12.07.2015 Views

Wireless Ad Hoc and Sensor Networks

Wireless Ad Hoc and Sensor Networks

Wireless Ad Hoc and Sensor Networks

SHOW MORE
SHOW LESS
  • No tags were found...

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

2 <strong>Wireless</strong> <strong>Ad</strong> <strong>Hoc</strong> <strong>and</strong> <strong>Sensor</strong> <strong>Networks</strong>simply require high throughput. With the use of the Internet diversifying<strong>and</strong> exp<strong>and</strong>ing at an exceptional rate, the issue of how to provide necessaryQoS for a wide variety of different user applications is also gainingimportance. This book attempts to clarify the QoS issues <strong>and</strong> examinesthe effectiveness of some proposed network solutions mainly by usingcongestion, scheduling, <strong>and</strong> admission control.In short, QoS depends on the statistical nature of traffic. An appropriateservice model should be defined, <strong>and</strong> some network QoS control methodsshould be engineered to meet a range of QoS performance requirements(e.g., throughput, delay, <strong>and</strong> loss), which are usually represented as a setof QoS parameters associated with the service model. There are two maintraffic types: delay-sensitive <strong>and</strong> loss-sensitive traffic. Delay-sensitivetraffic is characterized by rate <strong>and</strong> duration <strong>and</strong> may need transmissionin real-time. Examples include videoconferencing, telephone, <strong>and</strong> audio/video on dem<strong>and</strong>, which usually have stringent delay requirements butcan accept a certain data loss. Loss-sensitive traffic is characterized by theamount of information transmitted. Examples are web pages, files, <strong>and</strong>mail. It usually has stringent data-loss requirements but no deadline forcompleting a transmission.There are other traffic types, such as multicast traffic [e.g., conferences,distributed interactive simulation (DIS), <strong>and</strong> games], <strong>and</strong> traffic aggregation[e.g., from local area network (LAN) interconnection]. Observationsof LAN traffic (Stallings 2002) reveal that traffic on the network displaysself-similar or long-range dependent behavior. The rate is variable at alltime scales; it is not possible to define a duration over which the trafficintensity is approximately constant. These observations have been confirmedrepeatedly for every type of network. A plausible explanation forself-similarity is that LAN traffic results from a superposition of burstswhose duration has a heavy-tailed distribution.The networks have been evolving to provide QoS guarantees to users.For instance, ATM, widely adopted in the backbone network, can reservethe b<strong>and</strong>width <strong>and</strong> buffer for each virtual connection. Similarly, the Internetintegrated service (Intserv) can also provide QoS for each flow in theInternet Protocol (IP) network. Internet differentiated service (Diffserv)provides different treatment options for packets of different classes,instead of on a flow basis, so that it has better scalability than Intserv.Multiprotocol label switching (MPLS), a recent technology trend for theInternet, allows the network providers to have better control <strong>and</strong> provisionof QoS through traffic-engineering policies.The ATM forum summarizes the traffic parameters used in ATMnetworks. The constant bit rate (CBR) service category applies to connectionsthat require cell loss <strong>and</strong> delay guarantees. The b<strong>and</strong>width resourceprovided to the connection is always available during the connectionlifetime, <strong>and</strong> the source can send at or below the peak cell rate (PCR) or

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!