Stability of a type-II hybrid ARQ protocol for DS-SSMA packet radio systems

Qian Zhang, T.F. Wong, J.S. Lehnert
Proceedings. IEEE INFOCOM '98, the Conference on Computer Communications. Seventeenth Annual Joint Conference of the IEEE Computer and Communications Societies. Gateway to the 21st Century (Cat. No.98CH36169)  
The perform ance and stability of a slotted direct-sequence spread-spectrum multiple-access (DS-SSMA) packet radio network employing the type-II hybrid automatic-repeat-request (ARQ) protocol with ® nite-length transm itter buffers are considered. The equilibrium point analysis (EPA) technique is employed to analyze the system stability and to approxim ately compute the system throughput, delay, and packet rejection probability. It is found that the system exhibits bistable behavior in som e
more » ... uations. Issues of system design, such as the required length of the transmitter buffers and the desirable region of operation based on a predetermined perform ance requirement for packet rejection, are also investigated. of length 1 will reject too many packets attempting to enter the transmitter buffer. The goal of this paper is to investigate the performance and stability of a DS-SSMA packet radio network employing the type-II hybrid ARQ protocol with ® nite-length transmitter buffers. Although we know [1, 2] that the system dynam ics can be modeled exactly by a multidimension al Markov chain, this model is too complex to allow us to study the system performance and stability. In this paper, we employ the equilibrium point analysis (EPA) technique [3±5] to approxim ately compute the system performance and analyze the system stability. Approxim ate system throughput, delay, and packet rejection probability are obtained via EPA. The accuracy of the EPA approximations is checked by comparing the results given by EPA against the results obtained from Monte Carlo simulations. A key result obtained in EPA of the system is that the system stability can be described in terms of the number of stable (de® ned in Section 3.5) equilibrium points. Like some ALOHA systems [6, 7] , we observe that the system exhibits bistable behavior, i.e., there are two stable equilibrium points in some situations. When the system is bistable, it oscillates between the two stable equilibrium points, one of which corresponds to a sys-
doi:10.1109/infcom.1998.662945 dblp:conf/infocom/ZhangWL98 fatcat:jpf23oox7bethhgttejbppleye