At the OpenStack Summit last month, the Cloud Foundry Foundation's Devin Davis spoke to Telco Transformation about the importance -- and applicability -- of multi-cloud solutions and open source for carriers and content-delivery networks.
Earlier, in Part I, Davis, head of marketing at Cloud Foundry, delved into both industrial IoT and enterprise cloud agility, along with network infrastructure and DevOps complexity issues. (See Davis: How Cloud Foundry Helps Create Digital Businesses.) Now, in Part II, Davis discusses bandwidth and latency issues in video content delivery as well as how both colocation and the fear of vendor lock-in play into those issues.
Telco Transformation: Previously, in Part I, you mentioned video providers Comcast and
AT&T U-verse. What have you seen recently in terms of cloud impact on video-content delivery?
Devin Davis: There are definitely gains to be made from various little things that happen within different layers of the cloud stack, but that's about bandwidth largely in terms of quality and, to a large degree, location. This is why you'll hear colocation centers talk about the importance of having local data. And, of course, Amazon and all the other very large cloud providers make sure that you're pulling from someplace local. It has more to do with distance and bandwidth, I think, than it does with actual cloud solutions.
TT: So what I seem to be hearing is that, when it comes to picking a cloud solution, go with the local option primarily because of speed and velocity. Would you agree with that? Or what's wrong with what I've said?
DD: We're getting into a multi-cloud discussion. When it comes to multi-cloud, it's really a question of what's the problem that you're trying to solve first and foremost because there are different clouds that have different strengths at the infrastructure layer. And, frankly, a lot of companies -- particularly big companies -- are always going to have a fair amount of things that live onsite, and that's just not going to go away because they can't allow it to for regulatory reasons. When you look at financial services, when you look at healthcare, when you look at those types of verticals, some of those things will never be able to be in a public cloud. Many of them will with time, but certainly not in the near term.
And then when things do move to a public cloud, it's going to become a question of "What is it that you are trying to do?" Is it going to be data intensive? If so, maybe Google Cloud platform is your best bet. It could be that Azure is a good fit for another reason. It could be that Amazon Web Services Inc.
is a good fit for another reason. Of course, OpenStack, for one reason or another, could be the right fit. Or it could be that you just donít want to deal with it, and you're just going to send it all out to a co-lo (colocation center) and they're going to have to deal with the whole thing.
TT: In terms of what's "right," what are people really concerned about with vendor lock-in? And how do you address those needs?
DD: I think that, ultimately, the big concern when it comes to vendor lock-in is losing your work. It may be data. It may be applications. It may be some combination thereof. It may be custom services. It could be a million different things, but what it really comes down to is: "I have this thing that my business depends on, and I am stuck with this company because of it." This is the old Oracle Communications move from the 90s, and they still are making billions of dollars because companies can't leave and they've signed 30-year contracts with them. The solution is open source. OpenStack has done a wonderful job with this. The Linux Foundation and the various projects under the Linux Foundation are helping to solve this problem.
So the solution for us, when it comes to the platform layer with cloud, is that we have a certification program. If you are using one of the five certified clouds we have, and there'll be more by the end of the year, you know that your applications, and your services, and everything that you need to work will work no matter which one you choose. So if you choose IBM BlueMix and you decide two years down the road that that's not the right solution for you, you can move to Pivotal Cloud Foundry, and everything will just port right over; or Fujitsu [Cloud Service] K5, and so on and so forth. That's the solution for us at our layer. But it's a complex problem and part of it too is that, at a CIO level, you may be very concerned about vendor lock-in, but if the people who are actually doing the work are like, "Look, this is way easier," you're going to have an uphill battle with your personnel, which is a totally different issue.
TT: Sometimes you see the opposite, where the C-suiters want to consolidate vendor relationships simply to reduce complexity from the business perspective. And then it's just a matter of, "What's going to work for us?"
DD: Right. Right. That's very true. The thing is that when you're talking about technologies of this size and breadth, you're largely talking about large corporations. When you're talking about large corporations, there's so much that goes into making the simplest decision that it's just a hell of a process.
I think your point about vendor lock-in is really interesting. It's a funny world that we're living in and we've gotten to this point where open source is becoming the standard for large companies, and it's largely a backlash from all these long-term contracts we saw many, many years ago from a lot of the biggest technology companies in the world. And they're all learning to take a different approach to business, and that's really the only way you can be successful in this market, I think, which is a fascinating turn of events. It's just a complete flip from 20 years ago. Things change, and it's a good time to be in technology and watch that happen.
— Joe Stanganelli, Contributing Writer, Telco Transformation