Software-oriented architecture (SOA) is a promising paradigm for efficiently leveraging the functionality of individual IoT devices to build IoT applications. However, deploying SOA for IoT data-gathering applications requires spatial context-awareness and the ability to aggregate similar available services, which presents a challenge. To address this challenge, this paper proposes a formulation for spatial context-aware service composition with a novel quantitative model for spatial context. We demonstrate that incorporating spatial context into service composition is an NP-Hard problem and model it as an integer linear program. We propose two heuristic approaches capable of producing near-optimal solutions in real-time. We implement a simulation of the composition problem to study the performance of our approaches. Our experimental results show that the proposed methods are scalable compared to the branch-and-cut algorithm. These results set a precedent against which future work on solutions to this problem can be compared.

Real-time artificial intelligence (AI) applications mapped on edge computing need to perform data capture, data processing/intelligence extraction, and device actuation within some given time bounds. Synchronization across devices is an important problem that needs to be solved at different stages of an AI application. Synchronized data capture reduces the amount of time required in preprocessing data (data aggregation, data cleaning, missing data handling, etc). In the data processing phase, synchronization is key in ensuring convergence, accuracy, and speed of the distributed training process across multiple edge devices. The actuation phase in some cases requires certain actions to be performed at the same time on different devices. In this article, we develop a fast edge-based synchronization scheme that can time-align the execution of input-output tasks as well compute tasks. The primary idea of the fast synchronizer is to cluster the devices into groups that remain closely synchronized in their task executions and statically determine synchronization points using a game-theoretic solver. The cluster of devices uses a late notification protocol to select the best point among the precomputed synchronization points to reach a time-aligned task execution as quickly as possible. We evaluate the performance of our synchronization scheme using trace-driven simulations, and we compare the performance with existing distributed synchronization schemes for real-time AI application tasks. We implement our synchronization scheme and compare its testing accuracy and training time with other parameter server synchronization frameworks.

From road condition information to locating parking spaces, there exists a growing need for real time sensing information. This can be especially important to support the smart applications of the future. Vehicles are an ideal platform for such data gathering given its mobility, sensors and computing resources. In this paper, we focus on small scale sensing tasks, tasks which are highly personalized and are unlikely to be shared. Thus, sensing information is collected on demand rather than continuously and at all time. We refer to this as personalized vehicular crowdsensing. To achieve personalized sensing coverage, we must resolve the problem of vehicular participant selection. We propose two solutions based on push and pull for vehicular crowdsensing. We aim to maximize sensing coverage, improve load balance, error tolerance, and minimize costs. We compare our results with existing approaches and show superior performances for all metrics.

This paper presents a fog-resident controller architecture for synchronizing the operations of large collections of Internet of Things (IoT), such as drones, Internet of Vehicles, etc. Synchronization in IoT is grouped into different classes, use cases identified and multipoint synchronous scheduling algorithms are developed to schedule tasks with varying timing requirements; strict (synchronous) and relaxed (asynchronous and local) onto a bunch of worker nodes that are coordinated by a fog resident controller in the presence of disconnections and worker failures. The algorithms use time-based or component-based redundancy to cope with failures and embed a publish-subscribe message update scheme to reduce the message overhead at the controller as the number of workers increase. The performance of the algorithms are evaluated using trace-driven experiments and practicability is shown by implementing the time-based redundancy synchronous scheduling algorithm in JAMScript-a polyglot programming platform for Cloud of Things and report initial findings.

Machines controlled by cloud or edge resident coordinators are becoming an important trend for creating smart systems. The cloud provides a global perspective while the edge provides low latency and localized service to the machines. Coordinating these machines to work collectively to solve problems with strict timing requirements in the presence of disconnections is a challenge. Clock synchronization is necessary but not sufficient for such a system. We introduce a three-tiered hierarchical system model with the cloud at the top, devices at the leaf and fog in between. Tasks are differentiated into three categories: synchronous, asynchronous and local depending on the execution mode and requirements. We propose three scheduling algorithms; static, dynamic and batch synchronization scheduling algorithms, and evaluate the performance of the algorithms through simulations under varying operating conditions. Finally, we implement the dynamic synchronous scheduling algorithm on a polyglot programming platform for Cloud of Things to show its practicability and report initial findings.

