"Audio Networking Is Like Politics"
Now there’s a sentence guaranteed to alienate an audience! Bear with me though as I’m not going to get too far into the politics part of this simile. I read two posts on AoIP recently: “Time To Close The Chapter On AVB For The Pro AV Industry” by Roland Hemming of RH Consulting and “Cracking the Audio Over IP Nut” by Mikael Vest of DAD and the best sentence I could come up with, to describe my conclusion after reading them was the sentence at the top of this page.
How Is Audio Networking Like Politics?
The similarities I’m pointing out here could be that the development of AoIP involves changing behaviours across a large population who already have a way of doing their work - Audio already gets moved from place to place and computer networks already exist. There are advantages in bringing audio distribution and networks together, some types of work have immediate and obvious needs which are met with AoIP, others less so, some people are highly motivated to see AoIP implemented widely, others see it as having little relevance to them. All of those statements could, with a little re-working, be made about political issues.
But none of those is the point I wanted to make. The conclusion I reached was this - Whether you’re designing an AoIP system or a political system, it works best if everyone likes the same one.
But which one? Should we base our solutions on where we are, or where we want to get to? Idealistic vs pragmatic? Should we compromise our solutions based on what is easiest or quickest to achieve? Should we work within the rules or if the rules get in the way do we need to rewrite them?
In "Time To Close The Chapter On AVB For The Pro AV Industry” Roland Hemming suggests that many thought that the fact that AVB/TSN is an IEEE standard and has the backing of large companies like Cisco would somehow guarantee its success in spite of the changes in the network hardware such a change would necessitate, while in reality it appears that the pragmatic, quick to implement model created by Audinate has allowed Dante to penetrate the market so quickly, and crucially so widely, that the head start they have achieved seems, in the short to medium term, unstoppable. This graph from Roland's article shows the current numbers of products supporting each protocol:
Has AES67 Stolen AVB/TSN's Status As A True Standard?
In his piece “Cracking the Audio Over IP Nut” Mikael Vest suggested that one of the most attractive aspects of AVB - its status as a standard, has been somewhat negated by AES67, with its similar status as a standard but one designed from the ground up to be compatible with Layer 3 networks as they exist today, and as a result able to be implemented immediately. AES67 has already been widely adopted as an interconnection “safety net” for AoIP and with the introduction of SMPTE 2110 to the video world, if AVB/TSN is going to take over networking - and in spite of its difficulties in terms of propagation I personally think it has a great deal to offer - that takeover has been pushed further into the future by the explosive success of Dante and AES67.
There Is A Place For Closed Systems
The AoIP landscape is full of very capable, alternative systems. RAVENNA is popular in broadcast and while not enjoying the same uptake as Dante, is open, effective and in some areas is the dominant protocol. Some systems are closed by design. Some of the Layer 1 systems popular in live sound favour performance over network compatibility and systems like Soundgrid seek to differentiate themselves by remaining strictly proprietary but offer additional features such as the ability to offload audio processing.
However the odd situation AVB finds itself in is that it is an open IEEE standard which is best used within the same manufacturer's product line. For example I know that MOTU interfaces are a joy to use over AVB/TSN but I don't know whether I can use them successfully with Presonus AVB gear.
Short Vs Long Term
Which brings me back to my original comparison between AoIP and politics. Politicians work in the short term. If presented with a choice between two ways of fixing something and one is good enough, is quick and uses existing infrastructure, and the other fixes a fundamental issue but needs new infrastructure and takes longer than the next election... Well, we know what happens.
Both of the articles referenced in this post are well worth a read, to find out more follow the links below:
I needed an example of someone who is trying to address infrastructure issues by building quick fix solutions based on the way the world is in 2017 while also trying to rewrite the rule book. The best example I could come up with was Elon Musk. To move people around California he is offering the Tesla electric car, which works on normal roads and meets the brief of offering an alternative to internal combustion engine based transport but he's also developing Hyperloop - a very ambitious 700mph futuristic train system. It's potentially faster but definitely needs new infrastructure! This example doesn't hold up to close scrutiny, replacing network switches is going to happen pretty easily, building hundreds of miles of massive tubes is less of a certainty but it does illustrate the fundamental infrastructure issue which is common to transport and to AoIP.