Effect of Buffering Capability in Data Vortex Networks Based on 4-ary Routing Scheme

Abstract

Buffering capability is added within each routing node of the modified Data Vortex network with the 4-ary routing scheme. In two different cases, either with single packet or with two packets routing within the node, the effect of buffering capability is studied in such networks. The results have confirmed that with additional hardware cost in buffering, the performance can be enhanced as that in the original binary routing Data Vortex networks. While buffering based on single packet routing allows for modest performance improvement, buffering based on two packets routing allows for much better latency and throughput performance compared with buffer-less networks, and such improvement is also more significant than the performance enhancement in binary decoding Data Vortex networks with the same buffering capabilities.