. Recently one sender-side enhancement of TCP, called Veno TCP, is proposed to mainly eliminate TCP's suffering in wireless environments. Real network measurements and live Internet results validated Veno's throughput improvement and its harmonious co-existence with legacy TCP connections. In this paper, we revisit Veno TCP and evaluate its performance in more practical way. Specifically, we measure Veno from four metrics - compatibility, flexibility, robustness and deployablity. Our extensive arguments not only prove Veno's advantages, but also illuminate some basic philosophies behind Veno, which could provide helpful guidelines for future protocol design.