Federated Learning (FL) is a privacy-preserving technique that allows clients to collaboratively train models while keeping their local data private from both other clients and the central server. However, the presence of a malicious server or attacker can lead to model inversion attacks, which compromise clients’ private information. To counteract these threats, homomorphic encryption (HE) has been proposed as a solution, enabling encrypted model aggregation by the server without exposing raw data. This ensures that a malicious server or attacker cannot access model parameters or reveal clients’ local information. However, HE introduces substantial computational and communication overhead, making it impractical for many real-world implementations. To address this, recent research suggests selectively encrypting only a subset of most important model parameters while leaving the rest unencrypted. While it is promising, the challenge lies in identifying these key parameters to maintain security while minimizing the additional costs associated with homomorphic operations. In addition, even with a subset of parameters, none of the existing methods offer truly efficient and practical HE-based solution. In general, these approaches fall short when applied to large, complex deep learning models and are impractical beyond basic implementations. This paper introduces a straightforward yet effective parameter selection and encryption strategy for HE-based federated learning. Our goal is to minimize this computational, and communication overhead caused by homomorphic operations while maintaining robust security, ensuring practical applicability in real-world scenarios. We begin by evaluating different implementations of CKKS (Cheon-Kim-Kim-Song)-based HE algorithm (which allows floating point operations on encrypted domain) across various model architectures, systematically analyzing the additional burden HE imposes in terms of processing time and size of the model parameters. Next, we introduce a magnitude-based selective encryption strategy that not only provide same security guarantees compared to existing methods but is also practical for scaling to larger, more complex models in FL environments. Hence, this approach enables developers to make informed decisions that strike an optimal balance between performance and privacy/security, making it feasible for real-world deployments.

In recent years, machine learning as a service (MLaaS) has brought considerable convenience to our daily lives. However, these services raise the issue of leaking users’ sensitive attributes, such as race, when provided through the cloud. The present work overcomes this issue by proposing an innovative privacy-preserving approach called privacy-preserving class overlap (PPCO), which incorporates both a Wasserstein generative adversarial network and the idea of class overlapping to obfuscate data for better resilience against the leakage of attribute-inference attacks(i.e., malicious inference on users’ sensitive attributes). Experiments show that the proposed method can be employed to enhance current state-of-the-art works and achieve superior privacy–utility trade-off. Furthermore, the proposed method is shown to be less susceptible to the influence of imbalanced classes in training data. Finally, we provide a theoretical analysis of the performance of our proposed method to give a flavour of the gap between theoretical and empirical performances.

In machine learning, boosting is one of the most popular methods that is designed to combine multiple base learners into a superior one. The well-known Boosted Decision Tree classifier has been widely adopted in data mining and pattern recognition. With the emerging challenge in privacy, the personal images, browsing history, and financial reports, which are held by individuals and entities are more likely to contain sensitive information. The privacy concern is intensified when the data leaves the hand of owners and is used for further mining. Such privacy issues demand that the machine learning algorithms should be privacy-aware. Recently, Local Differential Privacy has been proposed as an effective privacy protection approach, which allows data owners to perturb the data before any release. In this paper, we propose a distributed privacy-preserving boosting algorithm that can be applied to various types of classifiers. By adopting LDP as a building block, the proposed boosting algorithm leverages the aggregation of the perturbed data shares to build the base learner, which ensures that privacy is well preserved for the participated data owners. Our experiments demonstrate that the proposed algorithm effectively boosts various classifiers and the boosted classifiers maintain a high utility.

For over many decades, relational database model has been considered as the leading model for data storage and management. However, as the Big Data explosion has generated a large volume of data, alternative models like NoSQL and NewSQL have emerged. With the advancement of communication technology, these database systems have given the potential to change the existing architecture from centralized mechanism to distributed in nature, to deploy as cloud-based solutions. Though all of these evolving technologies mostly focus on performance guarantees, it is still being a major concern how these systems can ensure the security and privacy of the information they handle. Different datastores support different types of integrated security mechanisms, however, most of the non-relational database systems have overlooked the security requirements of modern Big Data applications. This paper reviews security implementations in today's leading database models giving more emphasis on security and privacy attributes. A set of standard security mechanisms have been identified and evaluated based on different security classifications. Further, it provides a thorough review and a comprehensive analysis on maturity of security and privacy implementations in these database models along with future directions/enhancements so that data owners can decide on most appropriate datastore for their data-driven Big Data applications.

Community detection is of great importance for online social network analysis. The volume, variety and velocity of data generated by today's online social networks are advancing the way researchers analyze those networks. For instance, real-world networks, such as Facebook, LinkedIn and Twitter, are inherently growing rapidly and expanding aggressively over time. However, most of the studies so far have been focusing on detecting communities on the static networks. It is computationally expensive to directly employ a well-studied static algorithm repeatedly on the network snapshots of the dynamic networks. We propose DynaMo, a novel modularity-based dynamic community detection algorithm, aiming to detect communities of dynamic networks as effective as repeatedly applying static algorithms but in a more efficient way. DynaMo is an adaptive and incremental algorithm, which is designed for incrementally maximizing the modularity gain while updating the community structure of dynamic networks. In the experimental evaluation, a comprehensive comparison has been made among DynaMo, Louvain (static) and 5 other dynamic algorithms. Extensive experiments have been conducted on 6 real-world networks and 10,000 synthetic networks. Our results show that DynaMo outperforms all the other 5 dynamic algorithms in terms of the effectiveness, and is 2 to 5 times (by average) faster than Louvain algorithm.

Supervised learning techniques such as classification algorithms learn from training data to predict the correct label for newly presented input data. In many real-world scenarios, training data required by such techniques can contain personal information and data collection can be a significant problem due to privacy concerns. Cryptographic techniques have been used before to do training on encrypted data. However, such techniques are computationally expensive and they are not scalable most of the time. If a dataset in another party will be used for training, differential privacy technology can be used to preserve the privacy of the individuals in the dataset. When there is no such dataset and data needs to be collected from individuals directly for training, local differential privacy can be used. Local differential privacy is a technology to preserve privacy during data sharing with an untrusted data collector. In this work, we propose to use local differential privacy techniques to train a Naive Bayes classifier. Using the proposed solution, an untrusted party collects perturbed data from individuals that keep the relationship between the feature values and class labels. By estimating probabilities needed by the Naive Bayes classifier using the perturbed data, the untrusted party can classify new instances with high accuracy. We develop solutions that work for both discrete and continuous data. We also propose utilizing dimensionality reduction techniques to decrease communication cost and improve accuracy. We show the accuracy of the proposed Naive Bayes classifier achieving local differential privacy via experiments on several datasets. We also show how dimensionality reduction enhances the accuracy.

