Many academic and industrial research activities have recently recognized the relevance of expressive models and effective frameworks for highly scalable data processing, such as MapReduce. This paper presents the novel Quasit programming model and runtime framework for stream processing in datacenters, with its original capabilities of i) allowing developers to choose among a large set of quality policies to associate with their processing tasks in a fine-grained way, and ii) effectively managing processing execution depending on the associated quality indications. The paper describes the Quasit programming model, via the primary design/implementation choices made in the Quasit runtime framework (available for download from the project Web site) to achieve maximum scalability, flexibility, and reusability. The first experiences with our prototype and the reported experimental results show the feasibility of our approach and its good performance in terms of both limited overhead and horizontal scalability.
The increasingly use of wireless networks and mobile technologies has raised the desire not only to have a good quality access to the network, but also to seamlessly change the network when moving. Various handover algorithmus have been proposed to handle this situation. Unfortunately, many of these algorithms have been only evaluated in simulative environments using simplified models and network assumptions. They do not take the wide range of mobile devices with varying system parameters and capabilities into account which are offered on the market. For the practical deployment, handover algorithms are required which adapt to various device parameters and network characteristics. In this paper we present a fuzzy-based vertical handover decision algorithm which adjusts itself to the given device parameters and network capabilities. Starting with a discussion on the requirements to vertical handover, we present the algorithm and describe how it is activated during the various phases of the handover process. Thereafter we present several experiments which evaluate the accuracy of the handover decision, the quality of service guarantees for the application, and the resource consumption.
Today, handheld devices equipped with Wi-Fi interfaces are used intensively by a huge number of people every day. These devices can form intermittently connected mobile ad hoc networks spontaneously. These networks appear as a relevant solution to extend a pre-existing infrastructure-based network composed of several access points in view of providing nomadic people with application services in a wide area. In such hybrid networks, intermittent connections are prevalent, and end-to-end paths between clients and providers cannot be maintained all the time. Thus, the communications must be achieved following a ”store, carry and forward” principle. In this paper, we present a new soft handover mechanism dedicated to service delivery in such hybrid networks. This handover solution exploits several pieces of information, such as the message propagation time, the path stability, and the mobility degree of intermediate nodes in order to select the most appropriate access point(s) to forward a response to a given mobile client.
Spontaneous Multi-hop Networks (SMNs) are emerging as a novel networking and communication paradigm, strongly pushed by the widespread availability of smartphones equipped with heterogeneous wireless connectivity and powerful computing capabilities. SMN nodes can opportunistically exploit peer-to-peer contacts to seamlessly share resources/content in an impromptu and transient way. The paper presents a novel 3-layer modeling abstraction for multicast in SMNs, in order to characterize the different kinds of possible inter-node interaction based on different degrees of expressiveness and social-aware collaboration. In addition, we originally present the design and implementation of some novel semantic-based multicast mechanisms that efficiently target SMN nodes based on user interests and that are integrated into our SMN middleware solution. First preliminary results show the feasibility of the approach and its limited overhead.
This paper suggests an approach for formalizing Tuple Space based Mobile Middleware (TSMM) that contains a fully-decoupled reactive tuple space model as coordination medium. Formalization of TSMM is carried out using Mobile UNITY.
Cross-platform frameworks for mobile devices promise to facilitate the porting effort of applications between different smartphones. Our approach is to cross-compile Android applications to other platforms such as iOS or Windows Phone 7. Doing so requires to refactor the Android source code base in a platform-dependent and platform-independent part separated by a Common Platform API. This paper discusses the cross-compiling of Java-based Android applications and the design and implementation of the Common Platform API.
This article describes an easy, efficient way to manage context-aware applications with the help of metadata. We rely on CATS, our proposition for an application framework embedded on mobile devices. It is designed to host applications conforming to the SOA principles for achieving a flexible and dynamic architecture. Our framework provides non-functional capabilities for context management and for the adaptations required at context changes. In this article we focus on the use of iPOJO handlers and the advantages they bring to the OSGi technology.