Growing mobile data demand is pushing the limits of cellular networks. With the rise of smart cars, vehicular users are poised to become a substantial portion of this demand. To meet rising requests, road side units (RSUs) will play a key role in offloading traffic from cellular infrastructure. In this paper, we propose a deduplication hybrid cellular and RSU based approach where a vehicle can exploit the advantages of both cellular and RSU networks. Popular data items are broadcasted from RSU networks while unpopular items can be obtained from LTE. We propose three algorithms which takes into account mobility of vehicles to create a global broadcast schedule. We compare and evaluate performance using cellular usage and download speed as metrics. We utilized TAPAS Cologne, a large scale simulated vehicular mobility dataset. We find that our proposed algorithms reduce cellular usage and improves download speed.

We describe Wireless GINI, a wireless platform for hosting virtual networks. Wireless GINI Is Not Internet (GINI) allows each virtual network to define its own topology and network configuration, while amortizing costs by sharing the physical infrastructure. The platform also creates mechanisms to readily integrate physical wireless devices into a deployed network. The platform supports a diverse combination of network elements that are all integrated into one custom Internet. The platform seamlessly integrates process-emulated components running on the user machine, wireless mesh overlays deployed on the wireless platform, and generic wireless devices connected to the user's custom network. Wireless GINI provides a user-friendly interface that makes the physical setup process completely transparent to the user. A centralized server is used to provide this transparency, handle user requests, and automatically provision the shared physical infrastructure. We describe the design and implementation of Wireless GINI. Moreover, we evaluate the performance of the platform and suggest several educational experiments that can be conducted on this new platform. A detailed survey of the existing toolkits and platforms and a comparison with Wireless GINI are also provided. Copyright © 2016 John Wiley & Sons, Ltd.

This paper presents the Physical Object Model (POM), a data model for smart environments, that creates a smart spaces hierarchy to oversee spaces, things and users and the interactions between them. To demonstrate our design, we developed a POM server and a POM client. The server is in charge of creating such a smart space that can register, monitor and control many clients. The clients, running on smart things, use a pre-defined API to communicate with the server to query surrounding devices or to advertise services they provide to the space. To record clients' impact and activities inside the server, we also developed a logging component that saves every change to the space to a database. Using the logging component, we provided a temporal API to request the state of space at a given point in time in the past. To evaluate our system performance against real-life scenarios, we made a simulator that takes profile inputs and performs tests against the POM server.

In this paper, we propose a new model for smart spaces, things, and people contained within them. Our model is inspired by the document object model (DOM) that underpins the dynamic web. Like the DOM that enables programs attached to elements of a web document to manipulate portions of the document, we develop a physical object model (POM) that enables programs attached to things to manipulate other things or the smart space containing the thing. We describe the important pieces of the model and explain how it could enable interactions in smart spaces. Two important concerns with smart spaces are security and privacy. We discuss how POM facilitates the inclusion of policies to manage these concerns at the model level.

SpaceOS is a system software stack for smart computing environments. The major objective of SpaceOS is to make it easier to create applications and deploy them in smart computing environments. It provides several core services for managing smart devices in physical spaces and letting them interact with cloud backends and web-based users. SpaceOS also supports a programming language called Jade that leverages the facilities provided by the system software stack to provide a simpler programming model for the things.

This paper presents a chaotic map based image encoding scheme that allows pixel-level operations to be applied on an image while it remains encoded. Using this image encoding scheme, sensitive images can be manipulated by untrusted public cloud servers. Before the images are uploaded onto the public cloud servers they are encoded using user supplied keys by trusted-cloud servers. We fully describe the chaotic map based image encoding algorithm and perform a detailed security analysis of it. Our security analysis demonstrates that the encoding scheme is robust against a wide variety of threats. Additionally, we implemented many pixel- level filtering operations on images over the proposed encoding and found that the overhead is less than 20%.