Peer-to-peer (P2P) botnets have become one of the major threats in network security for serving as the fundamental infrastructure for various cyber-crimes. More challenges are involved in the problem of detecting P2P botnets, despite a few work claimed to detect centralized botnets effectively. We propose an enhanced PeerHunter, a network-flow level community behavior analysis based system, to detect P2P botnets. Our system starts from a P2P network flow detection component. Then, it uses “mutual contacts” to cluster bots into communities. Finally, it uses network-flow level community behavior analysis to detect potential botnets. In the experimental evaluation, we propose two evasion attacks, where we assume the adversaries know our techniques in advance and attempt to evade our system by making the P2P bots mimic the behavior of legitimate P2P applications. Our results showed that enhanced PeerHunter can obtain high detection rate with few false positives, and high robustness against the proposed attacks.

Today's data center networks (DCNs) tend to have tens to hundreds of thousands of servers that provide massive and sophisticated services. The architectural design of DCNs are usually over-provisioned for peak workloads and fault tolerance. Statistically, DCNs remain highly under-utilized, with typical utilization of around 30%. Network over-provisioning and under-utilization can be exploited for energy-saving. Most research efforts on DCN energy saving focus on how to save maximum energy but have little or no consideration to the performance of the residual network. Thus, the DCN performance can become degraded and the network left vulnerable to sudden traffic surges. In this paper, we have studied the energy-saving problem in DCNs while preserving network performance. The problem was formulated as mixed integer linear problem (MILP) solvable by CPLEX in order to minimize the energy consumed by DCN; meanwhile, safety threshold constraints for links utilization are met. To overcome CPLEX high computational time, a heuristic algorithm to provide practical and efficient solution for the MILP is introduced. The heuristic algorithm uses switches grouping and links consolidation to switch the traffic to a small number of network devices and turn off unused switches and links. Valiant load balancing is used to distribute the loads over active links. Simulation experiments using synthetic and real packet traces were conducted to validate the heuristic in terms of energy consumption and network performance. The results show that the heuristic can save up to 45% of the network energy and improve the average imbalance scores for links and switches by more than 50% with minimal effect on network performance.

As the analytic tools become more powerful, and more data are generated on a daily basis, the issue of data privacy arises. This leads to the study of the design of privacy-preserving machine learning algorithms. Given two objectives, namely, utility maximization and privacy-loss minimization, this work is based on two previously non-intersecting regimes - Compressive Privacy and multi-kernel method. Compressive Privacy is a privacy framework that employs utility-preserving lossy-encoding scheme to protect the privacy of the data, while multi-kernel method is a kernel-based machine learning regime that explores the idea of using multiple kernels for building better predictors. In relation to the neural-network architecture, multi-kernel method can be described as a two-hidden-layered network with its width proportional to the number of kernels. The compressive multi-kernel method proposed consists of two stages - the compression stage and the multi-kernel stage. The compression stage follows the Compressive Privacy paradigm to provide the desired privacy protection. Each kernel matrix is compressed with a lossy projection matrix derived from the Discriminant Component Analysis (DCA). The multikernel stage uses the signal-to-noise ratio (SNR) score of each kernel to non-uniformly combine multiple compressive kernels. The proposed method is evaluated on two mobile-sensing datasets - MHEALTH and HAR - where activity recognition is defined as utility and person identification is defined as privacy. The results show that the compression regime is successful in privacy preservation as the privacy classification accuracies are almost at the random-guess level in all experiments. On the other hand, the novel SNR-based multi-kernel shows utility classification accuracy improvement upon the state-of-the-art in both datasets. These results indicate a promising direction for research in privacy-preserving machine learning.

The power-saving opportunities of long-running application servers which execute on multi-core systems are studied in this paper. The research goal is to develop an efficient power-saving strategy of application servers with the minimum performance degradation in cloud environments. The power-saving strategy is based on the run-time information which is already available in a JVM, the base software component of application servers. Several key findings are revealed through this study. First, the particular behavior of application servers, also known as phases, can be related to the run-time information of a JVM. Thus the phases of an application server can be predicted before the applications actually execute on hardware. Secondly, some particular phases are observed in this study and used to establish the power-saving strategy, such as memory phases and execute phases. Finally, a new finding of idle phase is proposed to reduce significant energy wastage without performance degradations. Based on these findings, a set of power-saving algorithms is proposed and implemented with two widely used JVMs, Sun's Hotspot and Jikes RVM. With the experiments of five multi-threaded benchmarks and two web application benchmarks, the use of proposed power-saving strategy leads to the lowest value of energy-delay product among the other power-saving techniques, and the performance degradation is well below 6 percent.

In this paper, a fast kernel ridge regression (KRR) learning algorithm is adopted with ( ) training cost for large-scale active authentication system. A truncated Gaussian radial basis function (TRBF) kernel is also implemented to provide better cost-performance tradeoff. The fast-KRR algorithm along with the TRBF kernel offers computational advantages over the traditional support vector machine (SVM) with Gaussian-RBF kernel while preserving the error rate performance. Experimental results validate the cost-effectiveness of the developed authentication system. In numbers, the fast-KRR learning model achieves an equal error rate (EER) of 1.39% with ( ) training time, while SVM with the RBF kernel shows an EER of 1.41% with ( ) training time.

In the Internet era, the data being collected on consumers like us are growing exponentially, and attacks on our privacy are becoming a real threat. To better ensure our privacy, it is safer to let the data owner control the data to be uploaded to the network as opposed to taking chance with data servers or third parties. To this end, we propose compressive privacy , a privacy-preserving technique to enable the data creator to compress data via collaborative learning so that the compressed data uploaded onto the Internet will be useful only for the intended utility and not be easily diverted to malicious applications. For data in a high-dimensional feature vector space, a common approach to data compression is dimension reduction or, equivalently, subspace projection. The most prominent tool is principal component analysis (PCA). For unsupervised learning, PCA can best recover the original data given a specific reduced dimensionality. However, for the supervised learning environment, it is more effective to adopt a supervised PCA, known as discriminant component analysis (DCA), to maximize the discriminant capability. The DCA subspace analysis embraces two different subspaces. The signal-subspace components of DCA are associated with the discriminant distance/power (related to the classification effectiveness), whereas the noise subspace components of DCA are tightly coupled with recoverability and/or privacy protection. This article presents three DCA-related data compression methods useful for privacy-preserving applications: — Utility-driven DCA : Because the rank of the signal subspace is limited by the number of classes, DCA can effectively support classification using a relatively small dimensionality (i.e., high compression). — Desensitized PCA : By incorporating a signal-subspace ridge into DCA, it leads to a variant especially effective for extracting privacy-preserving components. In this case, the eigenvalues of the noise-space are made to become insensitive to the privacy labels and are ordered according to their corresponding component powers. — Desensitized K-means/SOM : Since the revelation of the K-means or SOM cluster structure could leak sensitive information, it is safer to perform K-means or SOM clustering on a desensitized PCA subspace.

