STAC Report: STAC-TS with MetaApp32 rev A5A & MetaWatch 0.5.2

Worst-case absolute accuracy  of -0.5 +/5.5 nanoseconds for timestamps from any port on a single device

17 October 2017

STAC recently performed relevant STAC-TS benchmarks on a stack consisting of two Metamako MetaApp32 rev A5A switches running MetaWatch 0.5.2 firmware on the MOS-0.14.0alpha3 operating system. The STAC Report is available here.

STAC-TS is a set of benchmarks and software tools that measure the accuracy and other characteristics of solutions used for time synchronization, timestamping, and event capture. STAC-TS embodies industry best practices formulated by leading vendor firms and the trading community. Independent research reports and self-service tools based on STAC-TS will help trading organizations demonstrate compliance with regulations such as MiFID 2 as well as design state-of-the-art timing and capture solutions for low latency trading.

The stack we tested consisted of two Metamako MetaApp32 rev A5A switches running MetaWatch 0.5.2 firmware on the MOS-0.14.0alpha3 operating system. Each device contained Finisar FTLX8574D3BCV SFPs and was synchronized to a common PPS signal using the TimeTech Pulse Distribution Unit (part number 10535).

Of the many results in the report, Metamako chose to highlight the following:

• Worst case absolute accuracy of -0.5 +/- 5.5 nanoseconds with respect to the PPS source, for timestamps obtained from any port on a single device (STAC-TS.NTE2)

• Worst case error of -1.5 +/- 4.5 nanoseconds between any two ports on a single device (STAC-TS.PSE1.TOTAL)

• Worst case error of 1.5 +/- 8.5 nanoseconds between any port on one device and any port on the other device (STAC-TS.PSE2.TOTAL)

• Best case random error between two ports on a single device (error that cannot be calibrated out) of +/- 1 nanosecond (STAC-TS.PSE1.RAND.BEST)