This book constitutes the thoroughly refereed proceedings of the 5th International Conference on Mobile Wireless Middleware, Operating Systems, and Applications, Mobilware 2012, held in Berlin, Germany, in November 2012. The 18 revised full papers presented were carefully reviewed and selected from numerous contributions. The papers are organized in topical sections on Internet of things and mobile sensing, mobile middleware platforms, mobile networks, systems support for mobile applications, and context awareness.
Delivery of multimedia services over heterogeneous wireless networks is a challenging proposition because of the diverse characteristics of the underlying wireless technologies. The problem is compounded further by the availability of a wide range of end user devices such as desktops, notebooks, tablets and smartphones. To provide a consistent and uniform Quality of Experience to the end user, we propose a context-aware service adaptation framework in this paper. The basic idea is to monitor the user and network context and leverage this information for adapting services to match the device and network characteristics.
The growth of “smart” mobile devices, such as smartphones and tablets, has been exponential over the past few years. Such growth was mainly attributed to the development of mobile applications. To date, mobile applications have been increasingly used to improve our productivity and also to provide the entertainment contents. However, with a huge number of mobile applications that appear in the application stores; in particular those that provide similar functionalities, users are often confused with the selection of trustworthy and high quality mobile applications. At the current state, there is a limited research embarked to provide solutions for measuring the trustworthiness of mobile applications prior to download. Thus, the aims of this paper are to review the current research in this area and to discuss several issues in measuring the trustworthiness of mobile applications. In addition, this paper also proposes MobilTrust, a similarity trust measurement method to solve the identified issues.
The current trend in operator networks is towards the deployment of high capacity radio technologies such as LTE accommodating a high number of devices and their data traffic. However, the current network architecture was designed for a lower level of communication in which scalability was achieved through uniform operator control. Connectivity for each mobile device was handled in the same manner, no matter of its characteristics, network location or resources required resulting in a high overhead in supporting a large part of the subscribers. This paper introduces a new self-adaptation concept realized as a subscriber oriented management layer enabling the customization of the control procedures and resources reserved to the individual communication requirements for each device. The concept is exemplified for access network selection and core network path adaptation use cases, adapted for the 3GPP Evolved Packet Core architecture and evaluated through a testbed realization based on the Fraunhofer FOKUS OpenEPC toolkit.
Internet of Things technologies are considered the next big step in Smart Building installations. Although such technologies have been widely studied in simulation and experimental scenarios it is not so obvious how problems of real world installations should be dealt with. In this work we deploy IoT devices for sensing and control in a multi-office space and employ technologies such as CoAP, RESTful interfaces and Semantic Descriptions to integrate them with the Web. We report our research goals, the challenges we faced, the decisions we made and the experience gained from the design, deployment and operation of all the hardware and software components that compose our system.
Delay Tolerant Networking (DTN) approaches based on the Bundle Protocol are commonly used within mobile IP based networks. Instead of being isolated applications, the Internet is often used to provide additional services or to route through other parts of the DTN network. A major drawback is that current DTN routing and discovery protocols are not generally applicable in the Internet as there is no common protocol to resolve DTN node names to convergence layer addresses outside a local network. We present Nasdi, an approach based on Distributed Hash Tables which can support naming, routing, notifications and service discovery in a heterogeneous DTN linked by the Internet. We present the architecture and initial evaluations of a Nasdi prototype system we built for the IBR-DTN software.