Heterogeneous networks have been pointed out to be one of the key network architectures that help increase system capacity and reduce power consumption for efficient communications. Although conceivably, high operational efficiency brings a high profit for mobile service providers, it is noteworthy that the potential for maximizing the profit has not been explored for the heterogeneous environment. This paper investigates profitability for network operators with the spectrum-energy efficiency metric on the downlink of LTE Advanced communication systems. We pursue optimal policies by employing the techniques of cell size zooming, user migration, and sleep mode in the deployment of different base station types. The problem is formulated as a quasiconvex optimization problem and it is transformed into an equivalent form of the MILP problem; the former is solved with a bisection algorithm and the latter is approached by an off-the-shelf software package. Since the formulated optimization problem is NP hard, a sub-optimal approach with a lower computational complexity is also proposed. Numerical analysis through case studies are presented to evaluate the efficiency improvements, and demonstrate the performance of the near-optimal solution.

Presents the introductory welcome message from the conference proceedings. May include the conference officers' congratulations to all involved with the conference event and publication of the proceedings record.

To steadily gaining benefit from the exponential growth in mobile traffic, operators are eager to find solutions to maximize profits. Two very attractive strategies have been proposed to complement the existing macro cellular architecture: deploying low power bases stations and offloading data traffic to other networks. Each strategy has different costs and yields different benefits for operators. The offloading option could be cheaper in the short run; nevertheless, it might be more expensive in the long run than cell densification due to the varying cost. On the other hand, small cells, since having to be deployed in advance, may be underutilized or not fully meet future demands. In the latter case, offloading techniques can be used to increase capacity with additional costs. Further, uncertainty of future data demands and electricity prices also impact operators profitability, making the best network strategy difficult to achieve. To address such problem, an optimal cell configuration algorithm is proposed by formulating a stochastic programming model that considers both network design and data offloading. This algorithm can maximize the profit, under future demand and price uncertainty. Numerical studies are extensively performed in which the results show that operators' profits can be improved with our proposed algorithm.

Continuous authentication for mobile devices using behavioral biometrics is being suggested to complement initial authentication for securing mobile devices, and the cloud services accessed through them. This area has been studied over the past few years, and low error rates were achieved; however, it was based on training and testing using support vector machine (SVM) and other non-privacy-preserving machine learning algorithms. To stress the importance of carefully designed privacy-preserving systems, we investigate the possibility of reconstructing gestures raw data from users' authentication profiles or synthesized samples' testing results. We propose two types of reconstruction attacks based on whether actual user samples are available to the adversary (as in SVM profiles) or not. We also propose two algorithms to reconstruct raw data: a numerical-based algorithm that is specific to one compromised system, and a randomization-based algorithm that can work against almost any compromised system. For our experiments, we selected one compromised and four attacked gesture-based continuous authentication systems from the recent literature. The experiments, performed using a public data set, showed that the attacks were feasible, with a median ranging from 80% to 100% against one attacked system using all types of attacks and algorithms, and a median ranging from 73% to 100% against all attacked systems using the randomization-based algorithm and the negative support vector attack. Finally, we analyze the results, and provide recommendations for building active authentication systems that could resist reconstruction attacks.

With the continuing growth of cloud computing services, power consumption has become one of the most challenging issues in data center environments. With the support of today's virtualization technology, the efficiency and flexibility of data center management can be greatly enhanced, creating great energy saving opportunities. However, effective energy aware design is a non-trivial task, considering the size of the data center, the dynamic fluctuation of workloads and the variation of computing resource requests. In this paper, we propose Cool Cloud: a practical solution for managing the mappings of VMs to physical servers. This framework solves the problem of finding the most energy efficient way (least resource wastage and least power consumption) of placing the VMs considering their resource requirements. Experiment result demonstrates our design can effectively improve data center energy efficiency and scales well to large size data centers. Comparing with industry leading product VMware's Distributed Resource Scheduler (DRS), our design offers better performance in both load balancing and power consumption.

Current system administrators are missing intrusion alerts hidden by large numbers of false positives. We propose an intrusion detection tool that effectively uses select data to provide a picture of "network health". Our hypothesis is that by utilizing the data available at the node and network levels we can create a synoptic picture of the network providing indications of many intrusions or other network issues. Our major contribution is to provide a revolutionary way to analyze node and network data for patterns, dependence, and effects that indicate network issues. Our first contribution in this vein is to present a method based on utilizing the number of packets sent, number of packets received, node reliability, route reliability, and entropy to develop a synoptic picture of the network health in the presence of a sinkhole.

Presents a listing of the Symposium organizing committee.

With today's limited bandwidth, high data rate services, and energy efficiency requirements, maximizing the spectral efficiency and minimizing the consumed power becomes essential. Investigating the issues impeding spectral efficiency maximization and consumed power minimization for mobile systems is crucial for solving this contemporary problem. This paper aims to optimize the utilization of the scarce mobile spectrum and the amount of power consumption in the multi-cell IEEE 802.16m networks. A return on investment model adopting a utility optimization technique is proposed; the model objective is to increase the revenue of the mobile service providers by maximizing the normalized spectral efficiency and decrease the operational cost by minimizing the power consumed in the network. Based on this model, we propose two phases scheme (distributed and centralized) to solve the joint spectral and power optimization problem. The problem is solved to identify the optimal distributed resources assignment and the central down-link frequency partition configuration that achieves the model objective. Simulation results show that the optimal solution significantly improves the system power consumption while maintaining the normalized spectral efficiency, yet with high computational complexity. Accordingly, an effective suboptimal solution utilizing a polynomial-time heuristic is proposed for practical implementations.

In this study a keystroke-based authentication system is implemented on a large-scale free-text keystroke data set, where cost effective kernel-based learning algorithms are designed to enable trade-off between computational cost and accuracy performance. The authentication process evaluates the user's typing behavior on a vocabulary of words, where the judgments based on each word are concatenated by weighted votes, whose weights are also trained to provide optimal fusion of independent judgments. A novel truncated-RBF kernel is also implemented to provide better cost-performance trade-off. Experimental results validate the cost-effectiveness of the developed authentication system.

