5th International ICST Conference on Broadband Communications, Networks, and Systems

Research Article

Efficient Mapping of VoIP Traffic in Wireless OFDMA Systems Using Semi-Fixed Allocations

  • @INPROCEEDINGS{10.1109/BROADNETS.2008.4769072,
        author={Itzik Kitroser and Eldad Chai and Yehuda Ben-Shimol and Nathalie Yarkoni},
        title={Efficient Mapping of VoIP Traffic in Wireless OFDMA Systems Using Semi-Fixed Allocations},
        proceedings={5th International ICST Conference on Broadband Communications, Networks, and Systems},
        publisher={IEEE},
        proceedings_a={BROADNETS},
        year={2010},
        month={5},
        keywords={WiMAX Efficient Mapping VoIP},
        doi={10.1109/BROADNETS.2008.4769072}
    }
    
  • Itzik Kitroser
    Eldad Chai
    Yehuda Ben-Shimol
    Nathalie Yarkoni
    Year: 2010
    Efficient Mapping of VoIP Traffic in Wireless OFDMA Systems Using Semi-Fixed Allocations
    BROADNETS
    IEEE
    DOI: 10.1109/BROADNETS.2008.4769072
Itzik Kitroser, Eldad Chai, Yehuda Ben-Shimol, Nathalie Yarkoni

    Abstract

    This paper addresses the problem of efficient broadcasting of resource allocations descriptors for CBR applications (e.g., VoIP) in OFDMA-based wireless systems. In such systems mapping is the process of allocating resources and representing them is the mapping overhead. Existing mapping algorithms require an allocation descriptor to be broadcast in each frame for each allocated resource for every served subscriber. Information concerning the allocations is crucial for proper system operation, therefore it is broadcast on the wireless down link with low modulation and high repetition rate. The mapping overhead which is a function of the number of served subscribers is usually large and consumes a substantial part of down link resources. We present the notion of semi fixed allocations that allows a substantial reduction of the mapping overhead. We show that by taking advantage of the periodicity of VoIP frames and by applying semi fixed allocations in the existing resource allocation process, the mapping overhead can be reduced substantially. We present an innovative system model for the problem on hand and for evaluating the solutions. Using this model we show that the problem is in NP hard and present a set of observations on the resource allocation process leading to good mapping decisions from the perspective of mapping overhead. We present new mapping algorithms implementing those mapping decisions and show that by extensive simulations with these mapping algorithms the overhead can be reduced to 10% of the overhead experienced in OFDMA-based systems such as WiMAX and IEEE802.16, while retaining a low computational complexity.