INFN: Reflections on the TIP announcement

Today Infinera announced their participation in the Telecom Infra Project (TIP). The TIP, lead by Facebook, is a consortium of telecom service providers and equipment manufacturers that seeks to open up interoperability between networking infrastructure:

Facebook hopes to create a more simplified network architecture using modular components that combine hardware and software. This would replace the expensive, proprietary gear that underlies telecommunications networks today.

You can read more about the TIP (from a non-technical perspective) here:

I think the timing on Infinera making this announcement was calculated (and shrewd). Here’s why:

  1. First, the TIP means to create a system of open standards and modularity so that the different parts of the telecom ecosystem can be ‘interchanged’ more easily. The three main parts to the ecosystem are access, backhaul and core.

  2. Traditionally these different parts aren’t as easy to integrate as one might think. That is because system houses like Cisco and Erikson have built up their own proprietary mechanism for interoperability. That is not to say system providers don’t have the ability to co-mingle with other vendor solutions. It’s just not a plug and play approach that can work out of the box. Rather, it requires complicated configuration management and field tests, and it becomes a real barrier to having other system vendors participate in your network. This has been long-running obstacle for Infinera when trying to get into established networks.

  3. Infinera has elected to participate in each of the subcommittee projects in the ecosystem (access, backhaul and core). That means Infinera will be involved in developing interoperability standards and design references across the ecosystem. This is of big interest to them. It means they’ll be able to get their gear in more places with less friction, and it will work with their current integration mechanisms. The list of sub-committee projects are located here:

  4. The deadline to participate in these projects is tomorrow. And if you’re not one of the members developing these standards - their rules will be yours to live by as more telecoms sign up to an open standards process. You can be sure any structural advantage provided by the participating vendors (such as Infinera’s PIC) will work seamlessly with these design standards (read: capacity, speed and instant bandwidth as a differentiator - with no barriers to get it into your network). From the Forbe’s article:

Even those who don’t play along are likely to adopt the principles of modularity and flexibility through software-defined architectures and buying gear from companies that support true interoperability. That’s why vendors like Cisco and others have been leery about the threat of open source switches, Open Flow, and other elements that have been forcing the network world to open up and embrace interoperability between vendors and their gear.

The system vendors participating in defining the new standards are Nokia, Juniper, ADVA Optical and now Infinera. The component vendors in the mix are Lumentum and Acacia. Equinix and Intel are also involved. The full list of participants are located here:

Exciting times.



Thanks for this, Kevin. Very interesting.

One question: the article mentioned something to the effect of opening up competition…do you have any concerns that this could adversely impact Infinera as time goes on?


I actually think this form of competition will make things better for Infinera.

If the industry adopts open standards Infinera will be able to get into any network. That is great. But the reciprocal of that is also true: so will everyone else.

But here’s where the benefits of vertical integration and the PIC come into play.

#1: Infinera is vertically integrated. Vertical integration means Infinera is in control of their cost structure, and they are the only company who has this control. Everyone else has 2 or 3 different tiers of suppliers to go through. Margins are going to be pressured for those companies.

#2: The PIC is an IP asset. With the new PIC coming out Infinera can pre-deploy a 2.4TB capable device into a network and have the customer turn it on when needed. They can also sell that device at an initial lower capacity (while being competitive because they control their costs per #1) and have the extra capacity built in and ready go.

To be truly competitive companies will need to move towards #1 and #2 above. Infinera will have a long window before that happens. Plus, if Infinera keeps up R&D at 20% per annum it will be a long time before competitors truly catch up.

Just my thoughts.



Thanks Kevin.

I agree with you…exciting times!


I know you just agreed with me, but you shouldn’t just take my word for it. Here is what Tom Fallon had to say about SDN/NFV in the last conference call transcript:…

And I think we – this summer you’ll see us make some more announcements, and in August you’ll hear a whole review of it. And I think, actually, SDN positions us very well to sell more optical gear. I think I’ve said this before, and I mean it. Our architecture was built to be SDN-ready. It’s not because we saw SDN but we saw what an intelligent optical network can do. And I do believe that the more pervasive SDN becomes the more applicable Infinera’s architecture is within that.

Unlike the router guys or the switch guys, our stuff’s not going to be subject to NFV. People aren’t going to virtualize what we do. A photon going from A to B takes what we do, and no software’s going to replace that. All of these things to me are exciting because it makes the optical layer more strategic. Having a giant pool, an infinite pool of intelligent bandwidth is what enables SDN and NFV to come to life. If you have constrained optical capacity, you can’t have a responsive, adaptive network. We have the most unconstrained optical capacity, scale of capacity, because of our PIC design. It pumps me up.



Yes, I have been reading a bit about it…not just now, but over the last several months. I’m not thoroughly versed in the technical aspects of it, but I definitely get the idea.
thanks Kevin.