Fifth generation mobile systems (5G) target an Average Area Spectral Efficiency (AASE) over hundred Gbps/km2/user for future mobile systems with an Energy Dissipation (ED) per unit area similar to the current ED levels. Heterogeneous networks (HetNets) with high density of deployed small cells are currently adopted to aid in achieving the target ED and AASE by 5G. Limited spectrum availability requires efforts to manage the spectrum utilization in such dense deployments. Development of new network architectures and Radio Resources Management (RRM) schemes is important to address such challenges. The objective of this work is to propose a new architecture that consists of a Decision Support System (DSS) and a data collection system to dynamically manage and control the spectrum allocation process. The DSS generates spectrum allocation patterns using non-parametric estimation and statistical analysis for the collected data. A new RRM model using a Plan, Do, Control and Act (PDCA) cycle is proposed as a new self optimization module in the self organizing network framework. The PDCA model utilizes the new architecture and the allocation patterns to dynamically predict future spectrum allocation. Results show improvement in the AASE achieved using the PDCA model compared to conventional spectrum allocation.

Rapid growth of cloud computing services have led to creation of large scale enterprise data centers which consume great amounts of energy. Data centers usually have an service level agreement (SLA) between the clients and the service providers, which specify the terms and quality of service to be provided. In this paper, we consider a situation in a data center where multiple user applications are executing on a multi-core system and each application may have a specified SLA requirement. We design a voltage and frequency scheduler (the “cool” scheduler) that can be used in enterprise data centers to provide CPU energy saving under the specified SLA requirement by exploiting the applications’ run-time program phases. Our design greatly improves the computation efficiency compared to other recently published works. The scheduler is built into the Linux kernel and evaluated against SPEC CPU2006 and Phoronix Test Suite on a quad-core system. Experiment result demonstrates that our cool scheduler achieves 25.8% energy saving on average with 8.7% performance loss under the given SLA requirement (10% allowed performance loss). Our design achieves 35.8% and 31.6% more energy saving compared to two of the most advanced related works.

Recently, the integration of smart antenna technology into existing wireless local area networks (WLANs) has been one of the hot spots of research work. In this paper, we design an IEEE 802.11-compliant medium access control (MAC) protocol, named M-HCCA, that fully takes advantage of multi-beam smart antennas equipped at the access point (AP) to not only boost the overall capacity of a WLAN, but also support quality-of-service (QoS) and power conservation for individual mobile users. Specifically, M-HCCA has the following attractive features: (i) since being a polling-based MAC scheme, M-HCCA can innately conquer the problems induced by carrier sensing or directional signals, including beam-synchronization constraint, receiver blocking problem, and unnecessary defer problem; (ii) M-HCCA achieves high real-time throughput by adaptively adjusting the sector configuration to quickly resolve contention/collision and to increase data transmission parallelism; (iii) M-HCCA employs beam-location-aware polling scheduling to not only solve the beam-overlapping problem and back/side-lobe problem, but also let real-time stations save as much energy as possible; (iv) M-HCCA adopts the mobile-assisted admission control technique such that the AP can admit as many newly streams as possible while not violating QoS guarantees made to already-admitted streams; (v) M-HCCA offers a location updating mechanism to promptly renew the beam-location information of a non-responsive station such that the miss-hit problem can be effectively alleviated. Extensive simulation results show that, in terms of throughput, real-time throughput, and energy throughput, M-HCCA significantly outperforms existing protocols even in uneven station distribution, imperfect beam-forming, and high mobility environments.

The popularity of smart mobile devices has brought significant growth of data services for mobile service providers. Mobile users of data services are charged based on the amount of data used. Raising served data amount seemingly increases the profit; energy consumption rises correspondingly. Besides, spectral resources are licensed and limited for mobile operators to allocate. Increasing data services over the spectrum for the profit does not count the cost of energy. To assess the profitability, considered is the revenue-to-cost ratio. Optimizing the ratio is an economic incentive for mobile operators. Revenue is regarded as efficiency in spectrum use, the cost as energy consumption; therefore we interpret the revenue-to-cost ratio as spectrum-energy efficiency. In this paper, we study the spectrum-energy efficiency optimization problem where BSs are with the ability to perform cell zooming, sleep mode, and user migration. We formulate the problem into an integer linear program which is solvable by CPLEX to maximize spectrum-energy efficiency; meanwhile traffic demands by associated users in multicell/multiuser networks are met. To avoid high computation time, a heuristic algorithm is proposed to efficiently solve the formulated problem. Numerical analysis through case studies demonstrates energy consumption and efficiency improvements, and comparisons between near-optimal solutions against optimality.

Cloud computing provides scalable computing and storage resources. More and more data-intensive applications are developed in this computing environment. Different applications have different quality-of-service (QoS) requirements. To continuously support the QoS requirement of an application after data corruption, we propose two QoS-aware data replication (QADR) algorithms in cloud computing systems. The first algorithm adopts the intuitive idea of high-QoS first-replication (HQFR) to perform data replication. However, this greedy algorithm cannot minimize the data replication cost and the number of QoS-violated data replicas. To achieve these two minimum objectives, the second algorithm transforms the QADR problem into the well-known minimum-cost maximum-flow (MCMF) problem. By applying the existing MCMF algorithm to solve the QADR problem, the second algorithm can produce the optimal solution to the QADR problem in polynomial time, but it takes more computational time than the first algorithm. Moreover, it is known that a cloud computing system usually has a large number of nodes. We also propose node combination techniques to reduce the possibly large data replication time. Finally, simulation experiments are performed to demonstrate the effectiveness of the proposed algorithms in the data replication and recovery.

No abstract available.

The latest generation of Wireless Local Area Networks (WLANs) is based on IEEE 802.11n-2009 Standard. The standard provides very high data rates at the physical layer and aims to achieve a throughput at the Medium Access Control (MAC) layer that is higher than 100 Mbps. To do that, the standard introduces several mechanisms to improve the MAC efficiency. The most notable ones are the use of frame aggregation and Block-ACK frames. The standard, however, does not introduce a mechanism to reduce the probability of collision. This issue is significant because, with a high data rate, an AP would be able to serve a large number of stations, which would result in a high collision rate. In this paper, we propose a Group-based MAC (GMAC) scheme that reduces the probability of collision and also uses frame aggregation to improve the efficiency. The contending stations are divided into groups. Each group has one station that is the group leader. Only the leader stations contend, hence, reducing the probability of a collision. We evaluate the performance of our scheme with analytic and simulation results. The results show that GMAC achieves a high throughput, high fairness, low delay and maintains a high performance with high data rates.

