While some greenfield players are able to immediately embrace cloud-native network design and reap the benefits that come with it, those cases are few and far between. For most operators this is an incremental transition that gradually shifts from legacy systems to cloud-based systems. Through that transition, NFV provides some degree of automation and serves as a jumping off point for forward-looking investments.
Shekar Ayyar, VMware’s EVP and GM of Telco and Edge Cloud, described NFV as one “ingredient of a telco cloud. This is the idea that a carrier can take what were otherwise core network functions that were generally deployed on purpose-built hardware and you can now make them available as software workloads on top of infrastructure that’s virtualized by, let’s say, VMware.” The advantage here, just like in a typical data center, is decreased costs. But, for telcos, there’s the added benefit of service agility.
NFV gives way to a “broader thesis around how you can operate the entire telecom network as a cloud. You can provision resources from a telecom network, access them more on demand and deploy services much quicker.” He said VMware is a “strong believer” in the idea that, “You can’t do 5G…without a virtualized infrastructure.”
With that virtualized infrastructure in place, a telco cloud is now part of the broader pool of public, private and hybrid clouds. And that’s important, he said, lest telcos suffer the ill effects of “cloud isolation.”
“There’s this massive wave occurring in the telco community,” VMware CEO Pat Gelsinger told a packed house at the Moscone Center. “5G is ushering maybe the biggest capital investment build out for the rest of our careers. Telcos will play a bigger role in the cloud universe.”
Gelsinger’s comments were followed by a customer testimonial video featuring Verizon CEO Hans Vestberg, CTO Kyle Malady and Verizon Business CEO Tami Erwin. Malady, speaking of VMware, said, “As we move into the world of hybrid clouds…I think you guys are well-positioned.”
VMware COO Sanjay Poonen said the company’s expertise in data center, cloud and SD-WAN, coupled with the move to 5G, provides “a great opportunity for VMware to step in,” noting that customers engaged with traditional telco vendors like Nokia and Ericsson are “forcing them to certify our NFV stack. That’s what’s happening with Vodafone,” which began working with VMware in 2017 on an NFV roll out.
Ayyar continued to say that carriers have “largely missed the boat on the cloud revolution. With 5G…it actually gives them a new entry point to come in and reassert themselves in this architecture and play a role in the next generation cloud architecture.”
VMware’s Gabriele Di Piazza said,“We do think telcos will be part of the cloud fabric. You could say that 5G, it’s almost a moment in time that becomes a calling card for carriers to transform themselves.”
Pervasive cloud computing capable of handling a variety of distributed workloads are needed to move the industry from where it is today to the full vision of network slicing–bespoke data pipes created automatically on demand. This gives users the appropriate connectivity for the necessary application while allowing operators to optimize the allocation of network and spectral resources.
“If you read through the overall corporate pitch, the corporate strategy from VMware, is really any application, any device, any cloud with intrinsic security,” Di Piazza said. “This is our corporate, core strategy. If you look at investments we’re making…it’s interesting how this matches well to some of what our customers in telco are doing right now. It is a roadmap of cloudification. You need to be able to program your network.”
Cloud-native network design enables edge services
Early commercial 5G deployments in the U.S., South Korea, the Middle East and Europe are focused on delivering an enhanced mobile broadband experience to consumers. However, as 5G networks scale and mature, operators will look to create new service revenues based on massive support for the internet of things and ultra reliable low latency communications. The latter–applications like mobile virtual reality or precision industrial control, will require a distribution of the compute power generally associated with a centralized data center out to the network edge, closer to the user or device.
As it relates to virtualized edge cloud infrastructure, “I would say exploratory conversations were happening late last year, the beginning of this year and we’re working at architecting and trial implementations,” Di Piazza said. “5G architecture requires a different type of latency, which requires separation of the control and user plane. 5G, especially the 5G core, is what’s being planned for today. But there are many other cases. We have examples where MEC, or multi-access edge computing, requires distributed data centers or distributed compute nodes. I’d say it’s passed beyond exploratory.”
Discussing how he sees carriers deploying virtualized edge infrastructure, Di Piazza paraphrased: “Will they still be the dumb pipe allowing services to ride on top of the infrastructure? I do think there’s a genuine interest in [operators]owning the edge infrastructure and opening that up for a variety of business models. Maybe the business model is just connectivity. I think the reality is they are starting to see the opportunity to build from there.”
Intel’s Caroline Chan explained how the combination of 5G connectivity and edge computing will allow enterprises to tap into a wealth of unused data, gaining new levels of visibility into potential revenue-generating processes.
Chan, vice president and general manager of the 5G Infrastructure Division within Intel’s Network Platform Group, told a packed room in the Marina Bay Sands Convention Center that the vision is to create a “data-centric world,” but right now only 2% of that data, regardless of the source, is actually used.
5G is currently serving in a limited capacity to provide enhanced mobile broadband to consumers. As the standards and technology continue to evolve, the goal is to use 5G to support massive internet of things and ultra-reliable, low latency communications. That’s where edge computing comes into play.
First off, to take advantage of the latency benefits 5G opens up, compute resources have to be closer to the user–whether that’s you, me, an industrial robot or a security camera; that’s just physics. Second, in order to use data to initiate an action, you need the ability to conduct real-time data analysis. But with these edge investments, what’s the business case?
“Let’s look at how we monetize 5G and edge,” Chan said. She gave the example of a smart venue, which she characterized as a “fertile ground” for innovation that also brings a global reach as “sports transcend all cultures.”
For example, Intel worked with Arizona State University on an IoT project (LTE was the connectivity medium, but use your imagination). Sensors collected a variety of metrics for retailers in the venue. The learning: people in the cheap seats spend more money per person on not just tickets but things like food and merchandise than people who buy premium seats. “That kind of data is very valuable for people in that micro-ecosystem.”
From LTE to 5G, Chan laid out a proof of concept Intel worked on in conjunction with the National Hockey League. Following a U.S. Supreme Court decision that allowed for state-level sports gambling, the NHL wanted to explore ways to make gambling more engaging.
“The league decided to put sensors in the puck and in the athlete’s skates, collecting 4,000 data points per second. What can you do with it?”
By using 5G as a transport mechanism to feed that data into edge compute infrastructure, gambling can now be more dynamic and based on instant, real-time game stats. “It’s a scenario that really shows you what can happen if you take in real-time data, calculate with an algorithm and feed it back–all done with very, very minimum latency.’
She continued: “When you look at that whole scenario…it really shows you what can happen with a wealth of data…with real-time, that is monetization.”