Console to Mixrack through a switch

Forums Forums dLive Forums dLive troubleshooting Console to Mixrack through a switch

This topic contains 3 replies, has 3 voices, and was last updated by Profile photo of Mr-B Mr-B 6 years ago.

Viewing 4 posts - 1 through 4 (of 4 total)
  • Author
    Posts
  • #68466
    Profile photo of BobA
    BobA
    Participant

    Hey all, I know it’s a dark topic, but I’m trying to connect my console to my mixrack through a switch. The switch I have is a Yamaha SWP1. I was wondering if anyone had a similar setup and got it to work. Would appreciate any advice.

    Yes I have read A&H’s guidelines about switches and the setup I have is fully compliant with those guidelines.

    dLive C1500 – CDM32

    TIA

    #68481
    Profile photo of Jay
    Jay
    Participant

    Bob,

    For GigaACE, you need layer 2 unmanaged or media converters if you want it to just “plug and play”. Once you get to a managed switch, it is no longer layer 2 transparent and some configuration has to be done.

    Everything below is “unofficial” since A&H doesn’t discuss protocol details.

    Here are the basics on the GigaACE packets:
    – Layer 2 only, destination on all packets is broadcast FF:FF:FF:FF:FF:FF.
    – Packet length 1276 (including headers), payload is 1248.
    – They are VLAN tagged (0x8100) for VLAN 1. The must stay tagged to be recognized.
    – Ethernet packet type is 0xe1e2.
    – Transmission rate is 96k packets/second.

    There also appears to be a multicast announcement packet that the surface uses to identify the connected mixrack, but I didn’t capture it. It appears to only be broadcast until the surface connects.

    To operate successfully, the switch must:
    – Disable EEE.
    – Accept tagged packets for VLAN 1 and send them back out tagged VLAN 1. This can be a challenge, since VLAN 1 is management and typically untagged.
    – Only 2 ports can be in VLAN 1. Because both connections are running broadcast, and you never receive your own broadcast packets, the data rate on these ports is fine. A 3rd port will receive ~2GB/sec data and have issues.
    – Be able to move management off VLAN 1. This is part of both “no other traffic” on the ports and restriction of 2 ports in the VLAN due to data rate.
    – Disable all other features on the surface/mixrack ports. CDP, GRVP, Spanning Tree, Bonjour to name a few.
    – Set qos priority on the ports to static and the highest priority queue. Make sure queues are running as a priority queue and not wrr.

    If you are trunking it across switches, the trunk ports:
    – Must have VLAN tagging enabled.
    – Must have qos enabled and trust the port (it uses the priority tagged on the packet).
    – Must have spanning tree, cdp, etc. disabled. This does create a risk of loops the switch won’t catch.
    – Should only have VLAN 1 enabled to make sure no other traffic crosses it, if you are on Gig ports. 10Gig should be ok to share with priorities.
    – LAG/LACP is a huge unknown here because it could significantly mess with clocking due to minor variances, though a 2nd trunk port that doesn’t carry the GigaACE data (no VLAN 1) should work. You get the data, no link redundancy. LAG/LACP also introduces some extra packets for management.

    Dante is significantly more forgiving than GigaACE, so the SWP1 may or may not work. I don’t have one to try. The Cisco SG300 series works fine, but you are limited to 1 GigaACE channel bridging through it with nothing else connected due to broadcast and being limited to Gig ports on trunking between switches. There are higher end Cisco switches (SG350 won’t) that allow VLAN remapping at the port level and have 10G interfaces that would work much better in this situation, but there is a significant cost with those.

    Jay

    #68574
    Profile photo of Jay
    Jay
    Participant

    To close a loop on this …

    I’ve been running a DM64 + S5000 through a Cisco SG300-10 for about 5 days. Only connection between them is through the switch (non-redundant) and the only 2 connections on the switch are the MR and Surface.

    I repeatedly receive an error on the home screen “Sync lock recovered”. When I clear it, it usually returns within 15 minutes. I have never seen this error with a direct connect.

    It would appear that one of two things is happening:
    – Other packets are getting on the wire. There is no way to prove/disprove this because I can’t packet capture at a 2gb data rate.
    – The switch can not maintain the data rate with no jitter for long periods of time.

    Either way, it doesn’t provide a usable glitch-free solution. While another switch may work better, the best option is probably to stay with layer 2 media converters and keep any “processing” out of the path.

    Jay

    #69309
    Profile photo of Mr-B
    Mr-B
    Participant

    I went the media converter route, two pairs kept the redundancy in place.
    I used the TP link MC200CM seems to be rock solid, think if I had to use a switch then the 10Gb route would be my choice.
    By going to fibre it allowed me to exceed the <100m limit (550m is the limit to my chosen converters in multimode). Not exactly sure on the cost but sub £50 per converter.

    Regards
    Ian B.

Viewing 4 posts - 1 through 4 (of 4 total)

You must be logged in to reply to this topic.