IEEE 802.16-2009 specifies two fast handover mechanisms, fast base station switching and macro diversity handover, to streamline communication for a mobile station (MS). Both operate with a diversity set that lists base stations among which an MS can move its connection readily. In view that an unduly chosen diversity set may cause prohibitive cost, we provide means to prevent the diversity set from including base stations that are less likely to serve the MS in the near future. Inspired from the working-set model, our approach develops predictive handover using numerical extrapolation to accommodate temporal locality of the MS. While a joint entry-replacement strategy is exercised to evict least-preferred entries in the diversity set, its capacity is reviewed and tuned periodically allowing for reasonable space demand. Simulation results show that our approach, compared with the counterpart scheme, reduces handover executions by over 48 percent, handover delay by over 51 percent, and diversity-set space requirement by up to 17.9 percent on average. As another salient strength, our approach conforms fully to the standard, keeping current protocols on the MS side operable without modification. Qualitative and quantitative performance discussions indicate the usefulness of our approach in pragmatic settings.

The performance and scalability issues of multithreaded Java programs on multicore systems are studied in this paper. First, we examine the performance scaling of benchmarks with various numbers of processor cores and application threads. Second, by correlating low-level hardware performance data to JVM threads and system components, the detail analyses of performance and scalability are presented, such as the hardware stall events and memory system latencies. Third, the usages of memory resource are detailed to observe the potential bottlenecks. Finally, the JVM tuning techniques are proposed to alleviate the bottlenecks, and improve the performance and scalability. Several key findings are revealed through this study. First, the lock contentions usually lead to a strong limitation of scalability. Second, in terms of memory access latencies, the most of memory stalls are produced by L2 cache misses and cache-to-cache transfers. Finally, the overhead of minor garbage collections could be an important factor of throughput reductions. Based on these findings, the appropriate Java Virtual Machine (JVM) tuning techniques are examined in this study. We observe that the use of a parallel garbage collector and an appropriate ratio of young to old generation can alleviate the overhead of minor collection and improve the efficiency of garbage collections. Moreover, the cache utilizations could be enhanced with the use of thread-local allocation buffer, and then leads to the performance improvements significantly.

In today's wireless networks, stations using the IEEE 802.11 Standard contend for the channel using the Distributed Coordination Function (DCF). Research has shown that DCF's performance degrades especially with the large number of stations. This becomes more concerning due to the increasing proliferation of wireless devices. In this paper, we present a Medium Access Control (MAC) scheme for wireless LANs and compare its performance to DCF and to other efficient schemes. Our scheme, which attempts to resolve the contention in a constant number of slots (or constant time), is called CONTI. The contention resolution happens over a predefined number of slots. In a slot, the stations probabilistically send a jam signal on the channel. The stations listening retire if they hear a jam signal. The others continue to the next slot. Over several slots, we aim to have one station remaining in the contention, which will then transmit its data. We find the optimal parameters of CONTI and present an analysis on its performance. More comprehensive evaluation is presented in the simulation results where we compare CONTI, DCF, and other efficient schemes from the literature. We consider the number of slots used, the collision rate, the throughput, the delay, and the fairness. The highest throughput was achieved by CONTI. Moreover, our results provide measurements from each of the schemes that we consider and provide the insight on each scheme's operation.

Frauds and attacks for online banking are increasing quickly. The major platform for current online banking, personal computer, has become untrusted especially under malware attacks. In this study, we design a smart card-based solution called secure online banking companion (SOBC) to address this problem. Portability and cost are also highly considered in the design. We have implemented a prototype of the solution on a Java Card simulator, which shows the solution can be implemented easily using current smart card technologies.

IEEE 802.16 based broadband wireless access network is considered as one of the most promising wireless access technologies. It employs a request/grant scheme in bandwidth allocation where each subscriber station (SS) can send bandwidth requests (BRs) to the base station (BS) before bandwidth can be granted to SSs. The 802.16 standard defines two types of BR strategies, namely incremental requests and aggregate requests. In this paper, we first present an analytic model to analyze their performance and show that both types of BRs have rooms for improvement when facing two common types of traffic patterns, called the uphill traffic pattern and the periodic/bursty traffic pattern, in terms of overhead and data waiting time in the queue. Then, we propose two enhancement strategies, called the aggressive strategy and the conservative strategy, that exploit the property of the two traffic patterns to improve the performance. The simulation results show that the two enhancement strategies can effectively reduce the overhead and data waiting time, at the price of somewhat less bandwidth utilization in the case of conservative strategy.

Broadband wireless access networks have received a tremendous amount of research and development in the recent years. There have also been pilot networks deployed in many cities around the globe. In the IEEE 802.16j standard, Relay Stations (RS) play a promising role of extending the range of a Base Station (BS). This architecture is suitable to areas with limited infrastructure, such as rural areas, since it is difficult to install many BSs, with each having a wired connection. In this paper, we present an optimization model that finds the number of RSs and their locations to serve a customer base. We also show how our model can be adapted to make the planning in real-life scenarios where there are obstacles, such as mountains and lakes, in the planning area.

In this paper, we study the upper bounds of memory storage for two different allocators. In the first case, we consider a general allocator that can allocate memory blocks anywhere in the available heap space. In the second case, a more economical allocator constrained by the address-ordered first-fit allocation policy is considered. We derive the upper bound of memory usage for all allocators and present a systematic approach to search for allocation/deallocation patterns that might lead to the largest fragmentation. These results are beneficial in embedded systems where memory usage must be reduced and predictable because of lack of swapping facility. They are also useful in other types of computing systems.

IEEE 802.16 standard was designed to support the bandwidth demanding applications with quality of service (QoS). Bandwidth is reserved for each application to ensure the QoS. For variable bit rate (VBR) applications, however, it is difficult for the subscriber stations (SSs) to predict the amount of incoming data. To ensure the QoS guaranteed services, the SS may reserve bandwidth more than the amount of its transmitting data. As a result, the reserved bandwidth may not be fully utilized all the time. In this paper, we propose a scheme, named Bandwidth Recycling, to recycle the unused bandwidth without changing the existing bandwidth reservation. The idea of our scheme is to allow other SSs to utilize the unused bandwidth when it is available. Thus, not only the same QoS guaranteed services can be provided but also the system throughput can be improved. Mathematical analysis and simulation are used to evaluate the proposed scheme. Simulation and analysis results confirm that our proposed scheme can recycle 35 percent of unused bandwidth on average. By analyzing factors affecting the recycling performance, three scheduling algorithms are proposed to improve the overall throughput. The simulation results show that our proposed algorithm can further improve the overall throughput by 40 percent when the network is in the steady state.

