The growing adoption of virtualization is having a profound impact in the satellite industry. Virtualization holds the promise of enabling significant benefits for the teleport, including scalability, cost savings, and better integration with customers, vendors, and partners. Robert Bell, Executive Director of the World Teleport Association (WTA) shares the findings from the recent report, “The Virtual Teleport,” and how teleport operators are embracing virtualization.
He shares insights as to virtualization’s benefits, the barriers as well as what the future holds. Constellation Podcast Host, John Gilroy: Robert, we’re hearing this term virtualization more and more, but it seems to mean different things to different people.
How do you define it as it relates to the teleport?
It actually only means one thing. It just has different flavors. Virtualization is all about abstracting the software level of what goes on in your computer from the hardware level. The physical infrastructure of the server is almost completely independent from what runs on it. And you can run a single program across multiple servers, and you can divide up that server into multiple bits and use each one independently.
So it’s really incredibly powerful technology that runs everything from email and internet to the very sophisticated systems running inside teleports. And it is of course the foundation of cloud services, so it’s in pretty much everything we do today.
Why is virtualization now becoming more prevalent in the satellite industry? Where was it before?
The satellite industry has always lagged behind IT and telecom. We’ve been in our own little niche, serving very narrow markets with proprietary solutions that nothing else really needed to work with. So, we’ve got sunk investments and the industry running these standalone, closed-loop systems using proprietary standards. But there’s been a sea-change in our industry to realize that rather than being 1 percent of global telecom, which what satellite is in terms of its revenue and services, that if you could plug into the rest of telecom, you could actually get to 2 or 3 percent or even more, and that’s a lot of money.
Let’s talk about the teleport operators themselves. How are they applying the power of virtualizations to their businesses?
It ranges from the cost side to revenue. On the cost side, automating network operations is first and foremost, which is something terrestrial telco has been doing for decades at this point. With all the things being done manually in the NOC for a long time, virtualization gives you the ability to monitor performance across a whole enormously dispersed set of assets, whether they’re in your local facility or thousands of miles away. It lets you go from being reactive in your network monitoring and maintenance to proactive because you can see trends evolving because of the level of information you have in everything.
A level above that is service orchestration, which is basically the manager of managers. So we have all these network operating systems, and service orchestration aligns those applications, the data, and the infrastructure with your business requirements. This gets to business rules and policies, and automated monitoring of service levels, and the ability to dynamically reconfigure systems automatically based on rules that take into account changes in what the customer might be doing.
But it goes on from there to one of the hot new areas of RF or radio frequency. That’s a very physical thing. We bring down RF from the sky and we run it through wave guides, get it into a receiver and then convert it with a modem.
The trend now is to push the conversion process as close to the antenna as possible so instead you’re just dealing with a stream of bits. The flexibility that gives you in your operation by virtualizing that physical signal is enormous. It gives you the ability to run multiple facilities from a single control center, to trading traffic with fiber in new and exciting ways. You go from there to the cloud.
Once we’re virtual, it becomes a lot easier for a teleport operator to integrate the cloud into their services. And now we have this new revenue model called ‘ground segment as-a-service’, which is going to transform operations over time.
From a practical standpoint of ROI, where are the opportunities for return on investment for virtualizing the teleport?
In our study, one teleport operator that runs hundreds of video channel monitors is monitoring 5,000 devices at local and remote sites around the world through this sophisticated network management system. Just imagine the army of people and equipment you’d need if you weren’t operating in this virtualized world. They’re able to serve their customer much faster, at a much lower cost and they’re able to scale. Which means you can make more money in terms of margin.
A simpler case is where an operator runs two teleports, one staffed and the other completely unstaffed. They’ve proven it works, so their plan for all future expansion is unstaffed facilities because they can run all of them from one central place.
This also helps build revenues because you can expand geographically with a very light footprint, accessing business you’d never have been able to access otherwise. And then there’s the customers of these teleports, particularly those serving telco, data, or enterprise customers, that increasingly demand that you not run a manual operation, that instead you give them a website so they can see what’s going on.
They want the teleport operator, the satellite operator, to integrate right into their operational support system, their OSS, or their business support system, their BSS, so that it’s all automated, so that it all looks the same as what they’re already running. And that’s the power that real success with the virtualization provides to a service provider in this industry.
That’s a lot of ROI. Let’s talk about the related topic of software-defined networking, which is included in your report. How does this concept of a software-defined network relate to virtualization?
It’s virtualization applied to the network. Software-defined networks are really, again, kind of magical. If you step back, it really reminds you of that Arthur C. Clarke law that sufficiently advanced technology is indistinguishable from magic. The idea is that a layer with SDN is monitoring multiple channels.
So let’s say you have a complex network with fiber, with a bunch of satellite hops, that’s maybe serving cruise lines, where we also have cellular operations in some of the ports. On a software- defined network, all these different channels can be monitored automatically, every few milliseconds, to find the ones that have the best quality service, lowest congestion, highest throughput, lowest error rate. Traffic, whatever it might be, is automatically routed to that.
And so instead of just sitting there with one circuit that’s doing well, and another that’s doing badly and trying to figure out what’s going on with it, the system has taken care of that for you. And the net of that, in addition to quality of service, is that you have optimized your usage. You actually need less overall bandwidth because the system is looking at optimization all the time.
It seems like it really increases flexibility for the operator to control their quality of service.
Absolutely. It’s very much like what you’ll hear from a cloud service company selling you on the concept of flexibility. They can bring it up when you need it and take it down when you don’t need it. It’s about optimizing the capacity, which is of course the most expensive part of any satellite service. Traditionally, the capacity you pay for to deliver the greatest value to your end customer, you don’t have fine grain control to know what you can get out of it. You sort of have to buy a whole lot to make sure you’ve got enough margin. And SDN really does a lot to strip that down, which makes satellite ultimately a more cost- effective solution.
