Open source advocates plead for greater cooperation in telco SDN

Last week’s Open Networking Summit proved to be a major coming-out party for the new-look Open Networking Foundation (ONF), which was formed from the October merger of the old ONF with ON.Lab.

The merger is yet to close formally, but the plans are being discussed openly. The new ONF is an increasingly powerful body, leading open source initiatives in some of the most strategic, and challenging, areas of SDN (software-defined networking) for carrier networks.

The Summit saw the ONF discuss its new structure and unveil the latest release of its
ONOS (Open Network Operating System), as well as discussing the roadmap for two other crucial elements of the software-defined telco, the SDN controller and the virtualized central office.

The merger has helped drag the ONF back into the carrier SDN limelight after a period when its efforts were competing with too many other open initiatives. Although it was a pioneer of SDN, and developed the OpenFlow platform, it had become overshadowed by ON.Lab itself and OpenDaylight. As a merged entity, it gains two very important projects which originated in ON.Lab and are now run autonomously under the auspices of the Linux Foundation – ONOS and CORD (Central Office Re-architected as a Datacenter).

Together with OpenFlow, these present an open source challenge to an alternative carrier SDN approach, that of the traditional standards organizations as epitomized by ETSI and its NFV working group. While NFV has been very successful in setting base standards for telco network virtualization, increasingly operators are looking to open source platforms as they formulate management and orchestration (MANO) strategies for virtualized network functions (VNFs), and as they look to implement SDN alongside NFV.

The main attractions of the open source approach are the relatively low investment, the strong innovation base, and in many cases, easier deployability than ETSI approaches – though most carriers agree that this comes at the expense of some telco-specific functionality. A hybrid approach will appeal to many telcos in the second wave of SDN/NFV, but for now, many are looking to open source to give them a relatively simple springboard.

ON.Lab founder, and now ONF head, Guru Parulkar said: “We believe that standards based on widely adopted open source projects such as ODL, ONOS, OPNFV, and CORD can be more widely and easily implemented within the industry. This emphasis on open source, supported by ON.Lab’s successful software development, will guide ONF’s ongoing standards work, including updates to OpenFlow.”

He insisted, when the merger was announced last fall, that open source will complement traditional standards (with the unspoken implication that this would require more cooperation from the standards bodies in future). He said: “You need end-to-end compatibility among these different platforms and that requires a certain amount of standards, at least for APIs or interfaces. Open source has been gaining a lot of influence but it does not eliminate the need for standards.”

Margaret Chiosi, formerly at AT&T and now VP of open ecosystems at Huawei US, raised similar issues in her Summit speech. “You have all these open source pieces. They are great initial pieces, but you can’t just clean it up and run it, because it’s not complete,” she said.

She cited SDN controllers, and the two main open projects in this critical area – ONOS and OpenDaylight. Huawei participates in both, but has created its own controller, combining the best of both, it claims. Other vendors will take a similar approach, with the benefit for operators being a finished, optimized and supported product – but the downside for industry being fragmentation and vendor lock-in. Huawei claims multivendor openness – in a project to interconnect China Unicom’s data centers with its Agile Controller SDN, it said, it works with other suppliers’ controllers and routers via common APIs.

That is how Chiosi believes the situation will resolve itself, with vendors providing commercial products within open source frameworks and using open APIs. She believes those initial closed offerings will “evolve from closed to partially open to completely open, if possible – and then operators can start picking best of breeds and mix and match, using common APIs.”

But the history of the telecoms industry, and the commercial imperatives of the large vendors, suggests they will not be a hurry to move along that path to openness, and the operators will need to make powerful common cause to force them to do so.

It is clear, however, that the telecoms network and its value chain will never be the same again, and that open source bodies will have significant ability to shape the way they develop. So all eyes were on the Open Networking Summit in Santa Clara. There were interesting demonstrations based on ONOS, including a packet optical solution shown off by Japan’s NTT, but perhaps the most buzz was around CORD.

Timon Sloane, VP of standards and membership at the ONF, said that 70% of operators are now planning to support CORD (or its mobile variant, M-CORD). He told LightReading that carriers are attracted to the solution because it is a top-to-bottom, cloud-native solution, not one which has been cobbled together from existing parts adapted for the cloud. That makes it very easily deployable.

He said: “CORD is the total stack: OpenStack plus extensions to make it work in the operator environment; a full DevOps that everyone is after in the cloud world; networking is in there as a critical component, along with others; and service creation, which is where XOS fits in … it is what provides the microservice capabilities. And then a whole pile of infrastructure stuff. You can leverage all of that and get started right away.”

