Fast Information Transport for an Instrument Enabled Grid

Grid-based computing frameworks leverage underutilized processing and storage resources. We present and evaluate a new high-performance, reliable middleware layer that can incorporate instruments into a grid. This Java based messaging system supports remote distributed control and operation of scientific instruments, such as sensors and probes, thereby significantly expanding the grid's capabilities. Various comparative measurements show that our system outperforms the top-ranked publish-sub scribe Java systems in the market. Our sofiware can reach a peak message exchange rate of 900,000 messages per second, with a latency of less than half a millisecond on a 1 GB Ethernet switch.