In IEEE 802.11 wireless networks, EDCA users' performance may be degraded because of the existence of legacy users and therefore would get a lower priority service. Such effects are mainly due to the fact that EDCA users are controlled by different contention parameters that are distributed in the beacon frames, but there is no control over legacy users as their contention parameters are PHY dependent, i.e. they have constant values. We discuss different aspects of the legacy DCF and EDCA users coexistence. Also, we propose a simple distributed management scheme (called NZ-ACK) to mitigate the influence of legacy DCF on EDCA performance in networks consisting of both types of users without any modifications to legacy users. Finally, we use Opnet simulation to evaluate the performance of NZ-ACK. Results show that NZ-ACK outperforms 802.11 in terms of maintaining the priority of service and delay bounds of EDCA users while providing acceptable throughput for legacy users.

Inter-Vehicle Communications (IVC) and Roadside-to-Vehicle Communications (RVC) in vehicular networks based on IEEE 802.11 are emerging technologies for future Intelligent Transportation Systems (ITS). This paper presents a new efficient routing approach, called RAR (Roadside-Aided Routing), that is the first one to exploit the unique characteristics of vehicular networks. A novel affiliation method is proposed to affiliate a vehicle to several Roadside Units based on road constraints. This scheme allows (a) agent advertisement to be broadcast in one hop (instead of multi-hop), (b) routing to be done in single phase, comparing to two phases, (c) to eliminate the use of hierarchical addressing which is commonly used in single-phase schemes. Simulation results of ns-2 show that RAR approach can provide a high packet delivery rate in vehicular networks with a low and constant overhead.

Recently, there has been extensive research interest in increasing the data rates supported by IEEE 802.11 wireless LANs. For this purpose, IEEE 802.11 formed Task Group N to develop specifications for high-data-rate wireless LANs. The medium access in the legacy 802.11 is not scalable as it exhibits a large control overhead when the data rates increase and a large collision rate when the number of stations is large. In this paper, we introduce a group-based medium access control (GMAC) protocol for wireless LANs with high data rates and a large number of stations. With GMAC, stations are divided into groups that are free of hidden nodes. Each group has a leader and only group leaders contend using CSMA/CA. Once a group leader wins the contention, it reserves transmission time for all the stations in its group and issues a polling packet specifying the group's schedule. Stations in the same group transmit after their leader according to the polling packet. GMAC achieves significant throughput gain over DCF by reducing the collision rate and the control overhead. Simulation studies show that GMAC maintains a high throughput as the data rates and the number of stations increase

Performance of modern computers is tied closely to the effective use of cache because of the continually increasing speed discrepancy between processors and main memory. We demonstrate that generational garbage collection employed by a system with cache and scratchpad memory can take advantage of the locality of small short-lived objects in Java and reduce memory traffic by as much as 20% when compared to a cache-only configuration. Converting half of the cache to scratchpad can be more effective at reducing memory traffic than doubling or even quadrupling the size of the cache for several of the applications in SPECjvm98.

As Java is emerging as one of the major programming languages in software development, studying how Java applications behave on recent SMT processors is of great interest. This paper characterizes the performance of Java applications on an Intel Pentium 4 hyper-threading processor. Using the performance counters provided by Pentium 4, we quantitatively evaluate micro-architecture metrics while running various types of Java applications. The experimental results reveal that: (1) Hyper-threading can indeed improve the performance of multithreaded Java programs; (2) The resource contentions within Pentium 4 are the major reason of pipeline inefficiency, which prevents better performance promised by SMT; (3) The static partition design of hyper-threading causes considerable performance loss for many single-thread Java programs; (4) Most multiprogrammed Java benchmarks can achieve decent combined speedups on hyper-threading processors

This paper presents a detailed design and implementation of a power-efficient garbage collector for Java embedded systems. The proposed scheme is a hybrid between the standard mark-sweep-compact collector available in Sun's KVM and a limited-field reference counter. There are three benefits resulting from the proposed scheme. (a) the proposed scheme reclaims memory more efficiently and this results in less mark-sweep garbage collection invocations, (b) reduction in garbage collection invocations improves cache locality and reduces the number of main memory accesses, and (c) reduction in memory access ultimately results in lower energy consumption, since a memory access can consume a large amount of energy when compared with an instruction execution. The proposed scheme has been implemented into Sun's KVM, and has been shown to reduce the number of mark-sweep garbage collection invocations by up to 100% in some cases, and the number of level-1 cache misses by as much as 87% when compared to the default garbage collector. We also find that in some applications, the proposed scheme can reduce the power consumption by as much as 27% when compared to the default Sun's KVM.

This paper investigates various issues of pairing Java applications for multithreaded execution on Intel's hyper-threading Pentium 4 processor. We first quantify the overall performance of multiprogrammed Java applications using a metric called combined speedup. Using the performance counters provided by the Pentium 4, we then quantitatively evaluate the performance of underneath micro-architecture components and their implications to the combined speedup. A statistical model is proposed to analyze the collected data. This novel approach reveals that trace cache is the major factor determining the pairing performance. In particular, we find that the trace cache miss rates of Java applications can be utilized to predict the combined speedups. Three new scheduling strategies are proposed based on these observations and then evaluated. The experimental results show that the proposed strategies have better performance than the conventional round-robin scheduling scheme. Overall, our best strategy enables a reduction in execution time of 10.5% over the serial execution, comparing with a reduction of 5.92% achieved by the round-robin scheduling. The improvement will be increasingly significant on future SMT processors.

Quality of service (QoS) support in wireless LANs is a challenging problem. In this paper, we propose a QoS enhancement over the legacy point coordination function (PCF) of IEEE 802.11, called enhanced PCF (EPCF). EPCF enables wireless LANs to send a combination of voice, data and isochronous data packets using the current IEEE 802.11 PCF. Due to the limitations of PCF, it was extended in the new IEEE 802.11e into the more complex hybrid coordination function (HCF). We found that HCF has several performance issues that may affect its anticipated performance. We also compare the performance of the proposed scheme (EPCF) with the HCF function through simulation. Simulation results demonstrate an enhanced performance of our scheme over the legacy PCF and a comparable performance to the IEEE 802.11e HCF in terms of the average delay and system throughput. However, EPCF is much simpler than HCF, provides service differentiation on flow level, and is easy to embed in the current IEEE 802.11 standard. Moreover, we extend the proposed scheme (EPCF) to work in a multi-hop wireless ad hoc mode.

