dLive Switch

[a.malo]

Hello:

I was wondering if anyone has used any switches and sfp+ with dlive?
What have their experiences been, what brands and models have they used.

Thank you for your answers

[Brian]

For which connection? Network, GigaAce, DX, ME etc? Please check out this article for some helpful information. https://support.allen-heath.com/hc/en-gb/articles/4403553378449-Audio-Networking-gigaACE-and-DX-networking-VLANs-and-fibre-optics

[SteffenR]

For what? Control network?
Every standard switch is working!

Ubiquiti, Netgear, TP-Link, Zyxel is what I tried so far.
And I’m aware that others used Cisco, HP and D-Link successful.

[msteel]

Besides what others have said and/or linked, I will add some personal experience.

We have a DX168 in one building that connects via fiber. We used one strand of a cable that already existed as part of the building’s network infrastructure. There is a box on each end to convert to/from copper. The box uses an SFP slot and I got the module our network people said was correct for the type of fiber. In our case the converter is Gigabit capable and we did not even have to force it to 100Mbps for the DX link (contrary what is suggested in the white paper @Brian linked). The link is rock solid to the point that I forget it is there.

Before I realized we had existing fiber I briefly experimented with a Netgear GS105e switch to extend the distance of copper DX. That was pretty much plug and play. In theory GigaACE should work the same way but I never tried.

[a.malo]

Thank you all for the answers and sorry for not responding.
My intention is to use a switch between stage and foh and separate the traffic into vlans. Vlan 100 for GigaAce, 200 audio control (IEM, wireless…), 300 sACN, 400 NDI, 500 guest…. and use 2 trunk ports with Link Aggregation to have backup, and 1 trunk to control the network.
I like Mikrotik crs310-8g+2s+in and Zyxel XGS1250-12

[TJCornish]

My day job is an enterprise network admin.

I would not recommend a setup as you describe. VLANs provide some network segmentation, but all of the traffic still has to flow on the link. LACP and the like do not failover fast enough to avoid audio disruption, and the extra complexity of switching loops and QoS will likely work against your end goal of reliability, not to mention at least two more powered devices in the chain that could fail.

If you need to go fibre for range – fine. Otherwise, a pair of redundant links between your mix rack and stage boxes is your best bet. If you don’t have redundant DLive hardware and are just worried about cable failure, run an extra Cat cable and be ready to swap on both ends.

[a.malo]

but if I only use the c3500 as a control table and I have a cable cut, everything still rings since the processing is done by the CDM48. I would only lose the audio that I played from foh, which is only music between performances.
This configuration would allow me to not worry about cable failures.

According to this post it is possible

GigaACE & Dante over switches and 10G Lightpipes

GigaACE & Dante over switches and 10G Lightpipes

[Brian]

My intention is to use a switch between stage and foh and separate the traffic into vlans. Vlan 100 for GigaAce, 200 audio control (IEM, wireless…), 300 sACN, 400 NDI, 500 guest….

While I can understand the desire to have as few wires running from FOH to the stage, I think you are asking for a lot of trouble by trying to combine all of those functions. Personally I would leave the GigaAce connection as a dedicated direct connection. The other functionality are all “normal network” communication, so VLANs should work, but keep in mind that it is very easy to saturate a connection with NDI so depending on how much NDI data you are trying to transmit, that may be another stumbling block.

That being said, if you are OK with the potential for the Surface to loose its connection to the Mixrack and you have a backup method of controlling audio, then the overall risk to the show seems small. I still wouldn’t do it personally, but if you really can’t run another dedicated network cable for the Surface/Mixrack connection, then what you are describing seems to work on paper at least.

[a.malo]

I am only going to use a single NDI stream (150Mbps 1080 60p) and not always, it is more only on rare occasions. In most cases it will be gigaace and sacn (2 universes, that’s less than 5Mbps)
If we add the gigaace which is 930Mbps it would be about 1085, if I remove the NDI and control (iem, wireless,…) I take it out through the gigaace tunnel I could take it out through a gigabit link. If I send it through ports of more than 1gbps there is no bottleneck and I shouldn’t have any major problems.

according to the white papers it can be done by vlan

https://www.allen-heath.com/content/uploads/2023/11/AH-dLive-for-IT-managers.pdf

[Brian]

As you have noted, the whitepapers say it should work. That being said, real world operations can sometimes be a far cry from whitepapers and “network designs on paper”. We aren’t saying that it can’t work, we are ALL saying it’s not worth the risk/hassle to save running another network cable or two from FOH to the stage.

