An Analysis of the Implementation of Kafka in High-Frequency Electronic Trading Environments

Abstract

Electronic trading amounts for the vast majority of all financial transactions with bonds and equities in the world. This type of trading is based largely on the brokering of messages from the buy side, brokers or financial institutions, to a sell side (usually an exchange). Since not all trading is done at the same time these messaging systems need to account for server down times, sequencing, high throughput and performance requirements. Therefore, most fintech companies employ queuing mechanism to manage the message flow. This paper analyzes the way in which Kafka, one of the premiere messaging systems currently in use, can simplify various systems in use such as message brokering, persistence or fault tolerance. The authors of this paper hope to demonstrate the use of Kafka as a messaging system, backup solution and alert broker for software operators and developers with low fault-tolerance and rigid up-time requirements.