Java has become a popular programming language used on different platforms, ranging from embedded systems to powerful servers. Since the memory management is one of the most time-consuming parts within Java virtual machine (JVM), various techniques have been developed to boost its performance. However, the JVM memory management still does not scale very well, especially for multithreaded server applications. In this paper, we study different aspects of JVM object allocation from thread's perspective, using the trace data we collected from Sun JDK 1.3.1. Additionally, we construct a heap simulator to study the potential memory contentions among different threads. The simulation results show that dividing heap into different subheaps is very effective in alleviating the memory contentions. The results imply the potential benefits of using subheaps in improving the Java memory management performance.

Previous research efforts have shown that pretenuring can potentially reduce the copying cost by creating long lived objects into the mature memory regions directly. To date, researchers often employ profiling and static analysis to accurately select the objects that should be pretenured. However, little research efforts have been spent on dynamic approaches for pretenuring objects. In this paper, we propose a novel approach that dynamically predicts object lifespan to assist with pretenuring selection. The proposed scheme performs dynamic pretenuring selection based on a feedback mechanism that records lifespan of objects from each class during garbage collection invocations. This information is then used to pretenure objects in subsequent allocation requests. We experiment with two approaches, jumpstart feedback and continuous feedback, to collect tenuring information. The experimental results of selected benchmark programs show that our schemes can improve the garbage collection time of IBM's Jikes RVM by up to 37%, and improve the overall execution time by up to 28%.

Supporting quality of service (QoS) in wireless networks is a challenging problem. The point coordination function (PCK) in IEEE 802.11 LAN standard was developed to support the transmission of real-time data. However, PCF was not able to meet the requirements of real-time data. Therefore, several IEEE 802.11e drafts, whose main task is to support QoS, were released after IEEE 802.11. The polling scheme of PCF is extended in IEEE 802.11e into the more complex hybrid coordination function (HCK). We found that HCK has several performance issues that may affect its anticipated performance. In this paper, we address these issues and propose a QoS enhancement over PCK, called enhanced PCF (EPCF) that enables wireless LAN to send a combination of voice, data and isochronous data packets using the current IEEE 802.11 PCF. We compared the performance of the proposed model (EPCF) with the HCF function of the IEEE 802.11e through simulation and we extend the proposed model (EPCF) to work in a multi-hop wireless ad hoc mode. The simulation results demonstrated an enhanced performance of our scheme over the legacy PCK and a comparable performance to the IEEE 802.11e HCF in terms of average delay and system throughput. However, EPCF is much simpler than HCK, provides service differentiation on flow level, and can be incorporated easily into the current IEEE 802.11 standard.

Dynamic memory management can make up to 30% of total program execution time. Object oriented languages like C++ allocate and free dynamic memory prolifically. Since computer memory is a limited resource its efficient utilization is required to minimize wastage and keep costs down. Memory management algorithms such as best fit seem to perform most efficiently in terms of space cost while simple segregated storage seems to minimize the time cost. There is a trade-off between time and space costs. We have developed a new general purpose hybrid algorithm that shows excellent performance with respect to both time and space in comparison to the Doug Lea version 2.7.0 dynamic memory allocator.

In the IEEE 802.11 WLAN standard, a positive acknowledgement informs the sender of successful arrivals of data frames. However unacknowledged frames could result from either unsuccessful delivery of data frames or positive ACK frame losses. Therefore the sender will simply retransmit the unacknowledged data frame, which may cause redundant retransmission in case of positive ACK frame losses. In this paper we propose an enhanced retransmission scheme that we call Dynamically Adaptive Retransmission (DAR), which uses modified RTS and CTS frames containing additional information on transmission status of unacknowledged frames. Based on that information, the sender is able to dynamically determine whether to retransmit or not. Experiments and analysis show that our proposed scheme efficiently decreases redundant retransmission by clearly differentiating the reasons for frame loss.

Java possesses many advantages for embedded system development, including fast product deployment, portability, security, and a small memory footprint. As Java makes inroads into the market for embedded systems, much effort is being invested in designing real-time garbage collectors. The proposed garbage-collected memory module, a bitmap-based processor with standard DRAM cells is introduced to improve the performance and predictability of dynamic memory management functions that include allocation, reference counting, and garbage collection. As a result, memory allocation can be done in constant time and sweeping can be performed in parallel by multiple modules. Thus, constant time sweeping is also achieved regardless of heap size. This is a major departure from the software counterparts where sweeping time depends largely on the size of the heap. In addition, the proposed design also supports limited-field reference counting, which has the advantage of distributing the processing cost throughout the execution. However, this cost can be quite large and results in higher power consumption due to frequent memory accesses and the complexity of the main processor. By doing reference counting operation in a coprocessor, the processing is done outside of the main processor. Moreover, the hardware cost of the proposed design is very modest (about 8000 gates). Our study has shown that 3-bit reference counting can eliminate the need to invoke the garbage collector in all tested applications. Moreover, it also reduces the amount of memory usage by 77 percent.

The Active Memory System - a garbage collected memory module - was introduced as a way to provide hardware support for garbage collection in embedded systems. The major component in the design was the Active Memory Processor (AMP) that utilized a set of bit-maps and a combinational circuit to perform mark-sweep garbage collection. The design can achieve constant time for both allocation and sweeping. In this paper two enhancements are made to the design of AMP so that it can perform one-bit reference counting that postpones the need to perform garbage collection. Moreover, a caching mechanism is also introduced to reduce the hardware cost of the design. The experimental results show that the proposed modification can reduce the number of garbage collection invocations by 76%. The speed-up in marking time can be as much as 5.81. With the caching mechanism, the hardware cost can be as small as 27 K gates and 6 KB of SRAM.

One major problem of using Java in real-time and embedded devices is the non-deterministic turnaround time of dynamic memory management systems (memory allocation and garbage collection). For the allocation, the nondeterminism is often contributed by the time to perform searching, splitting, and coalescing. For the garbage collection, the turnaround time is usually determined by the size of the heap, the number of live objects, the number of object collected, and the amount of garbage collected Even with the current state-of-the-art garbage collectors (generational and incremental schemes), they may or may not guarantee the worst case latency. Moreover such schemes often prolong overall garbage collection time. In this paper, the performance analysis of the proposed Active Memory Module (AMM) for embedded systems is presented Unlike the software counterparts, the AMM can perform a memory allocation in a predictable and hounded fashion (14 cycles). Moreover it can also yield a bounded sweeping time regardless of the number of live objects or heap size. By utilizing the proposed system, the overall speed-up can be as high as 23% over the JDK 1.2.2 running in classic mode.

