Stock Exchange – Market Data Monitoring (by Rohan Vartak, ITRC)

What is market data?

In financial markets, market data is the electronic streaming of data, which contains price and volume related trade data for a financial instrument traded by a trading venue. This data may be real-time or historical.

Market data consists of an instrument name, symbol, and latest bid and ask price. While firms which connect directly to an exchange receive market data through dedicated WAN links connected to the exchange, others take market data from market data vendors who resell it.

Need for real-time monitoring of market data

• Market data is used by trading and investment firms to buy or sell an instrument
• On time & accurate market data will instill confidence in the trading community to trade on an exchange
• The number of transactions executed on an exchange are directly proportional to the reliability of market data on that exchange

Use Cases for Market Data Monitoring in a stock exchange

So let’s have a look at how an exchange can setup monitoring of its market data with the help of some practical use cases.

Availability – The exchange should monitor, in real-time, if all the market data feeds are being disseminated. This means having a tool subscribed to all the data feeds at the end of the exchange’s infrastructure, typically at the end of the last market data switch (to mimic the end user).This tool can then measure and report on the availability of market data, typically via a dashboard, so that IT teams are always aware if there are any issues.

Metrics for availability monitoring include:

• Feed being received (yes / no)
• Data being received in bid/ask fields for highly traded symbols
• Sequence gaps
• Out of sequence messages received

Performance – Market data feeds are subscribed to by firms and brokers over WAN links connected to the exchange. Thus is it important for an exchange to generate market data with suitable total bandwidths so that the WAN links are optimally used.

Metrics for performance monitoring include:

• Latency between execution and market data dissemination
• Time interval between ticks for an instrument
• Bandwidth of each feed (in kbps or Mbps)
• Total Bandwidth against WAN capacity
• Middleware statistics (retransmissions, loss, buffering etc.)
• Message burst metrics to ensure smoothen market data at set bandwidths

Accuracy – Market data is highly time and accuracy sensitive. Since market data is a snapshot of the exchange’s order book, monitoring the market data enables the exchange to monitor its own execution order book’s health as well.

Metrics for accuracy monitoring include:

• Bid is less than the ask
• Data types in market data fields on a sampling basis

Resiliency – Like all components of a stock exchange, market data needs to be available at all times and hence built for resiliency. Typically, exchanges disseminate 2 feeds (feed A and feed B) for a particular market data stream. Brokerages then subscribe to both feeds and use the one which is fastest.

Metrics for resiliency monitoring are:

• Relative latency between feeds A and B
• Price comparison between feeds A and B

Challenges in setting up exchange market data monitoring

• Increasing market data volumes
• Monitoring downstream service provider’s quality of service
• One monitoring tool to collect and analyze data and then alert relevant personnel should an issue arise