BU CS 580S - A Survey on Dependable Routing

Unformatted text preview:

AbstractThe class of wireless and mobile networks features a dis-similar set of characteristics and constraints compared totraditional fixed networks. The various dimensions of thesecharacteristics/constraints strongly influence the routingsystem, which is often regarded as the glue of a network. Weintroduce the concept of routing dependability describingthe trustworthiness of a routing system such that reliancecan justifiably be placed on the consistency of behavior andperformance of the routing service delivered. We investigatethis concept by analyzing the basic characteristics of vari-ous networks. Subsequently, we derive the most importantattributes and impairments that contribute to routingdependability in sensor networks, ad hoc networks, andinfrastructure-based cellular networks. Departing fromstate-of-the-art network designs, we extend our survey tocover future network architectures as well. We finish bybriefly investigating possible directions and means thatallow mitigating the deprivation of dependability. 1. IntroductionMobile communications and wireless networking technol-ogy has seen a thriving development in recent years. Drivenby technological advancements as well as applicationdemands, various classes of communication networksemerged. In this paper, we are particularly interested in sen-sor networks, ad hoc networks, and cellular networks, eachof which class represents a solution to important chapters inthe mobile and wireless communications challenge.1.1. Sensor NetworksSensor networks for the collection, fusion, and communi-cation of environmental information are considered to havean outstanding potential for research and application [1].Basically, sensor networks are defined by the combinationof miniaturized sensors with communication technology.Possible applications for sensor networks include themeasurement of temperature/humidity [2], the collection ofpollution data, the monitoring of weaknesses in buildingstructures, and the detection of chemical agents, to name afew. A main advantage of distributed and collaborative mea-surements includes the non-obstrusiveness and the increasedaccuracy of the data collection [3]. These applicationsdemand for smart but cheap sensors, which operate self-organized even under harsh environmental conditions.Currently, sensor networks are considered to evolvetowards so-called “smart dust” if technological advance per-mits such miniaturization [4]. However, severe limits, forinstance, in energy supply, costs, maintenance of oncedeployed sensor nodes and reliability of operation persist.These and other limits are especially important for the com-munication aspects of sensor networks as we show later.1.2. Ad hoc NetworksThe visions of untethered communications and pervasivecomputing make a strong case for the self-organizing opera-tion of mobile and wireless nodes within ad hoc networks.Possible civilian application domains of such networksinclude inter-vehicular communications [5], disaster recov-ery, multimedia home entertainment, and zero-configurationpersonal area communications. Furthermore, there are fewproposals for wide area ad hoc networks [6]. All these applications have certain demands in common:either there is impromptu need for communication, or theabsence of infrastructure commands the network to be fash-ioned from whatever resources are immediately available.Moreover, the autonomous and cooperative operation isinherent to the network nodes, which are terminals (end sys-tems) and routers (intermediate systems) at the same time. However, the scope and features impose constraints on adhoc network operation such as, for example, limits for net-work size, high topology dynamics, unpredictability of sys-tem characteristics, etc.1.3. Cellular NetworksEnabled by cellular telecommunication networks, the suc-cess of mobile communications has been second to none—according to [7] the number of mobile subscribers surpassedthe one of fixed networks in early 2003. Until now, the num-ber one application of cellular networks has been personalvoice communication, which has been also reflected in thenetwork development. Application demands for these net-works include a high quality of service as well as a highgeographical coverage. These demands are usuallyaddressed using hierarchically designed and centrally man-aged infrastructures [8]. Emerging market opportunities andthe success of Internet applications such as electronic mailA Survey on Dependable Routing in Sensor Networks, Ad hoc Networks, and Cellular NetworksMatthias Hollick, Ivan Martinovic, Tronje Krop, and Ivica RimacMultimedia Communications Lab (KOM), Department of Electrical Engineering and Information TechnologyDarmstadt University of Technology, Merckstrasse 25, 64283 Darmstadt, Germanyhttp://www.kom.tu-darmstadt.de[Matthias.Hollick;Ivan.Martinovic;Tronje.Krop;Ivica.Rimac]@KOM.tu-darmstadt.deProceedings of the 30th EUROMICRO Conference (EUROMICRO’04) 1089-6503/04 $ 20.00 IEEEand the World Wide Web introduce various research chal-lenges to cellular networks. We observe a convergence ofservices while at the same time the heterogeneity of networktechnologies prevails. These concerns are partiallyaddressed in the International Mobile Telecommunications2000 (IMT-2000) framework. The limitations of cellular networks lie in the necessity foran expensive infrastructure. Moreover, the flexibility of thenetwork is highly restricted due to its centralized manage-ment and control [9].1.4. Future NetworksThe Internet has never been designed to support the heter-ogeneity, dynamics, and mobility it faces today with theintegration of a wide range of wired and wireless technolo-gies. Satellite and mesh networks extend the Internet’s coreof high speed fiber optics with a highly configurable wire-less infrastructure, allowing service mobility in differentscales ranging from mobile devices [10] to mobile networks[11]. The increasing dynamics and heterogeneity take theirtoll on breaking up trust and service relations of previouslywell controlled static networks. Moreover, the routing trans-parency of the end-to-end paradigm is broken by enablingmiddleboxes [12].These problems are addressed in the proposed futurearchitectures for the Internet (see, for example, [13], [14],[15], and [16]). The underlying concept is to separate theidentity resolution from the forwarding mechanism [17] toreestablish network layer transparency for services on top ofheterogeneous networks


View Full Document

BU CS 580S - A Survey on Dependable Routing

Download A Survey on Dependable Routing
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view A Survey on Dependable Routing and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view A Survey on Dependable Routing 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?