Recently, searching over encrypted data has become a hot research topic. Basic idea of privacy enhanced search is to generate an intermediate representation of the original text data and use it to perform the search. Prior research has used hash maps, tries, and other data structures to create the intermediate representation. We use a texture scheme for representing the characters. To enhance the privacy the textures are split and noise is added to each of the portions such that all portions of a texture needs to be collected to recover the original data in an unambiguous manner. One of the key advantages of our scheme is the ability to implement most of the search schemes (e.g., wildcard searches) that are performed with plain text searches. This paper fully describes our data representation scheme and presents the experimental data we gathered by implementing the scheme in a server cluster.

Template matching is a fundamental building block for image search operations. In this paper, we present a scheme that allows privacy-preserving template matching operations on images that are stored on clouds. Our scheme uses “ambient image data” (images that are found in social media sites such as Flickr) as well as a privacy-preserving encoding technique to encode a given image before it is stored in a cloud. We show a particular encoding strategy that allows template matching to take place in the cloud while not revealing any information about the image or queried template to the cloud. A simplified prototype of the image processing system was implemented and the experimental results are presented in this paper. Our prototype shows the feasibility of performing privacy-aware template matching on encoded images.

Cloud computing is ideally suited for hosting applications with large images because clouds can perform high performance computations and communications at very low upfront cost. While the benefits of clouds are compelling for image processing applications, security and privacy issues in using clouds are creating major problems. Because clouds are not passive storage devices, it is important to avoid encryption schemes that secure the data while preventing efficient processing of the data by the clouds. Therefore, this paper, is investigating approaches where segmentation is used to split the image data across multiple clouds. We show that segmentation, although a simple idea, can bring several benefits when deployed in cloud computing. A prototype of the system has been developed and initial performance results are reported here.

The amount of data stored in social media sites is humongous. It has very large in numbers and very diverse in variety. Researchers have already started using this information in various information security applications such as censorship resistance. When image data from social media sites are used in computational applications for hiding data or obfuscating data, it is important to retrieve the particular data used in the encoding process to retrieve the original data. Therefore, the availability of the image data is an important concern. In this paper, we investigated the availabilities of the images on Flickr. We examined nearly one million images hosted on Flickr and measured their availability in a two year period. We used the EXIF parameters normally generated by the cameras to further categorize the availabilities.

Social networks are emerging as the arteries for information flow on the web. In a previous paper, we introduced a new community-centric approach for information flow control for the social web. This paper introduces two key improvements to the previously introduced mechanisms. The first improvement is the ability to model user heterogeneity with regard to information relaying on the social web. The second improvement is the ability to reduce information flow to a specific user (i.e., block information flowing to a specific user). We evaluate the algorithmic ideas using traces of interactions obtained from Facebook and Flickr. Our evaluations indicate that the algorithmic ideas developed by us are useful in controlling the information flow in the social web.

In this paper, we present a novel community-centric confidentiality control mechanism for information flow management on the web. We use a Monte Carlo based algorithm to determine the potential spread of a shared data object and inform the user of risk of information leakage associated with the different sharing decisions she can make in a social network. Using the information provided by our algorithm a user can curtail sharing decisions to reduce the risk of information leakage. Alternatively, our algorithm can provide input to a fully- or semi- automatic sharing decision maker that determines the outcomes of sharing requests. We used datasets from Facebook and Flickr to evaluate the performance of the proposed algorithms under different sharing conditions. The simulation results indicate that information sharing can be effectively controlled by our algorithm.

We present an accountability framework for the Internet which ties a user''s action to her identity on an online social network. The framework is optional in that users do not need to be accountable at all times, but various web services can force accountability on the part of their users by only allowing accountable users access. Our design is general enough that higher level applications can place additional policies/restrictions on the basic accountability provided. In this paper, we introduce the design, discuss how various applications can be mapped onto our framework, and provide performance numbers from an experimental prototype.

