Q&A: EMC's Jeremy Burton On Dell's EMC Buy, VCE, Cisco And NetApp
Ramping Up For The Future Dell Technologies
While a lot of mystery remains around Dell's planned acquisition of EMC, which could close any time in the next few months, one thing is certain: Jeremy Burton will be there to help the new company prepare for a software-defined, converged-infrastructure world.
Burton, who serves as president of products and marketing at EMC but will become the chief marketing officer at Dell Technologies once the acquisition is complete, met with CRN at the recently concluded EMC World 2016 conference to discuss what life is like managing product development and marketing for a company that will shortly be absorbed into a new larger entity.
Burton also shared details about the future of VCE in Dell Technologies, EMC's continuing relationship with Cisco, the end of EMC's relationship with Lenovo, and more.
For a look at the future of Dell and EMC, it's hard to beat a conversation with Burton. Turn the page, and see why.
What's it like to be part of EMC during the acquisition by Dell?
The Dell stuff has obviously been ramping up. I went through the Veritas-Symantec merger, and maybe it's a comment on the way Michael [Dell, Dell's chairman] is, but he tees up these issues and like, ding, ding, ding, ding, ding. He's a guy that likes to make a decision. When you're going through something like this, it's almost like whether the decision's right or wrong is almost secondary to how fast you made it. It sounds weird --you've got to get more right than wrong -- but honestly, almost like the wrong decision made quickly is better than no decision. Right now, we've got what I think is a lot of clarity about where the company is going to go and focus, and [Dell at EMC World already announced the name] for the new federation …
Then he's, "OK, this [the enterprise entity] is Dell EMC." He's definitely a guy who likes to make a decision.
EMC has remained busy since its acquisition by Dell was announced.
Actually, the best thing about the first four months, for me, has been a lot of the stuff that we've announced. When you think about it, we did Isilon, we did ScaleIO and ECS [Elastic Cloud Storage], and then we did DSSD and the VMAX all-flash. Then we did the VxRail and EMC Unity. … You don't build this stuff up overnight, right?
No.
It takes years. ... For me, seeing the products go out the door and having them go out in the competitive state they are is a huge point of pride. I've had the product organization for pretty much a couple of years now. … I get blamed for all the things that went wrong that I inherited. It's good now to be coming out on the other side of that, delivering the new products.
How much has the pending merger impacted new product development since October?
Actually very little, because if you look at the overlap in the product lines, it's probably in the midrange storage, where Dell has Compellent and EMC now has Unity. But even there, if I look at where the Dell guys play, they do a pretty good job on server-attach at the low end, and we don't really get down to that [$15,000 to $5,000] area. Really, what you end up looking at is, what did that product line aspire to do? What did Unity aspire to do? Are those aspirations still valid given that we're [going to become] one company? That's what we [have] to work through.
Honestly, I'll tell you, in terms of grand tasks when doing integration, product line rationalization is not one of them. It's actually not that hard, I don't think.
Yet with the new Unity, EMC is talking about sub-$8,000 for the hybrid version. That is in that low-cost area.
It's getting down there. Yeah. It's getting down there.
Going forward, you'll look at the product lines and say, "Just like we've done with VMAX and VNX." If you go back three years, four years, you basically had VMAX coming downmarket with the [model] 100K and VNX going upmarket with the 7000 and 8000 series. We made a deliberate call on product strategy to say, "Look, we're not going to have one product. We're going to focus the VMAX much more on the high-end and the VNX, the next generation, much more on the mid- to low end." Which is why you see Unity really in the sweet spot that it is today. To my mind, at least, that's the kind of exercise that we've got to go through with Dell.
When EMC and Dell look at their midrange storage lines, people wonder what cuts will be made. EqualLogic could be a potential candidate because it has been around for some time …
I don't know the Dell product line well. ... Compellent seems to have taken a lot more of the limelight in their portfolio probably in the last couple of years. They're probably better off speaking to that than we are. I think that's one of the areas that, we've got to look at it and make the right call and move on.
Well, why do you merge at all? Today's storage business is tough. Overall storage revenue overall is falling, although flash storage sales are rising. The storage industry faces a rocky future.
More than anything, folks are going to buy better systems. … I think Dell EMC, the enterprise business, if it's not looking like a converged-infrastructure business in five years' time, then something's up. That is the way people want to consume their infrastructure, either that or as a cloud service. For that, you need compute and storage. Dell really doesn't have a big storage portfolio, and we don't have a server portfolio at all. We know that in converged, it's not just things like Vblock, where we're aggregating storage, switch, server and network. It's [also] hyper-converged, where compute and storage are one. It opens up an avenue to go down probably at a much higher velocity than maybe we could've done as separate companies.
As those technologies converge, how much room is there for Cisco at Dell Technologies?
Cisco plays a huge role in Vblock. … When you go to hyper-converged, typically, there what we deploy is a software-defined stack on top of an industry-standard server. Cisco is starting to play in that area, but that has not really been its sweet spot. There's still going to be a networking play [between Dell EMC and Cisco] without a shadow of a doubt. They've got to have a game plan for getting into hyper-converged, as do we.
