Seng Wai Loke

Mobile cloud computing

Despite increasing usage of mobile computing, exploiting its full potential is difficult due to its inherent problems such as resource scarcity, frequent disconnections, and mobility. Mobile cloud computing can address these problems by executing mobile applications on resource providers external to the mobile device. In this paper, we provide an extensive survey of mobile cloud computing research, while highlighting the specific concerns in mobile cloud computing. We present a taxonomy based on the key issues in this area, and discuss the different approaches taken to tackle these issues. We conclude the paper with a critical analysis of challenges that have not yet been fully met, and highlight directions for future work. Highlights? Problems in mobile computing can be solved when combined with cloud computing. ? Three types of mobile clouds are; offloading to a remote server, to a local cloudlet, and sharing work in a mobile p2p network. ? Existing research on mobile cloud computing is presented as a taxonomy of issues. ? We identify issues in operational, end user, and service levels, and in security and context awareness.

Verity: a QoS metric for selecting Web services and providers

With the proliferation of Web services, quality of service serves as a benchmark to differentiate the services and their providers. As of today, a wide spectrum of attributes have been identified to account for the quality of a service like availability, reliability, servability, performance, reputation and so on. Reputation has been measured as an average user rating and we argue that the user perception alone is not sufficient to indicate the reputation. It is necessary to measure how trustworthy the provider has been in complying with the agreed levels in the SLA (service level agreement). To quantify the consistency in compliance levels, we introduce a new QoS attribute termed verity and propose an architecture to quantify it. We argue that verity should be taken into account for a quality driven selection and composition of Web services. Reputation, when expressed as a vector of user rating, compliance and verity is a more intuitive indicator of the providers trustworthiness.

Towards a theory of context spaces

We propose initial work on a conceptual framework for context-aware systems, towards a general context model to aid thinking, describing, manipulating and utilizing context. Much work on context-aware computing have utilized context in specialized applications, with application-specific software (and perhaps hardware) and specialized context representation and manipulation. Context-aware systems are becoming increasingly important, and emerging research has begun to look at context-aware systems more generally, independently of specific applications, including context middleware and toolkits and ontologies for describing context.

Representing and reasoning with situations for context-aware pervasive computing: a logic programming perspective

Context-aware pervasive systems are emerging as an important class of applications. Such systems can respond intelligently to contextual information about the physical world acquired via sensors and information about the computational environment. A declarative approach to building context-aware pervasive systems is presented, and the notion of the situation program is introduced, which highlights the primacy of the situation abstraction for building context-aware pervasive systems. There is also a demonstration of how to manipulate situation programs using meta-programming within an extension of the Prolog logic programming language which is called LogicCAP. Such meta-reasoning enables complex situations to be described in terms of other situations. Furthermore, a discussion is given on how the design of situation programs can affect the properties of a context-aware system. The approach encourages a high-level of abstraction for representing and reasoning with situations, and supports building context-aware systems incrementally by providing modularity and separation of concerns.

Reputation = f(user ranking, compliance, verity)

The selection of Web services is typically based on both functional and nonfunctional attributes of the service, such as the quality of service (QoS) levels. Reputation, a widely acknowledged nonfunctional QoS attribute is currently expressed as the average of user ratings given to the service. However, this expression confines reputation to the subjective perception of the end user and is limited by the lack of an objective representation of performance history. In this paper, we address the need for a reputation mechanism that couples the subjective perception of the end user with the objective view of performance history. To represent performance history, we propose a novel QoS metric termed verity. Verity measures the degree of consistency exhibited by the service provider in delivering the quality levels laid out in the service contract, over a range of previous transactions. We express reputation as a composition of user rating, the compliance levels exhibited by the provider and the verity value. We contend that this reputation expression is a more viable attribute of quality than user rating alone.

An approach to data fusion for context awareness

We propose and develop an approach modeled with multi-attribute utility theory for sensor fusion in context-aware environments. Our approach is distinguished from existing general purpose fusion techniques by a number of factors including a general underlying context model it is built upon and a set of heuristics it covers. The technique is developed for context-aware applications and we argue that it provides various advantages for data fusion in context-aware scenarios. We experimentally evaluate our approach with actual use cases using real sensors.

Using On-the-Move Mining for Mobile Crowdsensing

In this paper, we propose and develop a platform to support data collection for mobile crowdsensing from mobile device sensors that is under-pinned by real-time mobile data stream mining. We experimentally show that mobile data mining provides an efficient and scalable approach for data collection for mobile crowdsensing. Our approach results in reducing the amount of data sent, as well as the energy usage on the mobile phone, while providing comparable levels of accuracy to traditional models of intermittent/continuous sensing and sending. We have implemented our Context-Aware Real-time Open Mobile Miner (CAROMM) to facilitate data collection from mobile users for crowdsensing applications. CAROMM also collects and correlates this real-time sensory information with social media data from both Twitter and Facebook. CAROMM supports delivering real-time information to mobile users for queries that pertain to specific locations of interest. We have evaluated our framework by collecting real-time data over a period of days from mobile users and experimentally demonstrated that mobile data mining is an effective and efficient strategy for mobile crowdsensing.

Multiple-Agent Perspectives in Reasoning About Situations for Context-Aware Pervasive Computing Systems

In open heterogeneous context-aware pervasive computing systems, suitable context models and reasoning approaches are necessary to enable collaboration and distributed reasoning among agents. This paper proposes, develops, and demonstrates the following: 1) a novel context model and reasoning approach developed with concepts from the state-space model, which describes context and situations as geometrical structures in a multidimensional space; and 2) a context algebra based on the model, which enables distributed reasoning by merging and partitioning context models that represent different perspectives of computing entities over the object of reasoning. We show how merging and reconciling different points of view over context enhances the outcomes of reasoning about the context. We develop and evaluate our proposed algebraic operators and reasoning approaches with cases using real sensors and with simulations. We embed agents and mobile agents with these modeling and reasoning capabilities, thus facilitating context-aware and adaptive mobile agents operating in open pervasive environments.

A fuzzy model for reasoning about reputation in web services

Reputation systems are typically based on ratings given by the users. When there are no mechanisms in place to detect collusion and deception, combining user testimonies as such to form a providers reputation may not give an accurate assessment, especially if the context of the ratings is not known. Moreover, such systems are vulnerable to manipulations by malicious users. Hence it becomes essential to establish the validity of the ratings prior to using them in formulating reputation based on such ratings. It is important to identify the rationale behind the ratings so that similar ratings (or ratings pertaining to a context) can be aggregated to obtain a reputation value meaningful in that context. We propose a fuzzy approach to analyze user rating behavior to infer the rationale for ratings in a web services environment. This inference of rationale facilitates the system to validate ratings, detect deception and collusion, identify user preferences and provide recommendations to users.

Context-aware artifacts: two development approaches

To create context-aware artifacts, developers can choose a self-supported or infrastructure-based approach. This article surveys these approaches and their implementations in various devices, also considering associated challenges and future research directions. Creating context awareness generally follows two main approaches. In self-supported context awareness, designers build the ability to perceive context, reason with it, and act accordingly into the device or its dedicated hardware support. In infrastructure-supported context awareness, designers obtain context-aware capabilities by harnessing a hardware and software infrastructure external to and associated with the devices space. Such an infrastructure might implement the context-aware behavior of specific objects and devices or act as a general context-aware infrastructure.