Social interaction is already a proven component of informal identification: humans are naturally skilled at recognizing other people and are unlikely to be duped by impersonation. Based on this premise, a fourth-factor, someone-you-know, has already been proposed as an emergency authentication method. This paper explores leveraging a user's preexisting social actions as a primary authentication tool, one that operates transparently and automatically without explicit user guidance. Specifically, we describe the feasibility of capturing a user's local social context using short range wireless devices and evaluate the uniqueness of that context in comparison to that of possible aggressors.

Phishing is a major problem on the Internet. The cornerstone of anti-phishing is detecting whether a given site is good or bad. Most of the approaches for anti-phishing rely on looking up centrally maintained repositories. In this paper, we present a decentralized framework called CASTLE that allows

In this paper, we investigate the problem of online data sharing on social networks from a game theoretic framework. We introduce blacklisting as trigger strategy to elicit cooperation among the players of a noncooperative sharing game. Using game theoretic analysis, we show the existence of an equilibrium in which the sharing conditions are honored when the involved players employ blacklisting strategies.

GINI (GINI Is Not Internet) is an open-source toolkit for creating virtual micro Internets for teaching and learning computer networking. It provides lightweight virtual elements for machines, routers, switches, and wireless devices that can be interconnected to create virtual networks. The virtual elements run as unprivileged user-level processes. All processes implementing a virtual network can run within a single machine or can be distributed across a set of machines. The GINI provides a user-friendly GUI-based tool for designing, starting, inspecting, and stopping virtual network topologies. This paper describes the different components of GINI, briefly discusses ways of using the toolkit in a computer networking course, and reports on user feedback on an early (incomplete) version of the toolkit.

This paper presents a case for exploiting the synergy of dedicated and opportunistic network resources in a distributed hosting platform for data stream processing applications. Our previous studies have demonstrated the benefits of combining dedicated reliable resources with opportunistic resources in case of high-throughput computing applications, where timely allocation of the processing units is the primary concern. Since distributed stream processing applications demand large volume of data transmission between the processing sites at a consistent rate, adequate control over the network resources is important here to assure a steady flow of processing. In this paper, we propose a system model for the hybrid hosting platform where stream processing servers installed at distributed sites are interconnected with a combination of dedicated links and public Internet. Decentralized algorithms have been developed for allocation of the two classes of network resources among the competing tasks with an objective towards higher task throughput and better utilization of expensive dedicated resources. Results from extensive simulation study show that with proper management, systems exploiting the synergy of dedicated and opportunistic resources yield considerably higher task throughput and thus, higher return on investment over the systems solely using expensive dedicated resources.

We present a personal data access control (PDAC) scheme inspired by protection schemes used in communities for sharing valuable commodities. We assume PDAC users are members of an online social network such as facebook.com. PDAC computes a ldquotrusted distancerdquo measure between users that is composed of the hop distance on the social network and an affine distance derived from experiential data. The trusted distance classifies users into three zones: acceptance, attestation, and rejection. User requests falling in the acceptance zone are accepted immediately while the requests in the rejection zone are rejected outright. Requests in the attestation zone need additional authorization to gain access. PDAC also tracks reposts to minimize the spread of data beyond the limits set by the data originator. PDAC was implemented on a social network emulator to demonstrate its viability. The performance of certain PDAC functions were examined using simulations driven by portions of social graphs obtained from myspace.com.

Epidemic protocols such as gossip have proven to have many desirable properties for information sharing. However, trust is one of the issues that is yet to be examined with respect to these protocols. In this paper, we present a trusted gossip protocol that uses trust estimates to impede spreading of rumors with reasonable message and processing overheads. We use traces collected from known social networks to estimate the performance of trusted gossip.

This paper presents a new bandwidth inference mechanism that can be used to predict bandwidth across two nodes on the Internet. We used simulation and actual implementation on PlanetLab to compare the performance of the proposed mechanism against an existing approach. The results indicate that our approach is lightweight and yields better performance.