Utilizing mobile devices for gaining a better understanding of one’s surrounding, physiological state and overall behavior has been argued for in many previous works. Despite the increasing usage of mobile devices for research in this space, few platforms developed are readily available for supporting the wider research community. This paper presents a mobile sensing platform that allows for exploiting the latest and ever-increasing capabilities residing in mobile devices. While we highlight the main design and implementation characteristics of this solution, we also outline our experiences with this platform for typical usage scenarios in lifestyle management.
enterprises that adopt BYOD (Bring Your Own Device) need to optimize network selection for refundable employees’ business usage. They can ‘force-on-net’ business sessions when employees are on-site and seek ‘best connection’ when employees are off-site, perhaps via hospitality partners that provide WiFi connectivity. For non-fundable, unproductive personal usage, service requests should be ‘forced-off-net’ and deferred back to the personal carriers. To achieve this, we propose that the Enterprise decides whether to accept or change the originating access network, having established the funding status via the eBC (enterprise Business Context) model. For each service request, the Enterprise evaluates QoE and Affordability vectors that are derived from prioritized STANDS and CART factors respectively and the results are used to select the optimal access network. An enterprise Access Discovery and Selection Function (eANDSF) is proposed to enable recommending preferred corporate hospitality partners to employees, instead of the carrier’s list.
In this paper, we propose Tracommender, a context-aware recommender system, which uses background tracking information from smartphones to generate location-based recommendations. Based on the automatically collected data that consist of locations with timestamps, the dwell time at certain locations can be derived in order to use it as an implicit rating for a location-based collaborative filtering. We further introduce two alternative path matching algorithms that utilize continuous location sequences (paths) to compute path patterns between similar users. In addition, in order to overcome the cold-start problem of recommender systems, clustering algorithms are used to calculate so-called Activity Zones - locations taken from an existing database of categorized points of interest. Synthesized movement data has been applied to perform evaluations on performance, scalability and precision of an implemented prototype of the proposed recommendation algorithms.
Note-taking applications today have features that imitate the behavior of real handwriting such as smooth strokes, responsive interface, and stroke thinning. However, the problem is that the writing tools of these applications do not fully simulate the feel of their real counterpart. Some of these applications have flaws like unnatural pen thinning effect, pixelation of strokes, and lack of highlighter blending. To solve these, an iPad application with a custom smooth writing algorithm was developed. It was compared to five commercially available note-taking programs in the App Store.
The proliferation of mobile applications for smartphones has been based on the presence of a data connection between those devices and the Internet. Therefore, users of entry level mobile phones as well as users who cannot afford a data plan are excluded from the current dynamics of mobile applications consumption. A possible approach to include those users is the usage of SMS as the communication channel. Although SMS is widely supported by carriers and extremely popular among users, it is not trivial to replace the data connection capabilities with SMS. This work addresses this issue with an integrated solution composed by a client platform and a server-side system. This solution enables GUI-based mobile applications to be transferred over a short number of concatenated SMSs. It also presents components for the creation, deployment and distribution of mobile applications. In order to attest the relevance of such solution, an experiment was conducted and its results indicate a substantial level of acceptance by target group.
The Insense VM is a specialised Java virtual machine for running Insense programs on wireless sensor nodes. The VM runs on top of InceOS, a purpose-built operating system. A split VM architecture is used, in which Insense programs are compiled to Java classes, then linked and compacted on a more powerful machine into a form suitable for execution by the VM. Measurements demonstrate that the virtual machine achieves good performance and memory usage for realistic Insense programs.
The widespread adoption of online social networking to upload and share user-generated personal content rises novel issues related to the management of content ownership and privacy. To retain content ownership, we have previously designed and implemented an original solution for social-driven content sharing in home-to-home federated and spontaneous networks. This paper relevantly enhances our solution by proposing novel mechanisms to tune the visibility of shared resources based on the dynamic evaluation of social relationship tightness, inferred through social data gathered from widespread online social networks. On the one hand, the proposal autonomously evaluates the tightness of social relationships based on the primary guideline that the more users interact the tighter their relationships are. Moreover, it determines a default decision tree suitable for many application scenarios and enables its dynamic personalization based on user's feedback. On the other hand, it defines and supports a grammar to define visibility filters: runtime visibility of shared resources is automatically tuned based on both relationship tightness and defined filters. The presented prototype demonstrates how to effectively design/implement the proposal and the feasibility of our approach in terms of performance.
