DISAGGREGATED PON: ONF HAS THE SOLUTION.Created April 9, 2023
SDN (Software Defined Networking) is not a new concept, but how to apply it to optical networks? As part of its work in the disaggregation of both fixed and mobile networks, the Open Networking Foundation (ONF) has come up with SEBA, a lightweight platform which supports a multitude of virtualised access technologies at the edge of the carrier network including PON, and VOLTHA (Virtual OLT Hardware Abstraction), a software solution which does exactly as the name suggests. Optical Connections editor Peter Dykes spoke to ONF’s general manager Timon Sloane to find out more about ONF and the development of these projects.
PD: How was ONF started and why the VOLTHA project?
TS: ONF is a pretty unique beast, it’s non-profit and it’s been around for more than 10 years. It came out of Stanford and Berkeley universities with collaboration from operators like AT&T, Google and Microsoft. Our mission from day one really started with SDN, which was a novel concept when it came out of ONF. We were standardising OpenFlow protocol and the vision was that networking devices had become too large, vertically integrated, expensive, high margin devices. At the time, Cisco routers were million-dollar devices, the size of refrigerators and the view was that cloud operators had started to build things that separated the software and the hardware, meaning they could procure simpler hardware devices and start to run the software in the Cloud.
That’s really what SDN was about. It occurred to our community that these techniques could be applied to the PON access network. PON devices are pretty complicated, expensive chassis, a lot of slots and vertically integrated, so what VOLTHA and SEBA actually do is take that PON chassis, and instead of making it a big, complicated device that’s configured with rules, they separate the hardware and software.
PON has special transceivers and everything needed for a multidrop environment, but take protocol elements like configuring a subscriber, DHCP, or assigning an IP address to a subscriber. In the old world, these were functions that were built into that complicated chassis, and if an operator tried to bring on a second chassis from a different vendor, that simple feature behaved slightly differently in each chassis. That creates multiple complications, especially up at the OSS and BSS layer, where minor changes can have a ripple effect. What VOLTHA does by separating things and running in the cloud and in our open-source stack, the hardware layer is simplified into a 1U pizza box device that houses the transceivers and the simple devices.
As much functionality as possible is abstracted out, making it easy to swap out the pizza boxes and get consistent behaviour across a diverse range of devices. Furthermore, today in the PON world, these big chassis are tied to the devices that run in the home. Customers buy a device from provider X, and have to buy a home gateway from the same provider, and the vendors make a lot of margin on the home gateway device, if you get locked in, but we’ve broken that as well.
The key protocol is OMCI [ONT Management Control Interface] and we have an open implementation of it that can be offered on home gateways. It’s software that runs in the cloud on the central office side, making it possible to mix and match gateways devices as well as vendors. So it’s about a couple of things. First, it’s about putting the operators in control of the network, rather than selecting a vendor and being beholden to their roadmap. Operators can choose different devices in their supply chain and change them on the fly without disrupting their whole network architecture. It’s also about cost, because you’re driving more towards whitebox economics and finally, it’s the about the cloud. It’s about really being able to embrace the best of cloud scale-out, cloud tooling and running as much as you can in a cloud-like environment on a generic compute, rather than on proprietary devices.
PD: Is VOLTHA gaining traction in the industry?
TS: Yes. It’s being embraced and deployed by tier one operators and the community of vendors that are building compatible devices continues to grow. Now that it’s being deployed at scale, the feedback from operators is very positive. It really is achieving everything we thought it would on all fronts, especially with the supply chain problems we’ve had over the last year. It was a perfect example when it became hard to get certain devices, but operators were able to get other devices and substitute them in with relative ease. And, it seems, there were considerable cost savings in the order of, say 25% to 35%, as well as giving operators a cloud native environment for the next generation central office, versus legacy architecture.
PD: Is scalability an issue?
TS: No, it’s not. In many ways, VOLTHA is more scalable than the most. It’s pushing a lot of this intelligence to the edge and you can replicate it as much as you want when you need to, so it’s very easy to scale. It’s being deployed by the likes of Deutsche Telekom and Jio and these are among the biggest operators in the world, so this is not a scale problem, if anything, it’s the major operators recognising its value that’s driving adoption. Scale is, in many ways, the least of our concerns.
PD: There are new PON architectures coming along such as XGS PON. Do they present any particular challenges for the VOLTHA project?
TS: Those are opportunities, not challenges. We already support XGS PON, which is already being deployed. We also support Combo PON, which allows you to mix and match ports off individual devices. We’ve been able to support it faster than most because disaggregation gives you a much more nimble, faster architecture. Operators are able to source new OLT’s that can support XGS PON more rapidly, because they don’t have to swap out their whole access architecture to do it. So if anything, XGS PON demonstrates the value and importance of leveraging an architecture like this.
PD: What’s the relationship between SEBA and VOLTHA?
TS: The way it works is VOLTHA is the software project where there are open-source artefacts you can download and run and there’s an ecosystem of hardware that’s up and running and deployed. SEBA was two things. There’s the reference design, which is an architectural document that you can find on our website and which defines how you can plug VOLTHA into the larger end and put together all the pieces, for example. SEBA is and was an open-source project implementing a lot of the pieces but that work is not as active, although there are open repositories and components available. What we’re really finding is that VOLTHA is the thing that’s taking the world by storm. People are integrating it into their own versions of SEBA-like architectures in their own ways using other tooling or other vendor components. But VOLTHA is really the piece that as a software project, is where the traction is.
PD: Is VOLTHA a completed PD project, or is there more to come?
TS: It’s definitely active and it’s reached a certain level of maturity, but that doesn’t mean it’s finished by any means. There are active groups at work, and there’s an active Technical Steering team that guides the project. Members gather weekly and talk about new releases like the one coming out this month with new feature functionality. We test the releases against a suite of different hardware devices and load environments, and there’s a whole test and CI/CD [Continuous Integration/Continuous Delivery] infrastructure that’s backing the project to support all the project consumers. There is also an area board called the Broadband Area Board at ONF, that guides the project from a business perspective and the direction of our open broadband work, which is predominantly built in SEBA. A set of both interested operators and vendors that really care about this space and are betting their businesses on it, are part of the project. We are actively looking at, working with and demonstrating more points of alignment with the Broadband Forum and the Cloud CO architecture, so that VOLTHA can be plugged into Cloud CO, so you can really use this next generation of disaggregated PON, but still work with the rest of the Cloud CO architecture. That integration was demonstrated at NetworkX in October last year.
There’s more work to come there, but I think there’s a lot of good and important work to be had there as well. I think that there’s a ton of bandwidth in broadband access and cellular backhaul and it’s the unspoken hero of the last two years, because while mobile and 5G gets all the press, broadband and mobile backhaul is actually carrying all the data. Suddenly we’re seeing this spike in data growth, huge investments and big government backing in investments, and VOLTHA has run into the epicentre of it all, so it’s a very, very important project.
PD: Thank you.
Timon Sloane, General Manager, ONF
This article was first published in Optical Connections magazine, Spring 2023.