In a recent study, we proposed a trusted gossip protocol for rumor resistant information sharing in peer-to-peer networks. While trust aware gossiping significantly reduced the rumor spread on the network, we observed that the random message spraying in trusted gossip creates too many redundant messages increasing the message overhead and error rate. In this paper, we propose a message targeting scheme that can significantly improve the performance of the trusted gossip. Our targeting scheme can be easily implemented in a social network setting. We performed large-scale simulations using traces collected from the Flickr social network and other data sets to estimate the performance of targeting in trusted gossip. Our experiments show that significant performance gains can be achieved.

Trust and security are two major issues in large distributed systems that are highly inter-dependent such that it is hard to bootstrap one without the other. In distributed systems that are built using private resources, security is bootstrapped using off-line trust relationships. Mandating off-line trust relations, however, has the undesirable effect of limiting the membership of distributed systems. Therefore, public-resource based systems need online trust modeling. This paper presents a two-layered framework that considers trust, security and incentives in an integrated fashion. Our framework is characterized by (a) a community-based process that is essentially decentralized for evaluating and assigning trust for peers (b) policy autonomy in administering incentives and trust, and (c) lowers the join time trust requirements for peers. Results from early prototyping and simulations show the proposed framework can be implemented with acceptable overheads and is capable of evolving trust and applying it for resource management.

Provisioning quality of service (QoS) in Web servers has gained immense importance because Web servers are a major part of the Internet. To deliver the pledged QoS, Web service providers need control over the allocation of the resources in their Web servers. Control is also necessary for reaching the optimal resource allocation through proper service differentiation. In this paper, we propose and investigate an analytic approach that enables the service providers to deploy a differentiated service policy that offers this control. The proposed service policy is configurable by tunable control parameters. We devise the relationships between the performance measures and these parameters by adopting a unique queuing theoretic approach. Once these relationships are established, we describe how these parameters can be set to their most appropriate values depending on the objectives of the service providers. We illustrate the usefulness of our approach by conducting the analysis on a real Web trace.

Summary form only given. We present a trust brokering system that operates in a peer-to-peer manner. The network of trust brokers operate by providing peer reviews in the form of recommendations regarding potential resource targets. One of the distinguishing features of our work is that it separately models the accuracy and honesty concepts. By separately modeling these concepts, our model is able to significantly improve the performance. We apply the trust brokering system to a resource manager to illustrate its utility in a public-resource grid environment. The simulations performed to evaluate the trust-aware resource management strategies indicate that high levels of "robustness" can be attained by considering trust while allocating the resources.

One of the key issues with decentralized online multi-player games is ensuring fairness of game play despite different network latencies. One approach is to follow a "lock step" scheme that prevents any player from acting upon a message until that message is received by all the players. We present a fair synchronization protocol (FSP) that enforces fairness and is more efficient than a lock-step scheme. The basic FSP protocol is susceptible to cheating. Therefore, we add a cheat prevention mechanism as an enhancement. We implement the enhanced protocol, the cheat proof protocol (CPP), in a simple game environment and deploy it on the PlanetLab. The results indicate that the CPP is effective in enforcing fairness while preventing cheating.

This work describes a design for a quality of service aware public computing utility (PCU). The goal of the PCU is to utilize the idle capacity of the shared public resources and augment the capacity with dedicated resources as necessary, to provide high quality of service to the clients at the least cost. Our PCU design combines peer-to-peer (P2P) and grid computing ideas in a novel manner to construct a utility-based computing environment. In This work, we present the overall architecture and describe two major components: a P2P overlay substrate for connecting the resources in a global network and a community-based decentralized resource management system.

Resource management is a central part of a Grid computing system. In a large-scale wide-area system such as the Grid, security is a prime concern. One approach is to be conservative and implement techniques such as sandboxing, encryption, and other access control mechanisms on all elements of the Grid. However, the overhead caused by such a design may negate the advantages of Grid computing. This study examines the integration of the notion of "trust" into resource management such that the allocation process is aware of the security implications. We present a formal definition of trust and discuss a model for incorporating trust into Grid systems. As an example application of the ideas proposed, a resource management algorithm that incorporates trust is presented. The performance of the algorithm is examined via simulations.

