FIX Protocol Ltd. is addressing the issues of skyrocketing market data volumes through a new data compaction methodology called FIX Adapted for Streaming, also known as the FAST Protocol.

At the FPL Electronic Trading Conference in New York in early November, the Market Data Optimization Working Group, which formed in 2004 and developed the FAST protocol, reported that in pilot tests FAST is slashing the bandwidth and latency of high volume FIX messaging.

The group is working with Pantor, a Stockholm-based software company specializing in products for building and monitoring transaction systems, and Spryware, a Chicago-based software company focused on providing low latency, standardized market data. Microsoft is contributing to the development funding and a version of FAST will be available in C#.

Rolf Anderson, the CEO of Pantor, and Dan May, director and cofounder of Spryware, demonstrated FAST, showing its latency and CPU utilization. Instead of sending an entire FIX message, FAST encodes it to indicate whether it is related to messages around it, and then sends just the changed data. In high-volume trading where users are hitting the markets with hundreds of orders and cancellations in tight price clusters, this cuts out significant redundancy of data. As a result, FAST uses about a third of the bandwidth of a regular FIX message and has under half the latency, even with the encoding and decoding which has minimal impact on the CPU utilization. Further reports will be available after additional testing.

The protocol could help alleviate problems of trade data volumes going through the roof – caused by a combination of decimalization and algorithmic trading. Arca reports receiving up to 1,700 orders per transaction as algorithmic engines hit the site with orders and cancellations at lightening speed.