Sync Monitor for PTP Networks: How does it work?

By Andreja Jarc.

Meinberg introduced a new monitoring tool called Sync Monitor, which is available in all Meinberg LANTIME Systems with firmware 6.22 or later.

Sync Monitor is used for monitoring the performance of NTP or PTP nodes in your network and can be accessed via the web interface. Moreover, it also enables monitoring of devices with 1PPS (pulse per second).  This feature might be useful in testing environments, for example.

The purpose of the Sync Monitor is to measure nodes’ time accuracy and show that they are within specification.

There are two main reasons for this.  One is to monitor your network to make sure that the timing is working correctly. The other is to create an archive measurement log that can be pointed to for regulatory purposes. For example, if there is ever a question about whether, or not, a server had the correct time during trading hours. You would have archived logs to prove that it was.  With Sync Monitor you are not just transferring the time but you can also measure it and know that it is correct.

In this post I will explain a general functionality and operation of the Sync Monitor. In the subsequent posts I will introduce use cases, where monitoring via Sync Monitor proves beneficial compared with standard self-reporting methods.

Let us consider the following PTP scheme:

Here is a fully redundant network with two grandmasters some switches and a several PTP slave devices. In this case the slaves are implemented as PCI express cards, but they can be any PTP capable device acting as a slave.  At the bottom of the network there is a monitoring device, called a PTP monitoring node. The PTP nodes can be monitored through the 1PPS signal but also through the network, which is often the way it is done.

Ideally you want a monitoring device with its own source of time, so it can measure the slaves against its own UTC traceable reference and make sure that slaves are reporting correct time. Secondly, if the monitoring is done through the network it should preferably be on different paths from the slaves than the grandmaster, to reduce common mode errors.

The Sync Monitoring does a kind of reverse PTP client server paradigm. The monitoring node would act as a client and a monitored device as a slave with a correct add-on to its PTP capability would act as a server.

In this case the monitoring node sends a delay request with a special TLV and gets a response back. This prompts the monitored device to send sync and follow up message if necessary.

The monitoring node now has the timestamps it needs to measure the time at the monitored device and compare it to its internal clock. It can log the measured difference and archive that log. In this case we are actually measuring the monitored devices, not just asking them if they have the correct time.

This monitoring scheme is a Meinberg proprietary, however we are talking to several other vendors to have their support.

The monitoring with TLV add-on is already supported on the following PTP Systems:

• All Meinberg IMS Systems with an HPS100 card and SyncBox -N2X (a PTP slave device)
• Oregano PTP aware devices (Switches, PCIe Card)
• Solarflare PTP adapters on Linux platforms (sfptpd)
• Domain Time II, PTP client for Windows OS from Greyware
• linuxptp open source (ptp4l) from Version 1.9.2 and following
• Meinberg PTP Client for Win and Linux OS

If you have any questions about the Sync Monitor, don’t hesitate to send me an email to andreja.jarc[at]meinberg.de or visit our website at www.meinbergglobal.com.