Efficient schemes for deploying and maintaining wide-area services and accessing such services are essential if the Internet is to sustain itself as the communication medium for business critical applications. This paper presents the architecture of an overlay service network called the Invisible Network. It is based on a two-stage architecture. The first-stage is responsible for service addressing, discovery, and mobility. The second-stage is responsible for quality of service aware resource allocation for wide-area services. The two stages together provide a unique service network that makes the wide-area services "invisible" by hiding the locational and implementational details of the wide-area services.

This paper presents three algorithms for co-reserving resources. A novel mechanism for specifying QoS constraints with the requests is introduced. The performance of the algorithms are studied through extensive simulations. The results indicate that depending on the objective of optimization different algorithms should be used. One of the distinguishing features of our algorithms is that they operate in a pseudo-online mode (called the “batch” mode).

Grid computing systems that have been the focus of much research in recent years provide a virtual framework for controlled sharing of resources across institutional boundaries. Security is a major concern in any system that enables remote execution. Several techniques can be used for providing security in grid systems including sandboxing, encryption, and other access control and authentication mechanisms. The additional overhead caused by these mechanisms may negate the performance advantages gained by grid computing. Hence, we contend that it is essential for the scheduler to consider the security implications while performing resource allocations. In this paper, we present a trust model for grid systems and show how the model can be used to incorporate security implications into scheduling algorithms. Three scheduling heuristics that can be used in a grid system are modified to incorporate the trust notion and simulations are performed to evaluate the performance.

Abstract The resource management system is the central component of distributed network computing systems. There have been many projects focused on network computing that have designed and implemented resource management systems with a variety of architectures and services. In this paper, an abstract model and a comprehensive taxonomy for describing resource management architectures is developed. The taxonomy is used to identify approaches followed in the implementation of existing resource management systems for very large‐scale network computing systems known as Grids. The taxonomy and the survey results are used to identify architectural approaches and issues that have not been fully explored in the research. Copyright © 2001 John Wiley & Sons, Ltd.

A system can be considered as an infrastructure that allows location independent access to the resources and services that are provided by geographically distributed machines and networks. One fundamental operation necessary to support such a system is performance discovery in wide-area networks. Dissemination of resource status information is an important component of any performance discovery algorithm. The data dissemination for performancediscovery in systems has several distinct features and requirements that can be exploited to make the process efficient. Recently, several data dissemination algorithms based on a new concept called the Grid potential were proposed. This paper introduces a new algorithm for data dissemination and compares it with previously introduced schemes through simulation studies. The results indicate that the new algorithm further improves the performance of data dissemination.

Two new QoS-based unicast routing algorithms for high-speed internetworks are proposed as part of this work. Both algorithms, Limited-Scope Probing (LSP) and Limited-Scope Hop-by-Hop Probing (LSHP) use the route update messages generated by the Open Shortest-Path-First (OSPF) algorithm to create hybrid routing tables. The LSP and LSHP are semi-distributed routing algorithms. They assume that the network "state" information maintained by the OSPF is accurate in "close" vicinity of the maintaining node. Therefore, the LSP and LSHP propose to use source-based routing techniques to route along paths that are "centrally" computed by the node. The extent of the vicinity can be set by an appropriate parameter. Extensive simulations are performed to compare our algorithms with existing QoS-based unicast routing algorithms.

By considering electromagnetic scattering problems as examples, a study of the performance and scalability of the conjugate gradient squared (CGS) algorithm on two MIMD machines is presented. A modified CGS (MCGS) algorithm, where the synchronization overhead is effectively reduced by a factor of two, is proposed in this paper. This is achieved by changing the computation sequence in the CGS algorithm. Both experimental and theoretical analyses were performed to investigate the impact of this modification on the overall execution time.