K-ary n-cube based off-chip communications architecture for high-speed packet processors

Abstract

Abstract-A k-ary n-cube interconnect architecture is proposed, as an off-chip communications architecture for line cards, to increase the throughput of the currently used memory system. The k-ary n-cube architecture allows multiple packet processing elements on a line card to access multiple memory modules. The main advantage of the proposed architecture is that it can sustain current line rates and higher while distributing the load among multiple memories. Moreover, the proposed interconnect can scale to adopt more memories and/or processors and as a result increasing the line card processing power. Our results portray that k-ary n-cube sustained higher incoming traffic load while keeping latency lower than its shared-bus competitor.