Besides, unless you are going to use two switches (one at FOH and one on stage) and use a faster than a 1gb link between them, you will need to run multiple network cables between the two locations which means you are saving nothing and simply adding extra complexity to the system design. As you noted yourself, the GigaAce connection alone will nearly saturate a 1gb connection. (The phrase “Giga” in GigaAce wasn’t used by accident).

So once again I would recommend running GigaAce connections normally, independant from anything else and without connecting to a switch. Then you can run all your other stuff using VLANs and likely have plenty of bandwidth on a 1gb connection to handle it all (assuming a low number of NDI feeds).

[a.malo]

I understand that there is a risk and that it is not acceptable for you. I’m not just talking about eliminating wiring, I don’t care that much about that. If I had the S series I would launch 2 cat and have redundancy. The tests that I have carried out on a couple of Ciscos, which they lent me, are that with the Link Aggregation Group I achieved redundancy by losing one of the trunks.
If you read both the switch models and my first post I talk about SFP+, to go to 2.5 and 10GB (the 1GB that I mention, which in most cases would be enough for me, is to illustrate that at no time will the switch go overloaded)
For me it is much more important to gain “logical” redundancy and not “mechanical redundancy”.

I don’t want to sound angry or rude at all, I appreciate your comments, but that’s why I was asking about your experiences with switches

[SteffenR]

As long as you are using the switches only for the GigaACE connection, there is no problem.
The issues will start with additional network traffic, even with VLAN separation.
No matter what kind of link you use between the switches and the amount of traffic in question.

So no recommendation for specific devices.

[Brian]

a.malo – first, I would say that you have changed your “goal” during this thread. You started out the conversation saying, “My intention is to use a switch between stage and foh and separate the traffic into vlans. Vlan 100 for GigaAce, 200 audio control (IEM, wireless…), 300 sACN, 400 NDI, 500 guest…. and use 2 trunk ports with Link Aggregation to have backup, and 1 trunk to control the network.” Now you are saying “I’m not just talking about eliminating wiring, I don’t care that much about that. If I had the S series I would launch 2 cat and have redundancy. The tests that I have carried out on a couple of Ciscos, which they lent me, are that with the Link Aggregation Group I achieved redundancy by losing one of the trunks.”

If you had started the conversation by saying something like, “I have CDM devices but I’m concerned about the lack of redundancy. Can I use a couple of switches (one at each end) with the GigaAce traffic on its own VLAN and also use LACP to provide redundancy?” you would have gotten a much different answer than you got. Instead IMHO, you obfuscated the goal of redundancy with all the other “traffic” you hoped to also include over the same two network cables running between the switches.

So yes, I believe you can use two switches, a dedicated GigaAce VLAN, and LACP to create a type of redundancy. However I would still strongly recommend that you don’t put any other traffic on this switch setup. I completely agree with Steffen that it’s the “additional traffic” that is going to cause issues and this has been the theme throughout our answers/recommendations.

If you insist on putting other traffic on the switch, I would at least add more network cables between the switches and set up your LACP to ensure the GigaAce traffic is being transmitted by itself without any other traffic on those connection. The additional traffic should be transmitted on another set of network cables/LACP ports entirely IMHO. I have no idea if that will prevent any problems with the GigaAce data, but it’s the least you should do to try to keep that data separate from the additional traffic.

[a.malo]

You’re right Brian, I expressed myself terribly. I took things for granted. In my head (with the rest of the information inside) it sounded good and if we add a fairly poor level of English… I’m sorry.
Thanks for your comment.
I think it will be the best option to achieve “automatic redundancy” in case of failure both in dlive and with the rest of the systems involved.
What do you think of the Zyxel XGS1010-12, it has 3 10G ports that will be very useful for the other vlans and I will use 2 1Gb ports for GigaAce.
Would it be possible to do the same configuration with SQ and a Dx168, forcing their port to 100Mb, and get redundancy?

[Brian]

a.male – no worries! It’s hard to ask technical questions sometimes and fully explain what your ultimate goal is. At least we got to the root of your goals in the end. I hope it helps!

Please report back your results if you do try this. I can’t say that many people are doing this and you are certainly outside of A&H ‘supported” network methods. It’s always good to hear what works and what doesn’t work for people.

[Author]

Posts