STAC measures Metamako’s MetaApp latency at just 60 nanoseconds

New York and Sydney - 27 July 2015

In a series of new STAC-N1™  benchmark tests carried out by the highly-respected Securities Technology Analysis Center LLC (STAC®), Australian-based Metamako has set a new record benchmark for hybrid Layer1+ aggregation switches, averaging just 60 nanoseconds for each switch hop. Metamako is the leading specialist in deterministic ultra-low latency devices for the trading community, exchanges and telco providers.

MetaApp is the only STAC-N1™-verified device of its class, leveraging layer 1+ switching technology, FPGA, and an x86 server. It’s also the first device to effectively combine these three components. MetaApp is exceptionally flexible, facilitating easy development of applications by third party vendors, financial institutions and of course Metamako. New features, applications and upgrades – whether developed by Metamako or third parties – are delivered via downloadable software updates.

The possible use cases for MetaApp broaden significantly through applications developed by third-parties and Metamako. For example, the devices may be used for edge network applications such as trading algos, data capture, exchange gateways, pre-trade risk checking, book building, intrusion detection, firewalls and feed handlers.

Dr. Dave Snowdon, founder and co-CTO of Metamako, said: “The use case in the STAC Report reflects a typical exchange scenario, where orders are aggregated into the exchange and the return traffic from the exchange is distributed to multiple servers via MetaApp. Once again, the results demonstrate that our devices are exceptionally fast. However, what’s almost as important for trading applications is that the latency is deterministic, predictable and has extremely low jitter1.” He adds: “Our clients have been testing these devices since February; they’re now generally available and a number of devices are already in production.”

STAC tested the MetaApp 32, announced in June, by comparing results for multiple hops through the MetaApp 32 to those with no switch. The benchmarks measured the roundtrip involving two hops - through the device and then back - which represents a typical use case of multiple trading algorithms aggregating their orders to an exchange link. The network path involved both the MetaApp FPGA and the in-built Layer 1+ switch. The tests were conducted using 6 different message rates and 4 different messages sizes. MetaApp demonstrated an average latency for each hop of only 60 nanoseconds across all message rates. The average variation was just 20 nanoseconds at a rate of 800,000 messages/second and was undetectable at lower message rates.

Become a bobsguide member to access the following

1. Unrestricted access to bobsguide
2. Send a proposal request
3. Insights delivered daily to your inbox
4. Career development