0

I have a newbie question on multicast delivery times. In this article on clock synchronisation they claim that the propagation time on a local network is close to zero.

What is the propagation time of multicast messages on a local ethernet network, wired or wireless?

closed as too broad by Mike Pennington, Ron Trunk, Adam Loveless, Teun Vink, Brett Lykins Mar 27 '15 at 4:31

Please edit the question to limit it to a specific problem with enough detail to identify an adequate answer. Avoid asking multiple distinct questions at once. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • Thanks for the comment, I have tried to rephrase it as a question. Ok? – Mellson Mar 25 '15 at 10:23
  • 1
    I'm still voting to close because as phrased, it's way too broad. There are zillions of different networks, all with different characteristics, depending on size, topology, type of equipment, media bandwidth, usage, etc. Essentially, there is no answer to your question. – Ron Trunk Mar 25 '15 at 10:29
  • Thanks for your input. I am mainly just trying to expand the claim from the mentioned article, which claims that the propagation time on a local network (wired and wireless) is close to zero. This to me seems like a bold claim, but I do not know where to look for evidence. I am trying to test this myself, but even if I time stamp incoming packets I have the problem of each machines clock not being in perfect sync. So it is really hard to test this claim myself. NTP and PTP does not provide a strong enough synchronisation to test the claim. Any ideas? – Mellson Mar 25 '15 at 10:33
  • @MikePennington you say "It depends, especially on wireless" - I guess I am looking for something like that. It depends on what? I need to multicast because I am constructing a synchronisation algorithm and I want to eliminate the senders delay (something similar to the work presented in the article). – Mellson Mar 25 '15 at 10:35
  • 1
    @Mellson, "it depends" on CSMA and the channel bit error rate (because the 802.11 data-link layer retransmits several times upon CRC errors, which effectively increases propagation time). In the face of high channel utilization and error rates, I have literally seen wifi ping times exceed 30 seconds to go to an AP 50 feet away – Mike Pennington Mar 25 '15 at 10:43
1

What is the propagation time of multicast messages on a local network, wired or wireless?

"it depends, especially on wireless", wired latency is normally in the low to mid-hundreds of microseconds range, depending on your specific switch hardware. If there are any ethernet frames queued ahead of your traffic, latency increases based on the serialization time of those frames ahead of yours in the queue.

...

Mellson asked @MikePennington you say "It depends, especially on wireless" - I guess I am looking for something like that. It depends on what? I need to multicast because I am constructing a synchronisation algorithm and I want to eliminate the senders delay (something similar to the work presented in the article).

Wifi latency depends on:

  • CSMA/CA
  • wifi channel bit error rate (because the 802.11 data-link layer retransmits several times upon CRC errors, which effectively increases propagation time).

In the face of high channel utilization and error rates, I have literally seen wifi ping round-trip times exceed 30 seconds to go to an AP 50 feet away.

Mellson asked I am mainly just trying to expand the claim from the mentioned article, which claims that the propagation time on a local network (wired and wireless) is close to zero. This to me seems like a bold claim, but I do not know where to look for evidence. I am trying to test this myself, but even if I time stamp incoming packets I have the problem of each machines clock not being in perfect sync. So it is really hard to test this claim myself. NTP and PTP does not provide a strong enough synchronisation to test the claim. Any ideas?

As for how to test propagation across any channel (wired / wifi): test equipment vendors Ixia, Spirent, etc... all offer precision timestamping assuming you have a good clock reference for the test set.

This equipment is expensive (and complicated). Bottom line if you want precision, you will pay for it... fortunately, you can lease this equipment. I used to do this as part of my $dayjob, but it's been quite a while and I've lost touch with capabilities and model numbers. Call Ixia or Spirent and tell them what you are trying to do; you'll get good guidance.

Not the answer you're looking for? Browse other questions tagged or ask your own question.