LeapRate's Daily Forex Industry Newsletter
Join now to receive first access to our EXCLUSIVE reports and updates.
Screenshot of a breaking news alert e-mail from Q2 2017
The technological arms race to bring latency to zero remains in full swing, as Australian technology company Matamako completes benchmark testing
Metamako, a supplier of ultra-low latency devices to exchanges and the trading community which is based in Australia, has announced the results of STAC-N1 Benchmarks carried out by the Securities Technology Analysis Center (STAC).
STAC has reported last week that it tested two systems that were identical except for what connected the servers. In the first test, the servers were connected by daisy chaining ports on three MetaConnect 16 switches, resulting in 48 total switch hops for each round trip message.
In the second test, the servers were connected by the same cables as the first system linked together by simple LC couplers rather than a switch. The test results showed that the mean per-hop latency of the MetaConnect 16 device was undetectable at the base rate of 100,000 messages per second and only 2 nanoseconds at the maximum rate tested of 600,000 messages per second. The STAC reports for the two systems are available here.
Metamako’s MetaConnect switch, launched in December 2013, is unique in combining ultra-low-latency performance with high levels of functionality which are usually only available on much slower devices. In benchmark tests it demonstrated excellent deterministic latency.
The industry-standard benchmark in accordance with which the testing was carried out uses the STAC-N1 Test Harness, which simulates the conditions in a typical trading network using data from US equities order-book feed handlers.
Scott Newham, co-Founder at Metamako, made a commercial statement last week to the effect that “Although in real life we would only expect to see one or two hops, the latency on MetaConnect 16 was so low that we had to put it through 48 hops to be able to measure the latency. The fact that the device is able to handle such a large number of hops means that its signal recovery capabilities are excellent. This in turn means that when MetaConnect is used over long distances the performance is exceptionally good, with the latency remaining close to zero.”
Mr. Newham further added that: “We are very pleased with the benchmark results, which demonstrate the exceptional performance of MetaConnect 16. We are confident that our performance will not be beaten in a meaningful way any time soon. ”
Having 16 ports on the device lowers the cost per device, potentially making MetaConnect 16 an attractive proposition for customers who only need a small number of ports, because they are, for example, using MetaConnect as a kill switch. Customers who need more ports will be able to purchase MetaConnect 48 which the company has stated represents a very closely related design, with similar latency and 48 ports.
In addition to extreme low latency, the key features of MetaConnect include packet statistics for diagnostics and troubleshooting; functionality for remote rewiring; network taps for splitting signals such as market data, without adding latency or compromising signal quality.
MetaConnect allows users to replace a significant number of devices in their server rack (patch panels, taps, signal regenerators, media converters and switches) with a single intelligent box which has negligible latency. Each of these devices introduces latency overheads, possible points of failure and ties up valuable rack space, as well as adding significant costs.