We’re seeing a new generation of satellites going up now that are much more dynamic than previous ones. How does virtualization and software-defined networking enable the teleport to better support these operations?
It’s going to be crucial when we talk about more dynamic satellites, which is a pretty big bucket because we’re talking about LEO configurations, where you have high-capacity spacecraft moving across the sky, but also electronically configurable satellites that are being developed, which create complexity for service. You’ll need a highly virtualized network management system overlooking everything just to be able to deliver the quality of service that’s needed, and to get the advantage of that massive influx of bandwidth that’s being provided.
Earlier, you mentioned this concept of ‘ground station as a service’ and the early adopters of virtualization in the cloud. What lessons can be learned from these pioneers that can be applied to the wider satellite industry?
Quite a few. Instead of all these little fiefdoms sitting independent, the idea of the virtual teleport is that you link these all together so they actually can exchange traffic, they can serve as each other’s backup facility and so forth. When you talk about this ground segment as a service, for most of these companies, their secret sauce is the operating system they’ve developed. And it’s really kind of an amazing idea. It’s based on the idea that you’ve got a lot of facilities sitting around the world, antennas, power, and connectivity, that’s underutilized.
There’s a lot of spare capacity out there and wouldn’t it be great if it could be used. And so what these operators have are systems, basically partnerships, with a lot of owners and operators of physical infrastructure to use that capacity on their systems when it’s available and appropriate. From a business standpoint you sort of think of them as competitors, but they’re actually teleport customers because they’re paying to use the facilities when they need them, as well as being business generators, bringing new kinds of business to teleport operators. So, it’s a really interesting business model.
One hidden problem of virtualization is that when you put everything into data and you’re connecting all these disparate facilities, you are now in the land of gigantic data pipes, really enormous amounts of data connectivity, which costs something. So they’re constantly trying to manage that balance between what they can do with the one or two facilities they operate versus remote facilities.
The cost factor is fluid and can be difficult to manage. And the same thing applies to when they’re using cloud compute capacity, as part of their solution. The cloud companies are growing fast and are profitable because they charge good fees for ingress and egress, getting onto and off their network. Those costs can run up pretty fast and produce an uneconomic solution if you don’t really design it carefully. So it’s a breakthrough idea, but you have to get the new fundamentals of the business.
You listed a lot of the benefits to virtualization, and some of the challenges with the data pipes. What are the barriers to adopting this new technology?
There’s always barriers to something new, and they’re the things that you would expect. It is expensive to get started. I mentioned big data pipes. Even though most teleports already have some form of data center or other, you’re probably going to need a bigger one because again, there’s always going to be a need for local compute capacity and storage capacity and other things to make the whole thing work. You’ll need a new set of skills from the ones you have.
One CEO said he doesn’t actually have any RF engineers. And that’s because they’re at his partner facilities. “They’re the experts about that…that’s not what we do,” he said. So there’s new skill sets, new equipment, the fact you’re going to need big capacity pipes and you’re going to have to think through your business models to make all of that work.
It’s not a slam dunk, but ultimately it opens a doorway to optimize teleport operations, wherever they might be. And no matter how traditional the facility is, there’s a lot of play here for business improvement.
The traditional way of managing satellites and communications really changes when applying virtualization and cloud technology, doesn’t it?
When we go through this process of virtualizing and more importantly, automating any set of processes, you get into serious change management. One of the more advanced operators that I know has just put a tremendous amount of time, energy, and effort into building out their change management process so that it’s very rigorous, because you don’t want to get it wrong. All teleports have tried to be good at this. I think it’s a step change in operations.
Across the federal government and the DoD, you hear more and more about interoperability. Seems that virtualization can play a big role in this idea of integrating satellite and terrestrial networks.
Yes, and it all comes down to those things I mentioned before, the BSS and the OSS. If you can get what looks like a seamless connection between your satellite, your satellite systems and your terrestrial systems, you’ve accomplished something that has been standing in the way of the industry for its entire history.
The idea that we’re going to remain in our little silo, with our special processes, with terrestrial thinking satellite is expensive, unreliable, or they don’t understand it. But if it’s just a port on their router that follows all the standards that they’re used to, then those concerns go away. And standards are a really big part of this.
Again, our industry has always been resistant to open standards. Generally speaking, we’ve always tried to play the proprietary standards game, and the IT and telco industry left that behind a long time ago. This is an opportunity for our industry to catch up and plug into the standards, not only to help us get business, but quite frankly, to make the satellite and teleport business better as well.
With all the benefits you’ve listed with the adoption of virtualization, what’s a teleport of the future going to look like in four or five years from now?
Like I say, teleports are already data centers with dishes because they have to be. Even if you’re carrying broadcast quality video, ultimately they’re carrying bits. So I think you’re going to see that optimal kind of connection between terrestrial telco, terrestrial IT and satellite.
One executive said he wants his satellite service to be sold by a telecom partner, so that it’s an invisible part of their solution. Making the teleport more invisible is kind of an odd thing to say. But once it becomes invisible, which has been kind of a problem for 50 years, it will turn out to be a huge advantage because it will become part of accepted global infrastructure and everybody will make a lot more money.
Listen to this and more than 100 other podcast interviews on Constellations — www.kratoscomms.com/constellations-podcast
Review the full list of interviews and subscribe at www.kratosdefense.com/constellations-podcast
Interested in a particular topic for discussion? Email Constellations at Podcast@KratosComms.com