The industry is seeing a huge upswing in the amount of data introduced by Internet of Things (IoT), edge computing, artificial intelligence (AI), virtual reality (VR)/augmented reality (AR), and other applications—on top of all the traditional workloads associated with a service provider network—and 5G will only exacerbate the data explosion. What will the network infrastructure of the future look like, and how will it handle these new demands? What are the benefits and risks associated with cloud-centric environments, and what role will the cloud play in network transformation?
In this interview, Kevin Shatzkamer—vice president of enterprise and service provider strategy and solutions at Dell EMC—discusses the role of cloud computing in 5G, the concept of cloud-generation mobility, and how cloud generation mobility tackles the challenge of ever-exploding data traffic.
SDxCentral: From your perspective, what role does cloud computing play in 5G?
Shatzkamer: Even today, cloud computing foundationally powers the vast majority of experiences that consumers and enterprises have on the Internet, whether we’re talking about gaming, social media, video, and so on. Going forward in a 5G world, as services and applications become driven by more real-time and interactive requirements— things like Internet of Things (IoT), augmented reality (AR)/virtual reality (VR), connected vehicles, and smart cities—the role of cloud computing to deliver these differentiated experiences will be even more critical. These new services will not be driven by massively scaled, centralized data centers but intelligently distributed systems built at the network edge while still leveraging the operating principles of cloud computing.
We’re looking at a world where 5G and cloud computing are actually synonymous with each other. To meet the needs, expectations, and growth of the Internet from a mobile perspective— whether in terms of number of devices, the amount of traffic generated by those devices, the amount those devices will be moving, and the rate at which they move— cloud computing is essential. The status quo, the way we built networks in the past, just doesn’t work anymore. We have to get to the new operating paradigm, the new service-delivery paradigm.
SDxCentral: You’ve used the term “Cloud Generation Mobility”— can you expand on how you define this and how it differs from 5G mobile?
Shatzkamer: I would define “Cloud Generation Mobility” as the ability to deliver new mobile experiences through a cloud operating model. I would compare the phrase with the “5G” term that we hear in the industry quite a bit—I think they’re one and the same. 5G represents the transformation of the access network to meet the needs, services and experiences that I talked about earlier. The 5G architecture is built foundationally on the cloud— cloud radio access network (RAN), edge cloud, NFV, service function chaining— these are all cloud technologies intersecting with a telco operating model. It’s really the intersection of IT and cloud-like technology infusion into the network itself.
SDxCentral: The amount of traffic traveling over a given network has exploded and is unlikely to stop growing anytime soon. How does cloud-gen mobility address this issue?
Shatzkamer: When I talked about cloud generation mobility earlier, I talked about edge cloud specifically. If we look at the law of large numbers in terms of how we’ve built the Internet, when the amount of anything has exploded, we’ve reacted the same way. That reaction was to decentralize and distribute resources.
When we saw too much content and requests for video on the Internet, we built the concept of the content-delivery network. When we saw DNS requests rising to levels that were unsustainable from standard DNS servers, we started to geographically put DNS servers in place. Now, with the rise of services like IoT, AR/VR, and connected vehicles, we’re reacting with this cloud-generation mobility model of building clouds at the edge. We’re taking the operating model of massively scaled, centralized facilities, and applying them to the edge.
SDxCentral: As networks become more distributed between the edge, core, and cloud, what new challenges arise?
Shatzkamer: That’s a great question. I think about the conversations we have with our customers from a telecommunications perspective. When we talk about the core, when we talk about the transformation of their existing facilities, the conversation is very technology-driven: When do I leverage virtual machines? How do I think about the networking stack? How do I think about SDN? When do containers come into play? How do these applications become more cloud-native?
As we move that same conversation to the edge, it shifts from technology-driven to an operations conversation: How do I start to think about an operating model in which I centralize my control plane and distribute my user plane, or also centralize my management plane? It’s a recognition that as we move more toward the edge, facilities are more power-constrained and space-constrained. They’re often lights-out facilities: unmanned and unmanaged—there are no people to handle operations. So how do I think about remote operations and the capabilities that are required on the infrastructure to be able to manage these facilities remotely? Because every time I need to do a truck roll, every time I need to send a person to a facility, the costs are dramatically higher than at my centralized core facilities where I have people working every day. Moreover, the skillsets and technical know-how between the people at my centralized core facilities and those needed for maintenance at the edge may vary greatly.
And then as we move to the next layer of the edge—what we call the far edge—the conversation can shift even further from operations to logistics. So the questions become: How do I start to handle multi-billion endpoints that are connected to my network? When I look at the gateways and nodes that I need to set in my own managed domain, how do I get them to the places they need to be when they need to be there so that I can offer the services my customers need?
SDxCentral: What is Dell EMC’s approach to solving those challenges?
Shatzkamer: When we look across the landscape, it’s easy to see why Dell EMC and Dell Technologies are so well positioned to play a core role in this cloud-generation mobility and 5G journey.
First, we bring the leading technology company to the table, with massive experience and expertise building out these clouds from an IT perspective. Second, we’re leading in terms of thinking about the operations, management, automation, and orchestration of this infrastructure and helping drive that transformation. And, third, Michael Dell has built a company that’s second to none from a supply chain and support perspective. So, as we start to see 5G, NFV and SDN scale, we have the ability to bring the logistics capabilities to the table to ensure that the infrastructure is ready and available for that scale.
It’s a complete mindset approach for Dell EMC to play a key role in this 5G transformation, and that includes working early on with service providers to define the capabilities of the infrastructure going forward, and how that infrastructure will be operated. It means not just being there at a time to deliver infrastructure but being a strategic partner with the telecommunications industry and our vendors to define what the next-gen infrastructure looks like. It means playing a key role in proofs of concept and what we call first-off applications or first-off architectures that allow us to demonstrate technology readiness for the industry. It’s about playing a role in the design, architecture, and deployment of these end-to-end infrastructures. And finally, it means playing a role in helping out with the logistics and scale that’s required for this business to be successful.
SDxCentral: What cost advantages or risks are associated with cloud-centric environments?
Shatzkamer: When we’ve traditionally looked at the way we’ve built infrastructure in the telecommunications environment, we’ve focused on the concept of five-plus-nines of availability—SLAs with very limited downtime. I think there’s a view in the industry that cloud-centric environments and cloud-based architectures are inherently unable to offer five-nines of availability. So, from an industry perspective we need to recognize that the statement is only true if you don’t change the way you operate your infrastructure. In that regard, as we start defining which SLAs are really required to offer those five-nines and think about it less from an individual node perspective and more from a service-availability perspective, I think we’ll start to realize the true cost advantages.
As an industry today, we still build very siloed environments where we’re leveraging cloud-like technologies such as NFV. I think we’re often seen as being in this trough of disillusionment because we haven’t realized the cost advantages of cloud computing yet. Those advantages come when we start to put multiple applications, multiple services, and multiple environments on top of shared infrastructure—creating a true multi-tenant edge-core-cloud architecture. To do that, there’s a massive shift in terms of the operating structure of telecommunications companies and the skill sets within the across various departments of these service providers.
The risk in the industry right now is that a lot of the available skill is being consumed by well-known large Web-tech companies at a time when the broader industry needs to see that skill set be deployed, adopted, and readily available across their own operational environments as well.
SDxCentral: Any final thoughts?
Shatzkamer: I’m excited! The industry is facing a new set of challenges that like minds have the ability to explore and innovate around. I’m also excited from a consumer perspective that the new services and applications will become available to me—and my children! —as we start to see these networks get deployed and operated in the right way and more and more services get deployed.