Performance evaluation of a multicast ATM switch with multiple common memories under bursty traffic. | |
| Author | Kanmanee Kulawanich |
| Call Number | AIT RSPR no. TC-03-13 |
| Subject(s) | Multicasting (Computer networks) Telecommunication -- Switching systems Asynchronous transfer mode |
| Note | A research submitted in partial fulfillment of the requirements for the degree of Master of Engineering, School of Advanced Technologies |
| Publisher | Asian Institute of Technology |
| Series Statement | Research studies project report ; no. TC-03-13 |
| Abstract | In this research work, the performance of an 8x8 structure of a common memory type two stage crossbar ATM switch using address assignment algorithm is evaluated by simulation under unicast and geometrically modeled multicast traffics in terms of cell loss ratio and average delay. The influences of input traffic type, burstiness of traffic environment and switch parameters to the switch performance are studied. The switch composes ofK common submemories, first-in first-out address buffer per output port and read/write controller. Replication at sending approach is applied for multicasting. Two traffic types, random traffic modeled by Bernoulli process (BP) and bursty traffic modeled by intenupted Bernoulli process (IBP) are considered. Traffic loads are uniformly distributed to output links. Simulation results show that compared to random traffic, the performance under bursty traffic is worse because of the unbalanced memory usage by the bursty traffic, especially in multicast environment. The longer average burst length introduces more cell loss and delay. In heavy multicast environment at cell loss ratio of 10-5 effective offered load of 0. 7 Erl./ slot with the SM capacity of 3 cells requires 160 submemories with average delay 80 cell times, for bursty traffic, but only 12 submemories is needed with average delay 2 for random traffic. It is observed that with random input traffic the multicast perf01mance is better than the unicast performance, but on the contrary, for offered bursty traffic multicast traffic can obtain lower utilization with more average delay. The switch with larger K can support higher utilization at the relevant cell loss ratio with the cost of higher delay. To satisfy the required utilization the switch has to be properly dimensioned by considering the delay effect especially in heavy multicast bursty traffic. |
| Year | 2003 |
| Corresponding Series Added Entry | Asian Institute of Technology. Research studies project report ; no. TC-03-13 |
| Type | Research Study Project Report (RSPR) |
| School | School of Advanced Technologies (SAT) |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Telecommunications (TC) |
| Chairperson(s) | Erke, Tapio J.; |
| Examination Committee(s) | Ahmed, Kazi M. ;Teerapat Sanguankotchakom; |
| Scholarship Donor(s) | TOT Corporation Public Company Limited; |
| Degree | Research Studies Project Report (M. Eng.) - Asian Institute of Technology, 2003 |