Dublin-based technology vendor Corvil will this week release Version 6.1 of its CorvilNet latency management system, which includes a module that allows users to monitor application latency and performance within the same display console that the vendor developed for monitoring network latency, officials say.
Corvil rolled out the application latency monitoring module for beta testing with four unnamed clients—including a stock exchange, an options exchange, a global market maker and an execution management provider—at the end of March, and is now making it generally available as a software upgrade to all existing CorvilNet appliance users. To support application latency monitoring, as part of the software upgrade, Corvil will provide pre-built decoders for a range of the most popular trading and market data protocols—such as the FIX Protocol—used by marketplaces in Europe, North America and Asia.
Although Version 5.1 of CorvilNet provided some support for latency monitoring at the application messaging layer (IMD, June 1, 2009), the new release offers enhanced decoding capabilities that allow users to monitor even the most complex binary messaging protocols. “We already had some capabilities to measure latency on FIX and other simple ASCII protocols by picking out particular fields, but we now have full generic binary decoding capabilities built-in, which supports all protocols, no matter how complicated the encoding,” says Donal O’Sullivan, vice president of product development at Corvil.
The new software also includes specific support for market data and trading applications, including feed handlers, smart-order routers, FIX gateways and matching engines, with analytics that monitor events on an ongoing basis, such as the time taken from sending an order to receiving an order acknowledgment, and then for the order to be fed back as a market data update.
In addition to supporting a wider range of message encoding and analytics, CorvilNet 6.1 also includes an enhanced display console that allows users to interrogate data and run analytics on both application and network messaging latency at the same time.
“If the network guys are seeing one set of metrics and the application guys are seeing another, then they’ll be speaking different languages,” O’Sullivan says. “An independent, appliance-based application and network monitoring system allows everyone to talk the same language and saves money by reducing the number of systems you need to deploy, and by speeding up the time [needed] to diagnose the underlying root cause of a latency spike,” he adds.
By the end of next month, Corvil also plans to release a software development kit (SDK) that will enable users to decode messages used by in-house or third-party applications, thus allowing them to monitor the latency of any application, even those using proprietary data formats. “Some customers will use proprietary protocols internally. With the SDK you could take one of our decoders out of the box and then modify it to monitor your own internal protocol,” O’Sullivan says.
Jean-Paul Carbonnier