> From: shashank varshney <shashankva...@gmail.com>

> As mentioned, PTP switchover from one NIC port to another is not seamless
> when ports belong to different NICs with different hardware clocks.

That seems like a good application for devices which can synchronize their
hardware clocks.

> I am looking to utilize PTP for very time sensitive use cases like 5G
> wireless, stock markets, etc. which desire High Availability and
> redundancy at NIC port level while utilizing PTP.

That is why 5G wireless and stock markets typically use hardware with
support for external clock input (e.g. Intel 10Gb or 25Gb cards with
external clock and PPS inputs), and/or custom FPGA based designs.

> This is going to be a significant use case if we would like to move to
> open sourced environment and LinuxPTP is going to have significant
> role to play to achieve this aspect.

I'm not sure who you are thinking is going "to move" to an open source
environment, but in my previous employment I worked on an early roll out
of 5G equipment from one of the largest telco operators in the US, and
that environment was using nothing but open source.
See for example
https://www.redhat.com/en/en/topics/5g-networks/what-is-open-ran
or any of the other articles on OpenRAN and vRAN which have been published
in the last several years.

> Can someone help me with getting this functionality of seamless switchover
> by synchronizing Active Port PHC with PTP GM and then synchronizing
> system clock and standby PHCs with active PHC?

If you are purchasing large quantities of hardware for 5G or financial
trading applications then I am sure your device or system vendor would be
willing to help you, or perhaps Red Hat, SuSE, etc. if you are purchasing
commercial support for your OS installation.

First you will need to define your synchronization requirements (if it is
truly a single PHP used between multiple ports that is not really a
concern, but if you have two separate PHP that are synchronized you will
need to define your acceptable timing budget.

You then need to define specifics about acceptable switchover behavior
(timing, which software entity makes the decision about if/when to switch
paths, etc.).

After you have those requirements defined, then you can begin to evaluate
different hardware architecture choices and see what can possibly meet
your requirements.

After you have hardware in hand which should be able to meet  your
requirements based on specifications, research, communication with the
vendors, etc. you can setup some proof-of-concept systems and start
working with the software to see if you can actually achieve in practice
what the specifications imply you should be able to.

By that point you may have some specific questions about ptp4l that you
have not been able to understand from reading the documentation and
looking at the source code.  Someone on the list can probably help answer
your specific questions.

The open ended nature of your question does not inspire a lot of
confidence that you have gone through the first steps outlined above yet.

-- 
Chris Caudle



_______________________________________________
Linuxptp-users mailing list
Linuxptp-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linuxptp-users

Reply via email to