Communications service providers (CSPs) globally are being challenged to transform their networks to meet the demands of the 4th Industrial Revolution and move from connecting people to connecting 50 billion devices. Traditional networks built with expensive proprietary platforms from incumbent vendors are unable to quickly and cost-effectively scale to meet bandwidth demand, while keeping CSPs locked into a single supplier. New open architecture business models have emerged as a disruptive force to enable CSPs to add network capacity and accelerate service innovation, while eliminating vendor lock-in, reducing complexity and lowering costs.

From Industry Standards to Open Source

Traditionally, the telecom industry was driven by large standards bodies such as 3GPP, ATIS (Alliance for Telecommunications Industry Solutions), the ITU (International Telecommunications Union), ETSI (European Telecommunications Standards Institute), which defined standards for everything in telecom right down to the telephone poles. These standards bodies have had dedicated individuals from across the industry working together for years to develop industry standards that are comprehensive to meet stringent requirements across many use cases.

However, increasing pressure from over-the-top players and exploding network traffic has the entire industry re-thinking how it should operate and looking to the enterprise model of using virtualization, disaggregation and DevOps in order to accelerate innovation. Rather than spend years to develop industry-wide standards, enterprise has more typically worked with de facto standards, often through open source projects such as Linux.

Enter OCP, TIP, xRAN, CORD – The Rise of Open Source Projects in Telecom

The open source model offers some benefits over a standards-development approach that meets CSPs’ requirements for accelerated innovation. The idea behind a project driven DevOps methodology is to develop quickly, fail fast, and re-iterate quickly. In this approach, CSPs can test drive new solutions quickly and course correct, rather than wait years for an industry standard to be finalized.

A number of open source projects have emerged for the telecom industry creating a shifting landscape of options. CSPs and the vendor community are trying to figure out how these emerging open architectures relate to each other, where there are overlaps and gaps, potential alignment with industry standards, and perhaps most important: which open source projects will achieve mass adoption. 

The Open Compute Project (OCP) is one example of an enterprise/web initiative that is being embraced by the telecom industry. OCP was founded by Facebook to open source its hardware designs for the data center to promote a broader vendor ecosystem, with customization options and cost savings. Radisys worked with CSP and manufacturer partners to adapt the OCP base model to meet telecom requirements and bring OpenRack concepts and benefits into the service provider networks. The resulting CG-OpenRack-19 framework specification gained OCP-ACCEPTED status and allows CSPs to leverage a scalable, carrier-grade rack-level system that integrates high performance compute, storage and networking in a standard rack.

The Telecom Infra Project (TIP) is another open source consortium with a mandate to deliver open hardware for telecom network infrastructure. A sister project to OCP, TIP’s initial focus was on simplifying and disaggregating the Radio Access Network (RAN) of the network, but has expanded to include projects for wireless backhaul and for the network core and management.

Another approach to disaggregating the RAN is xRAN. xRAN emerged as an open source project to challenge proprietary RAN infrastructure with an extensible software-based solution that can rapidly respond to real-time user needs. One of the key attributes of open source projects such as xRAN is that they’re not trying to solve everything. They have a specific idea and focus, and thus are able to move quickly to bring the idea to fruition. In this case, xRAN was founded to extend the RAN with specific north and southbound SDN interfaces.

CORD (Central Office Re-architected as a Datacenter) integrates multiple open source projects focused on addressing the needs of service providers’ mobile, residential and enterprise customers. It was created to develop an open reference architecture that will allow CSPs to leverage SDN, NFV and cloud technologies to transform their traditional Central Offices into virtualized telco data centers. A CORD architecture is comprised of open source software and hardware; for example, it could be built leveraging an OCP-based CG-OpenRack-19 platform or other COTS servers. CORD takes a pure DevOps approach to iterate quickly.

Consolidation and Fragmentation?

By their very nature, open source projects can be formed very quickly. All that is needed is an idea, and for then a few companies to come together to begin trying that idea out. However, this has also led to an ecosystem of open source projects overlapping in scope or leaving significant gaps in implementation. At the same time, traditional standards bodies are recognizing that their rigorous standards process is moving to slow to keep up with industry disruption. Standards organizations such as the Metro Ethernet Forum (now just MEF) and the Broadband Forum are reinventing themselves to stay relevant and are shifting from workgroup driven to project-driven approach without hierarchy.

There is beginning to be some consolidation in the industry. ONF (the Open Networking Forum), which developed OpenFlow, and ON.Lab, which launched CORD, recently completed a merger under the ONF brand to drive SDN adoption. And CORD has fully embraced xRAN from a functional point of view in its architecture as it delivers a well-defined SDN interface into the RAN for CORD. OCP-based hardware can be deployed in a CORD architecture or COTS servers can be used.

However, it still unclear which open projects will be deployed and where fragmentation may occur. TIP is taking a new approach to telecom networks and disaggregation, but it is not necessarily trying to fit in with other open architecture models.  While TIP and OCP are backed by many of the same people, it remains to be seen if they will align. Open19, initiated by LinkedIn, is another new open source project that is enterprise centric, but also very similar to OCP’s CG-OpenRack-19. Time will tell if these two groups will converge, but regardless, the convergence of ideas and methods between enterprise, web, and telecom spaces is accelerating.  The Broadband Forum is working to develop an open reference architecture that could be complementary to CORD or could be competitive.

From Open “Science” Project to Real-World Deployment

In order for these projects to go from the lab to real-world deployments, strong backing from tier-one CSPs, in addition to the vendor community, is needed. For example on the orchestration front, ONAP, with support from AT&T and other service providers joining such as Vodafone, seems to be gaining critical mass as the de facto architecture.

The open projects described above have a healthy balance of CSPs and small and large vendors to help create the market. Moving from vendor-locked proprietary solutions to open telecom solutions built with a variety of open source hardware and software components requires a new business model where agile systems integrators aka “middlemen” can fill the gap left by single-solution vendors. These new systems integrators must be able to balance both:

  • Neutrality, leveraging open source components from multiple vendors to deliver a turnkey solution that meets the CSPs’ requirements rather than push a specific company offering; and
  • Having a stake in the outcome, where their company survival depends on the success of the open project. A systems integrator that is not actively engaged in an open source project, won’t care if it succeeds or not.

Successful real-world deployments of solutions from CORD, TIP, OCP, xRAN and more will require the backing of tier-one service providers, a thriving vendor community and engaged systems integrators. With all these pieces in place, the Industrial 4.0 Revolution will thrive in an open source landscape.