We don't have a VxRack based on Cisco UCS today. But it's not to say that we wouldn't have one at some point in the future. The thing is, the customer ultimately decides. If there's enough demand for a hyper-converged system that's going to be based on UCS, we'd be fools not to build one.
The VxRack today is based on what?
Actually it would be based on Quanta servers. Again, as a stand-alone company, we don't have a server supplier. We have things that we OEM for our storage appliances. Clearly going forward, we've got a pretty good server supplier coming into the fray.
Could VxRack someday be based on Dell?
I would almost categorically say that it will absolutely be based on Dell.
With few exceptions, the EMC product line is software-defined storage sitting on top of commodity hardware. Is Dell going to be the hardware base going forward for all the EMC products?
Yeah. It's not to say that we won't do systems through VCE that will be non-Dell. They're going to be based on customer demand, and obviously Cisco being a big one there. … We've got access to incredibly good economics in compute now. The size of Dell's supply chain is second to none. I think it's a pretty high priority that we replace all of our server infrastructure in the storage portfolio with Dell. By the way, I think five or six years ago, when we had the Dell partnership, it was Dell [servers]. That path, we've been down before.
And then you went to Lenovo ...
Yeah. [And now we] moved away from [Lenovo] for obvious reasons. [Dell's server portfolio is] going to be a pretty high priority. Because ultimately, the customer benefits, because they get better service and they're going to get a lower-cost server environment. At the end of the day, I think that's going to matter the most: servers and price.
Is EMC's partnership with Lenovo finished?
At this point. The difficult thing with the Lenovo partnership, we had absolutely the best of intentions, but the U.S. federal government made the determination that critical [infrastructures] -- that's not just government stuff, [but other areas like medical], that would swing like 60 percent of the domestic market -- that they won't let folks like Lenovo to play on grounds of national security.
When you say "like Lenovo," you mean China-based companies …
In that kind of environment, from our standpoint, that means we would have to dual-source. We would have to do everything twice. At some point, then it just doesn't make sense. That's more cost for us. ... I think we made the right call, and I think when we get this deal closed we'll move as much as we can over to Dell.
Is there a big cost benefit from moving EMC technology to the Dell hardware platform?
Without a shadow of a doubt there is. We're thinking compute here at the moment, so look at the volume of buying Intel processors. Can [Dell] get a better deal? Of course we can. Look at volume of SSDs that they buy. Can we get a better deal? Of course we can. I also look at it from a service standpoint as well: availability of parts. Can we get the parts the customer needs to fulfill a service request? We're going to have a fairly extensive set of depots to pick from, so we can provide a better service to the customer as well. It's certainly a cost issue. I think it's an availability of parts and service issue as well.
While EMC and Dell were discussing the acquisition, was software-defined storage a factor in the conversation?
It's one factor. You start to look at, we're going to bring these two companies together. Fine. What are the opportunities? I think one thing we both saw was converged infrastructure. [Converged infrastructures] today, which is this aggregated systems packaged like a Vblock with a sugar coating of service for the customer, that's state of the art today. Tomorrow, essentially server and storage come together in hyper-converged, and then you need software on the top to deliver the functionality. Even beyond that, once you get to rack scale, how can we deliver the kind of performance we think a lot of the next-generation applications are going to need? As much as Dell talked a lot about their evolving platform, they actually do a lot of custom work as well.
How will software-defined storage impact Dell Technologies?
You start to think, "Are there points of leverage between these two companies?" Certainly from a cost-based and supply chain standpoint. But right up to if we want to go build a custom compute platform to work with DSSD, for example, to deliver a real high-end system, that's a possibility too. Software-defined is only one factor of probably many.
Is business ready for software-defined storage?
Software-defined today is still in its infantry. The data center folks have got to make an architectural shift. It's like, I'm not going to go server, switch, storage, network. I'm going to have hyper-converged nodes, and I'm going to rack-and-stack these things to the ceiling. Then I'm going to be able to provision software onto them, depending on what the application is trying to do. That's a pretty big architectural shift, which is why I think very few folks have gone there today. It shows in our numbers. We do a massive volume of the Vblock. We do a very low volume of VxRack. Over the next three to five years, that'll flip. We'll end up doing more on VxRack than we do on Vblock over the next few years as that architectural shift kicks in. It's a compute-centric architecture.
Do you expect VCE to remain a separate entity going forward?
I think over time it becomes the heart and soul of Dell EMC. Chad [Sakac (pictured), VCE's president] is running it today. The important thing for us is we still need a focal point. If you say, "Now that's just part of EMC or Dell EMC," what you risk is, it's a part of everything, but then who's driving it? …
My product teams essentially build the building blocks for his [converged-infrastructure] systems. If you're not careful, if you don't do it the right way, it will be easy to get resentment in product teams, who might say, "Hey my route to market used to be, here's my shiny toy, can I buy it? Now I've got to package it in a converged infrastructure, so I'm a component." You're a part of the car. You're not the car.
How important will VCE be to Dell Technologies?
