As 2019 winds down, Volta has been asked several times about our thoughts on what was significant in 2019 and our predictions for 2020.
In 2019, we saw the tipping point in service providers moving to open networking from legacy infrastructure.
The key was the availability of carrier-grade switching ASICs from vendors like Broadcom. A just a couple of years ago, it was common for service providers to have a couple of white-box switches in the lab and it seemed like they weren’t going to make it out. However, the Dune family of ASICs from Broadcom changed that. Dune, more formally known as the StrataDNX family, includes Qumran and Jericho chips which have the ability to scale both tables and buffering with external lookup engines or memory. Qumran is also available in smaller bandwidth at a lower cost.
This enabled service providers to give Original Design Manufacturers (ODMs) specifications for white box switches tailored to their network applications. One example is the TIP’s disaggregated cell site gateway project with Telefonica, TIM Brasil, and Vodafone – who represent roughly 15% of global wireless subscribers. They are in active tests now and expect to be deployed within the year. A separate project from AT&T specified virtually the same hardware for their small cell site gateways.
This trend toward open networking will accelerate with Tier-2 service providers as they follow the lead of the Tier-1 providers. However, only the largest service providers will be able to use open source because there is not enough support available in the market. The large service providers will have the resources and human power to support it themselves, but other smaller providers will not be able to do this. The only “community” that has really developed is Free Range Routing.
Looking forward to 2020, disaggregated router architectures will disrupt the service provider router market as CapEx shifts away from legacy routers to investments in greenfield buildouts that set the stage for new revenue opportunities.
Service providers will begin significant field deployments of open networking as the only cost-effective approach for greenfield applications, such as 5G transport networks. (In a recent blog post, the Telecom Infra Project DCSG sub-group announced field trials by Telefonica and Vodafone.)
For service providers, major technology shifts that enable new services take a long time but deliver big returns. We believe it will be the same for 5G – especially for enterprise services. It’s like the old saying “If you build it, they will come.” Service providers need to get to a certain critical mass of 5G service availability before they will start to see significant 5G service revenue. But enterprise customers need time to figure out how they are going to use new 5G services – such as network slicing – before it really starts to take off and service providers see the revenues. It will be easy to question how quickly the revenues will come. But make no mistake the train has left the station!