How to Future Proof White-Box Switches in Carrier Networks
Global carriers, from Telefónica to AT&T to NTT, are making major investments in their networks. Spurred by the new services made possible by 5G, networks are being revamped to support more connections, to operate at higher speeds and to leverage new resources like edge computing. This expansion pushes the service edge much closer to the customer, which creates a need for a major expansion of the number of routers. Like with any other technology, an increase in the number of access capacity points requires a reduction in cost.
There are a number of factors that keep legacy routers from meeting this cost reduction requirement. Their basic architecture has not changed for 20 years. Carrier-grade routers have been closed appliances, built with custom silicon only and running the vendor’s own software. This locks an operator into the router vendor and limits their ability to innovate.
Advances in commercial ASICs have extended the benefits of open networking to carrier grade routing. Operators are breaking free from vendor lock-in using a combination of white box switches, disaggregated software, and open-source software. They’ve embraced this new approach and are testing software-based disaggregated routers as a way to reduce costs, boost bandwidth, and support new functionality to improve their network economics.
As a result, white-box switches have emerged as an affordable, high-performance way for service providers to drive greater scale and capacity in their networks. And sales of white-box switches and routers have grown in many areas over the years. In 2018 AT&T announced plans to install more than 60,000 white box routers across its network over the next several years, as part of a radical realignment of its network architecture to support 5G.
However, service providers face a challenge: how to future-proof white box switches so that they can be deployed in carrier networks for several years without needing to be replaced. Installation costs can be significant, as can cost of new and additional switches. However, there are ways to maximize the lifecycle of white-box switches, to minimize cost and to prevent disruption to service provider networks. Here are a few things to consider.
Solve the CPU Bottleneck Issue
In service provider networks, the data plane´s forwarding needs can be addressed by a very cost-effective white-box switch, but the potential issue there is that the white-box switch doesn’t have a lot of compute capacity for control plane functionality. Carrier-grade routing is a very CPU and memory intensive activity, because these are complex networks. This CPU issue can also become more prominent as routing tables get bigger over time, which can often be the case as more devices and people join the network.
Service providers need to realize that they do not need to have all the routing protocols running on a white box. The best place to be able to scale processing and memory, and do so in an affordable manner, is in the cloud, and if the solution is architected correctly it will work fine.
Move the Control Plane to the Cloud
Once the vast majority of the control plane runs in the cloud, an agent that runs on the white-box switch can take care of the communications between the cloud and the ASIC. There are some things that have to be done locally in hardware, particularly for failover, because they have to be done very quickly. However, building a routing table is not anywhere near as time-sensitive, because updates are only being sent back and forth when the network changes.
Another big advantage is that service providers can spread the cloud elements over multiple data centers for redundancy, reliability and high availability – these are all the things that customers really like about carrier-grade operations.
Leverage the Right Networking Protocols
IP and Ethernet are going to be critically important to the transport networks that drive the migration to 5G. For any given point in the network, one needs a low-cost router that’s going to be able to support multiple virtual routers. With lots of different services running on white-box switches, service providers need to be able to do some segmentation. Carriers can lower the cost of the network by expanding virtualization in order to address the new opportunities from 5G.
Use Virtualization to Drive Service Agility
Virtualization has many advantages, not just because it can reduce costs, but it also improves service agility. And improved service agility means service providers can meet their customers’ needs much faster, and that can drive revenue for their business. For example, service providers may be seeing a big spike in gaming and video conferencing traffic. During the pandemic, more people are stuck at home, so they are using their (service provider) network more for entertainment and for working from home. If anything, the situation with COVID-19 has underscored how dependent we are on networks in order to be able to continue to operate in this environment.
Embrace the New Ecosystem
To build transport networks that effectively deliver 5G, we need to look beyond a one-size-fits all approach. What's going to work well in a dense urban core may not work in a suburban or rural area. However, these differences are okay – 5G was designed to have a lot of flexibility in how you do it. Industry organizations such as the Telecom Infra Project (TIP) are trying to build an ecosystem that will help manage different network options, factoring in the role of white-box switches. On the routing side they have two major projects: disaggregated cell-site gateways (DCSG) and disaggregated open routers (DOR), which will be closer to the core router, with bigger, faster interfaces leveraging different ASICs that can power higher-speed ports.
In all, service providers are going through a major shift in how bandwidth is being used in their networks. Many have been able to respond to the changes, but they've used up a lot of their capacity in the process. Going forward, service providers need the footprint that they were planning on for 5G to be able to deliver more bandwidth and services, so architectural changes are needed in how routing is done. We believe the combination of white-box switches with disaggregated and open software to enable the creation of multiple virtual routers per device is going to play a major role in the new generation of transport networks.
About The Author:
José Miguel Pulido
José Miguel is Sales CTO at Volta Networks. Prior to Volta, JM co-founded Promofarma.com, the leading consumer health and beauty e-commerce marketplace in Spain, where he built and delivered a full-fledged marketplace platform on the AWS cloud platform, serving up to +1M daily visitors. Prior to Promofarma.com, he was one of the first members of the engineering team at RouteScience in Silicon Valley where he contributed to the creation of route optimization, a technology that exploited path diversity in IP networks to increase application availability, and that ultimately resulted in the company’s acquisition by Avaya.
He has an Eng.D. in Electrical Engineering from Stanford University, an M.S. in Computer Engineering from the University of California, Irvine, and a Telecommunications degree by UPC, Barcelona Tech. JM was a laCaixa fellow, a Balsells fellow and holds a portfolio of patents in networking technologies.