SP5
From AEOLUS IST Project
Wiki for AEOLUS SP5 - Extending global computing to wireless users
| SP 5: Extending global computing to wireless users |
|
The main objective of the subproject is to provide practically efficient algorithmic solutions for high quality, reliable, stable end-users services to heterogeneous wireless mobile networks; this will be necessarily accomplished by appropriately abstracting over the common characteristics of the diverse component devices. |
SP5 is structured into three Workpackages |
Participants List
|
Outline implementation plan
Wireless and mobile computing rapidly emerges from the integration among personal computing devices that compute and communicate in a distributed manner, cellular technology and the Web. This is possible due to the continuously increasing interaction between communication and computing, which is changing the information access from the current reactive “anytime anywhere” into the incoming proactive “all the time everywhere” approach. Such global computing scenaria are supported nowadays by a large variety of networks spanning from the well-known cellular networks to non- infrastructured wireless networks such as mobile ad hoc networks and sensor networks.
Such scenaria raise a number of interesting and complex algorithmic issues in diverse areas such as location management, resource allocation, ubiquitous information, network connectivity, reliability and security, and energy consumption. Easy, efficient, reliable communications on heterogeneous wireless networks are a key enabler for successful transparent extensions of overlay computing to mobile users. This need is combined with the increasing need for Quality-of-Service (QoS) guarantees. Indeed, mobile hosts and wireless networking hardware are becoming widely available, and extensive work has been done recently on integrating these elements into traditional wired networks, such as the Internet.
A main obstacle that has to be overcome for the full and profound accomplishment of the strategic goal of easy, efficient, reliable communications in heterogeneous wireless networks is the complexity of the challenges rising from the domain. This complexity stems from several factors. On one side, the target user base is very large and sophisticated applications are to be supported. On the other one, heterogeneous systems (with different computational resources) have to interoperate, while wireless stations have limited energy. Needless to say, the abstraction process over the common characteristics of the diverse component devices requires coping with new, unexpected computational problems.
User requirements and sophisticated applications that make innovative and extensive use of communication (such as video-on-demand) require more and more resources that are not always available to some of the wireless devices constituting an heterogeneous wireless network. Therefore, advanced and effective algorithmic tools and solutions for the careful and efficient management of the network resources are to be developed and used, otherwise the huge potential offered by new communication media will not be fully exploited. Indeed, we have reached a point in the technological evolution of the communication domain where any further progress on the transmission technology (e.g., new communication technology with a tenfold increase in bandwidth) does not have an equivalent impact on the capabilities and services offered to the information and knowledge-based society, unless it is coupled with adequate structural and algorithmic tools for the associated design and provisioning issues.
A special case of high importance arises whenever mobile users want to communicate in situations where no fixed infrastructure is available (either because it may not be economically practical or just because it is not physically deployable or was damaged), then the hosts with wireless network interfaces may set-up a temporary network without the aid of any centralized administration. This type of network is known as an ad-hoc wireless network [RT99]. Ad-hoc networks have been attracting a growing attention world-wide due to the remarkable potential applications in a variety of fields such as inter-vehicle communication, sensor networks, environmental control and defense. In the USA, DARPA heavily supports research in this area and the NSF panel on networking put a strong emphasis on all-wireless systems. In EU, telecommunication companies and defense services are also strongly interested. Ad-hoc networks are also expected to play and important role in environmental monitoring (radiation, toxic wastes, and other emergency situations), driving safety, cellular networks coverage extension as well as in defense. Technology makes indeed possible an emerging scenario in which different kinds of ad-hoc wireless networks are parts of a larger, heterogeneous wireless network, possibly comprehensive of even different kinds of wireless devices, e.g., mobile phones, sensors, PDAs, laptops, etc.
