Specifications

High Performance Trading/Algo Speed with Wombat Design and Implementation Guide

OL-15617-01

Testing

Methodology

Test Setup

Six servers are arranged as in Figure 8. The components consist of recorded OPRA (options price

reporting authority) data (from April 2, 2007), the Wombat playback mechanism mcpub using the

papareplay library, the Wombat OPRA feed handler (OPRA FH), and Wombat’s preferred performance

measurement client, mamaperf. Mcpub replays data from previously captured OPRA files. The goal of

the papareplay library is to mimic the original timing of the market data. This is important, since market

data is notoriously spikey, which can affect performance. The job of papareplay is to preserve the

relative timing of updates even as the playback rate is increased. The OPRA feed handler normalizes

and redistributes content, keeps a current cache, and otherwise manages the feed on behalf of consuming

clients. mamaperf subscribes to content and calculates and logs latency statistics (see

Data

Observations, page 16 for more detail).

Three mcpub instances were run on one server, each listening to a single OPRA line (multicast channel)

plus its backup (i.e., performing line arbitration) and publishing on a two multicast channels. These

transmitted exchange data over a GigE switch. Downstream was a single server running three instances

of the Wombat OPRA feed handler. Each instance was dedicated to a single OPRA FH channel. Twelve

mamaperf clients were distributed across four machines, three to a machine. Each mamaperf instance

subscribed to all of the items from one of the three channels. Across the system, this meant that there

were four consuming applications for each of the three OPRA FH channels.