Due to the increasing popularity of java in clienthemer environments, most of today's server applications are multithreaded. Thus, research focusing on the performance analysis of multi-threaded environments has become increasingly important. Since per-thread information can be crucial in such analysis, measuring tools are needed to provide perthread information that may include cycle-based timers and filters to eliminate tracing overhead. In this papel; a Cycle Accurate Thread Timing for Linun Environment (CAlTLE) is presented. This approach provides a cycle-accurate timer with functions to filter out tracing overhead by coordinating efforts from both kernel and user applications. In this scheme, the kernel keeps track of accurate thread timing, while applications inform the kernel which part of the execution is to be measured. To demonstrate the tool$functionality, two case studies are provided, which include measuring latencies incurred by malloc calls and monitoring potential memory heap contention in multithreaded-multiprocessor environments.

Recent advances in software engineering, such as graphical user interfaces and object-oriented programming, have caused applications to become more memory intensive. These applications tend to allocate dynamic memory prolifically. Moreover, automatic dynamic memory reclamation (garbage collection, GC) has become a popular feature in modern programming languages. As a result, the time consumed by dynamic storage management can be up to one-third of the program execution time. This illustrates the need for a high-performance memory management scheme. This paper presents a top-level design and evaluation of the proposed instruction extensions to facilitate heap management.

In order to improve the execution speed of Java in hardware, a new advanced instruction folding technique has been developed. In this paper an instruction folding scheme based on an advanced Producer, Operator and Consumer (POC) model is proposed and demonstrates improvement in bytecode execution over the existing techniques. The proposed POC model is able to detect and fold all possible instruction sequence types dynamically in hardware, including a sequence that is separated by other bytecode instructions. SPEC JMV98 benchmark results show that the proposed POC model-based folder can save more than 90% of folding operations. In this research, the proposed instruction folding technique can eliminate most of the stack operations and the use of a physical operand stack, and can thereby achieve the performance of high-end RISC processors.

In the last thirty years, several dynamic memory management schemes have been proposed. Such schemes include first fit, best fit, segregated fit, and buddy systems. Because the performance (speed and memory utilization) of each scheme differs, software engineers often face difficult choices in selecting the most suitable approach for their applications. In this paper, a quantitative simulator for dynamic memory management and memory tracing techniques are presented. This simulator receives dynamic memory management traces and performs allocations according to schemes (first fit, best fit, buddy systems, and segregated fit) defined by the user. At the end of each simulation run, different performance metrics are reported to the users. By using this approach, software engineers can evaluate system performance and decide which algorithm is the most suitable for their applications.

Recently, most research efforts on garbage collection have concentrated on reducing pause times. However, very little effort has been spent on the study of garbage collection efficiency, especially generational garbage collection which was introduced as a way to reduce garbage collection pause times. In this paper a detailed study of garbage collection efficiency in generational schemes is presented. The study provides a mathematical model for the efficiency of generation garbage collection. Additionally, important issues such as write-barrier overhead, pause times, residency, and heap size are also addressed. We find that generational garbage collection often has lower garbage collection efficiency than other approaches (e.g. mark-sweep, copying) due to a smaller collected area and write-barrier overhead.

Object-oriented programming languages tend to allocate and deallocate blocks of memory very frequently. The growing popularity of these languages increases the importance of high-performance memory allocation. For speed and simplicity in memory allocation, the buddy system has been the method of choice for nearly three decades. A software realization incurs the overhead of internal fragmentation and of memory traffic due to splitting and coalescing memory blocks. This paper presents a simple hardware design for buddy-system allocation that takes advantage of the speed of a pure combinational-logic implementation. Two binary trees formed by anding and oring propagate information about the allocation status of blocks and subblocks. They implement a nonbacktracking search for the address of the first free block that is large enough to satisfy a request. Although the buddy system may allocate a block that is much larger than the requested size, the logic that finds a free block can be augmented by a "bit-flipper" to relinquish the unused portion at the end of the block. This effectively eliminates internal fragmentation. Simulation results show that the buddy system modified in this way uses less memory in most, though not all, programs than the unmodified buddy. Hence, the hardware buddy-system allocator is faster and uses memory more efficiently than the standard software approach.

Object-oriented systems exhibit a very high rate of object creation, but most of the objects are short-lived. As a result, memory management overhead is significant. The paper evaluates an application-specific coprocessor architecture to speed up object creation and memory reclamation in object-oriented systems. The architecture supports a bit-vector approach to dynamic storage allocation and liberation. Newly created objects reside in a cache which is reference counted. The paper presents measurements of the performance of this coprocessor design. Simulation results show that 50% to 70% of objects die before they age out of the cache, greatly reducing the number of references to main memory. Overall, more than 60% of memory traffic is saved by the proposed scheme, and the interval between main-memory garbage collections is extended by more than 60%.>

Object-oriented systems exhibit a very high rate of object creation, but most objects are short-lived. As a result, memory-management overhead is significant. An application-specific coprocessor architecture to speed up object creation and memory reclamation in object-oriented systems is described. The architecture supports a bit-vector approach to dynamic storage allocation and liberation. Novel created objects reside in a cache that is reference counted. Most objects are expected to die before they age out of the cache, drastically reducing the number of references to main memory. Many existing computer architectures would require only minor compiler modification to incorporate and benefit from this coprocessor.>

In application-specific integrated circuit (ASIC) testing, the time consumed by DC parametric tests is much greater than that needed for functional tests, taking a significant part of the total testing time. A study on how to optimize DC parametric testing is presented. An optimal algorithm for selecting preconditioning addresses which results in an efficient use of automatic test equipment (ATE) is introduced. The idea is to minimize the total time spent using the parametric measurement unit, i.e., to do parallel measurement as much as possible. Benchmark results comparing this approach with a traditional method are presented. In addition, several key factors which affect the actual testing efficiency are discussed. Through the power of ATPG (automatic test pattern generation), not only is the speed of generating test programs increased, but the quality of the programs generated is improved.>

A modified timing set priority strategy and a lookahead register-replacement algorithm is presented for automatic test program generation. These methodologies are introduced to satisfy the testing requirements in complex VLSI chips, and are made possible by the increasing timing flexibility of modern ATE (automatic test equipment). It is shown that by applying optimization algorithms in allocating timing resources and a published timing set determination algorithm, that testing capability of modern ATE can be greatly extended. Further enhancements to these approaches are contingent on further improvements from ATE manufacturers in the number and flexibility of timing resources. However, the development of test programs will be greatly aided by this enhanced hardware.>