The teams also recognize, they see it today, that almost 20 percent of our storage revenue goes through VCE. Over the next few years that's going to go like 30, 40, 50 percent, without a shadow of a doubt. I still think we need the focal point. But let's be clear. I think all the time the brand for converged and hyper-converged needs to be associated with the mother ship, EMC and Dell EMC. We've got to be seen first and foremost as a systems, a solutions, or a [converged-infrastructure] company.
Given that, then, what do you think when you look at how poorly VCE has been treated in the press?
The thing is, someone has to create the market. So I still think the right call was with a separate entity, separate DNA. And look, it was a joint venture. Like any joint venture, everybody's always asking, "When's this going to end?" Every joint venture since the beginning of history has ended at some point.
We probably got the most grief when [EMC absorbed VCE] and did what everybody thought we would do. Immediately folks think, "It's just going to get absorbed and disappear." We don't do that at EMC. We like to have a line of sight to the business and put a pretty senior executive in charge, and then scale the thing over a period of time as it becomes bigger and bigger. Then it's almost safe to make it part of everything.
Going back to the building blocks idea, I notice that there's little talk about VSPEX reference architectures. Will VSPEX be disappearing?
It's still a program. … Reference architectures, I think, were almost the forerunner of converged infrastructure. Fewer people are doing reference architecture, and certainly when we talk with channel partners, they still do them, but they're less interested in doing reference architecture than in getting into the full-on [converged-infrastructure] business.
I think it's just the state of where the market is. Reference architectures were the big thing. Now they're the second biggest thing. And look, as hyper-converged ramps up, [reference architectures are] probably going to be the third biggest thing. It's not to say that they don't have a place. They do. It's just not at the level of interest now that maybe there was three or four years ago.
Are there any parallels between the Dell-EMC merger and NetApp's recent acquisition of SolidFire?
I think there's no parallels, honestly. I think if I was NetApp, I'd want to do half as good a job as EMC did with XtremIO. Theirs may be a bit easier because SolidFire has a product in market, whereas when we bought XtremIO it didn't. NetApp's history in acquisitions is terrible. This is one they can't afford to screw up. I hope they do screw it up. Their history would suggest that they will screw it up. But look, we're not underestimating them. We're going to have to compete in the marketplace with them. …
I think [SolidFire is] critical for NetApp. Maybe for once and for all it puts a bullet in their dream of creating one storage product to win them all because now they've got two.
Did you know that NetApp was holding its annual sales kickoff, also in Las Vegas, at the same time as EMC World?
No one cares about NetApp anymore. I don't think they do. We stopped caring about NetApp right around 2012. It's not that we don't care about competition. We do. But do I lay awake at night worrying about NetApp? Not really. I actually worry more about the startups that are out there. [And] some of the bigger players that are getting into [converged infrastructure] worry about those guys. I think anyone who's in the tech industry worries about [Amazon Web Services]. NetApp was a battle that we fought 2010, 11, 12. I think they lost. Can they recover? You never underestimate a competitor, but I think it's very hard. A lot of talent left the company.
EMC said all of its storage platforms will connect to the Virtustream Storage Cloud. How easy was that to implement, and what's its significance?
It wasn't easy, because you're trying to get every storage platform [that] wants to connect to the cloud to do it in a consistent way. That was hard to do. And look, until 10 months ago we didn't have Virtustream. It wasn't trivial. We made a couple acquisitions along the way that folks missed because we didn't really acquire them as businesses. We acquired them as technology. TwinStrata was one, and Maginatics was another. It really has been [those technologies] that have given us the on-ramps to the cloud. …
[We have to] treat the cloud as a storage feature of the backup platform. Yeah, it's object storage, and yeah, it's a hyper-scale object store, but the way we're looking at it and the way we can build a business is, it's tightly coupled with our storage and backup platforms. If we do a decent job there, I think we can build the business.
Does that mean Virtustream is not as cohesively coupled to non-EMC storage?
Actually, the Virtustream Storage Cloud has got an Amazon S3-compatible API. Anyone and everyone can use it. Again, the mindset of the Virtustream team is, we want to focus. We don't want to do everything. We can't just build it and they will come throw it out there like Amazon S3 and hope the developers come. We're really trying to build a business of something we know well. We know storage and backup, and if we can hook up to the Virtustream cloud, we should be able to build critical mass. Once we get there, then maybe we'll get fancy and do some additional things as well. But I think we've got to be focused. I've seen too many companies go out there, kind of shotgun trying to compete with Amazon. You can't do it. You've got to be very focused in what you do. That's the main goal.
Can other storage vendors connect to the Virtustream Storage Cloud?
If more storage vendors want to build it, great. We'll take it all. ... We support Amazon and we support Google and all that stuff as well. We think from a customer standpoint, if we can give them a package -- "You get Data Domain, it's got 500 TBs and 100 TBs of cloud" -- we can make it easy for them. You plug it in and off you go. Rather than buy a Data Domain, now buy a cloud gateway, now go deal with Amazon, and try and cobble the thing together. If it's too hard they're not going to bother. If we can tightly couple that then we can probably do all right.