The ONF also announced a new Open Innovation Pipeline, designed to make it easier for operators to join the body and participate in its processes. It will focus on specific use cases, harnessing the ONF’s various platforms and creating targeted packaged solutions. These combine the appropriate CORD variant, such as M-CORD, with specific VNF and services, and operators can then download and install these packages.

With ONOS, the approach is the same – to use it as the foundation for individual combinations of technologies for specific scenarios, saving the operator a great deal of integration work. For ONOS, the first two solutions focus on packet optical and spine-leaf.

The new release of ONOS is more focused than its predecessors on what the ONF calls ‘incremental SDN’, where the technology is added into legacy or brownfield networks rather than being deployed from scratch. In the past, ONOS has been stronger in pure-play SDN, but it is clear that for many operators, coexistence with legacy networks will be vital to the economics. So ONOS can now manage white-box and legacy equipment together.

However, there are still key changes that are needed in open source processes to ensure their success in the telco industry, according to Linux Foundation boss Jim Zemlin. Zemlin called for greater harmonization across open source efforts to reduce fragmentation and accelerate uptake – in particular, he said projects addressing similar topics must collaborate better to reduce overlap; open source efforts must harmonize better with standards bodies; and there should be education for companies in how to work within the open source community.

“We need to harmonize these projects up and down the stack. If we have three or four different projects at each level, how do we harmonize” at least at the level of common APIs and languages, he asked.

The ONF’s Parulkar said his body was seeking to address some of these issues by integrating different open source technologies to make them easily deployable; and, through open data and information models, making it “easy for vendors to plug innovation into this pipeline with significantly reduced R&D and lower time to market”.

It is key, he said in his address, that vendors recognize the value proposition. He acknowledged that the industry currently suffers from “misaligned incentives”, since the main beneficiaries of open source are the operators, but most of them lack development resources to bring products to market, while vendors see the disaggregation of software and hardware as “very disruptive to their existing business”.

Another open source group showing off its latest jewels in Santa Clara was the OPNFV (Open Platform for NFV) project group. Its objective is to create an integrated stack that makes NFV easy to deploy and run, and the latest release of its software, Danube, promises a unified approach to testing and roll-out. This is the fourth release of the software, and the project’s director, Heather Kirksey, said it was making OPNFV “the engine room for NFV, delivering performance and ease of use for network operators and solidifying its collaborative processes with upstream organizations.”

Danube helps improve NFV performance and supports integration between the NFV Infrastructure/Virtual Infrastructure Manager (NFVi/VIM) and the Open-Orchestration (OPEN-O) platform. OPEN-O was itself involved in the ongoing consolidation of open source efforts in this field, when it merged its efforts with those of the AT&T-initiated ECOMP earlier this year, to form the ONAP (Open Network Automation Platform) group.

With the new OPNFV capabilities, members can test the MANO stack with various NFVi stacks, as well as testing use cases which require the entire stack, using real world VNF onboarding and provisioning. The next step from that could be standard approaches to VNF onboarding, to simplify the process further.

The way that open source initiatives are taking precedence over those of standards bodies is highlighted in the OPNFV priorities – it is working first with ONAP, and only then looking to form ties with the other main MANO group, Open Source MANO (OSM). Despite its openness, this has ties in ETSI NFV, and so is rather less naturally integrated with the Linux-rooted technologies.

Danube also includes work done with to improve integration with OpenStack, via a cooperation with the OpenStack Gluon project. This will help operators to use SDN to carry out a larger proportion of network control functions. Other efforts include progress in automated testing, and ongoing work to improve support for IPv6, BGP virtual private networking and others. Many expect to see commercial OPNFV deployments by the end of this year.

The next OPNFV release, labeled Euphrates, will expand on the MANO support via collaboration with ONAP to integrate the ECOMP and Open-O code base as it matures. It will also focus on VNF interoperability and application enablement.

Meanwhile, ONAP itself announced several new members at the event, including Microsoft, Indian operator Reliance Jio, and Ciena. The latter has been advanced in developing SDN/NFV orchestration, with its Blue Planet platform, and has adopted a microservices approach, so its contributions will be highly relevant to the ONAP objectives. Ciena already contributes to the ONF’s technologies, and has produced a hardened version of ONOS for CORD.

A diversity of new members is crucial to create a snowball effect for any open source project. Both ONAP and Open Source MANO will be keen to lure major operators and vendors which remain on the fence, with Verizon being an obvious one. The telco’s VP of global technology and supplier strategy, Srinivasa Kalapala, told Light Reading that Verizon is doing due diligence on both MANO options, but has concerns about both. On the OSM side, he questions whether the technology is more than a VNF orchestrator, rather than a full service management platform; while “the concern we have with ONAP is whether it is truly open. How many groups are contributing?”.