The main objective of the subproject is to provide practically efficient algorithmic solutions for high quality, reliable, stable end-users services to heterogeneous wireless mobile networks; this will be necessarily accomplished by appropriately abstracting over the common characteristics of the diverse component devices. In turn, the provision of such algorithmic solutions requires facing a number of strongly related issues that can be grouped in the following categories, directly corresponding to the workpackages of this subproject:
- WP5.1 Resource management and quality-of-service
- WP5.2 Network design and topology control
- WP5.3 Mobility and fault tolerance
In the context of overlay computing, the following key issues must be necessarily addressed by our solutions:
- Heterogeneity transparency: protocols have to correctly execute whatever devices are actually composing a network. Users are not required to know the actual network composition in order to use it.
Reliability/stability: protocols operate in a mobile, dynamically changing environment, where mobile devices frequently join/leave the network. Such changes should be transparent to the users; hence, protocols' running times and quality of the solutions should be as independent as possible of the dynamic nature of the network.
- Scalability/adaptivity/fault tolerance: since mobile devices frequently connect/disconnect the network and the network size has to be transparent to the users, protocols must run efficiently independently of this size and be able to adapt appropriately to conditions by changing in an on-line fashion. Fault tolerance is strongly related with the concept of adaptability.
- Distributed issue: no centralized protocol can be realistically considered in an ad-hoc wireless network relying on no established infrastructure.
- Practical Efficiency. If short running time is an appreciated characteristic in every application, in the heterogeneous wireless networks setting it becomes crucial since protocols must potentially run on several different devices having specific features and limitations to be regarded. Hence, we are not interested in, say, complex linear programming based algorithms; rather we shall often focus on simple and fast heuristics which work well on the average using local/limited knowledge. Finally, notice that the high quality services requirement is often in contrast with practical efficiency.
- Energy Efficiency. We are interested in solutions requiring low energy consumption. We want to remark how, similarly to the practical efficiency, energy efficiency is really crucial in the heterogeneous wireless networks setting. Transparency still requires that protocols run independently of the actual network composition.
Detailed implementation plan – first 18 months
The design of resource-efficient algorithms for mobile networks as one extension of an overlay computer is a challenging task. Since such networks frequently change topology, and often without any regular pattern, it is important to consider topology maintenance to enable fast, reliable and scalable communications within this mobile and spontaneous network. In addition, the nodes of an ad hoc network may have heterogeneous capabilities. They differ, e.g., in efficiency, bandwidth, energy, communication media, and mobility. Open questions are, for example, what is realistic mobility and how do we handle fault tolerance in these heterogeneous wireless networks?
The main goal of this subproject is to model all these characteristics and to design, analyze and implement algorithms for resource-efficient network management which make computing and communication on such a mobile ad hoc network (like an extension to an overlay computer) possible.
Needless to say, the abstraction process over the common characteristics of the diverse component devices of a heterogeneous wireless network requires to cope with new unexpected computational problems in that user requirements and sophisticated applications require more and more resources that might be not always available to the (different) wireless devices. Hence, new advanced applications that make innovative and extensive use of communication networks need a careful and efficient management of the network resources. Therefore, advanced algorithmic tools and solutions are to be developed and used, otherwise the huge potential offered by new communication media will not be exploited. Indeed, we have reached a point in the technological evolution of the communication domain where any further progress on the transmission technology (e.g., new communication technology with a tenfold increase in bandwidth) does not have an equivalent impact on the services provided to the information and knowledge-based society, unless it is coupled with the adequate structural and algorithmic tools for the associated design and provisioning issues. The main objective of the subproject is to provide practically efficient algorithmic solutions for high quality, reliable, stable end-users services to heterogeneous wireless mobile networks; this will be necessarily accomplished by abstracting over the common characteristics of the diverse component devices. In turn, the provision of algorithmic solutions for high quality, reliable, stable end-users services to heterogeneous wireless mobile networks requires/means to face with a number of strongly related issues that can be grouped in the following categories, directly corresponding to the workpackages of this subproject: