High Frequency Trading on a Low-Latency Fabric in a Cloud Environment

To attract order flow and effectively match orders, Matching Engines have been designed to handle very high message rates at deterministic low latencies. Historically, these environments were built with proprietary technology deployed on dedicated on-premise infrastructure to minimize latency and control jitter. However, as outlined in this paper, we believe that innovations in the cloud service provider infrastructure enables configuration and implementation of an equally low-latency trading platform in a cloud environment. We have demonstrated this by implementing a simple foreign exchange (FX) trading system communicating via the Remote Direct Memory Access (RDMA) protocol on the InfiniBand fabric in the public cloud environment Azure. Our study concludes that public cloud can provide a robust, reliable, and efficient environment for high-frequency, low-latency trading.