Corvil » News Coverage

The first phase of the rollout, which is already underway, involves deploying the vendor’s CorvilNet monitoring system and CorvilClear peer-to-peer latency metrics sharing technology to monitor latency across multiple datacenters in London, where Morgan Stanley hosts a distributed market data and trading technology infrastructure that supports its institutional equity and equity derivatives trading, says Kevin Twitchen, executive director at Morgan Stanley’s institutional equity division.
Once the London rollout is complete, the bank will embark on a second phase to deploy the latency monitoring systems in its New York trading operations, and will consider rolling out the technology to other European cities where it also maintains datacenters, Twitchen says. 
Morgan Stanley will initially use the Corvil systems to monitor latency introduced by components such as feed handlers and exchange trading gateways, as well as network connectivity between its own datacenters and exchange hosting centers, although Twitchen says that,

“as we go through this phased rollout, we will be expanding the devices and venues that we monitor.”

Though the bank has measured roundtrip order latencies for some time, the Corvil technology enables it to focus on individual latency components, such as market data delivery from each trading venue-for example, by using the   CorvilClear peer-to-peer latency monitoring technology in conjunction with market venues such as European multilateral trading facility Turquoise, which deployed CorvilClear earlier this year (IMD, April 20).

“The first step is to work with the exchanges, MTFs and third-party service providers so that we can achieve complete visibility on the latency of our European market data and trading environment,” Twitchen says.

Corvil is in talks with other exchanges that use its technology to monitor internal latencies, to promote more widespread adoption of CorvilClear, says Corvil chief executive Donal Byrne. But for venues that do not support peer-to-peer latency monitoring, Corvil can also offer workarounds, such as placing its monitoring technology at a co-location or proximity site next to an exchange’s systems, Byrne adds.
The ability to derive a full latency profile that shows how each venue performs under different market conditions is an increasingly important factor for determining where to direct order flow,

Byrne says. “Exchanges mostly just publish the average latency of the speed with which they can execute an order, and very few publish the speed of their market data. But traders need to know the performance of the venue under peak loads for market data delivery and order execution,” he says.

Firms can then use venues’ latency profiles as an input to their smart order-routing engines, enabling them to adapt to different market conditions.

“Understanding the speed with which different parts of our European trading infrastructure are processing market data and executing trades, can help us complete a ‘probability of fill’ matrix that we can then use in our smart order routers,” Twitchen says.

Jean-Paul Carbonnier