LAS VEGAS -- The pending Dell acquisition cast a long shadow over EMC World 2016 this week, and Dell will have even more influence on EMC World 2017. That will be the first EMC World with EMC as part of Dell Technologies.
We spoke with Jeremy Burton about what the changes might mean to EMC technology. Burton, currently EMC's president of products and marketing, will serve as chief marketing officer for the combined companies when the $67 billion acquisition closes. In this interview, he discussed potential synergies between EMC technology and Dell, flash and the company's ongoing open source efforts.
Has the pending Dell acquisition influenced EMC technology decisions in any way?
Jeremy Burton: Such is the beauty of enterprise technology. Things that have been released this year, we've worked on probably for at least the last year and maybe the year before. So, the current product strategy is not influenced at all by Dell. But, obviously, going forward, you can see where the leverage points are. With something like a VxRack, it's really a hyper-converged. It's a server with storage in there. Now, Dell can bring to the table a great server platform, and so you think about a VxRack and those Neutrino nodes, they're really going to be Dell servers, somewhat performance-intensive, somewhat capacity-intensive, running EMC software-defined storage for one, but Neutrino for another. So, you can see over time the combination of server, storage and software is really what is going to make a new system.
EMC has been saying 2016 is the year of all-flash for primary storage. Do you think hybrid arrays go away and, eventually, we just go to all-flash?
Burton: I think, eventually, yes. For how long? That's going to be a few years. But we've been very clear, like this year, we think that flash is now economically cheaper than performance hard drives because of the benefits to be had. No mechanical parts, so you save on power and cooling and all these kind of things. I think it will be some time before flash is cheaper than SATA drives.
So, I think for the transactional prime of storage systems that that changeover is starting to happen this year, because the economics are better for flash. For the more capacity-optimized systems, I think it's going to take a bit longer.
Are there any other trends you've noted with respect to flash?
Burton: For the medium, obviously density. Look, there are questions around NAND as a technology. Where does that go next? And certain companies are starting to talk about new technologies that will ultimately replace NAND. And you get a little bit closer to nirvana, which is really people want DRAM performance at NAND cost.
Jeremy Burtonpresident of products and marketing at EMC
I think what's going to be interesting in the next couple, three years is technologies emerging that will supplant NAND flash. Now, it could be four or five years before we see them mainstream, but if you look at the research community, there is every sign that something will emerge that will give you an order of magnitude improvement in capacity, an order of magnitude improvement in performance at a lower cost. Then, it gets really exciting.
How has the EMC technology around open sourcing worked out during the last year? It was a big deal at EMC World in 2015.
Burton: It's going good. The ViPR controller was the first project that we open-sourced. The open source project was called CoprHD [pronounced copperhead]. We have many organizations contributing to that. I always said though that the ViPR controller was really the first step, in that when we open-sourced that, we made mistakes along the way. We've changed the license under which we open-sourced it. We knew that we were going to hit issues along the way. The good news is if you look back about six months ago, we open-sourced RackHD. RackHD is the technology that is in Neutrino. RackHD open source technology can elastically provision and reprovision resources to Hadoop and OpenStack.
You're going to see many, many more open source technologies from EMC, because I think part of the way these new applications are going to be built is on a predominantly open source-based infrastructure. And people are either going to go handcraft their own -- so they're going to get a ScaleIO or a RackHD, or get Docker and Mesos and this kind of thing and assemble it -- or they're going to go with something like Cloud Foundry, which is a much more structured, opinionated way of building applications, but at the same time is also an open source project.
Are you seeing anything surprising with respect to object storage?
Burton: It's early days. There are some very, very massive opportunities, but it tends to be what we would have called in the past ISVs. They've got an application that they're selling as a service. That application is collecting masses of data, and they want an economic way to store it. We've seen some massive deployments there. In fact, we have almost an exabyte at one single customer. But I think object stores will see their kind of heyday once the volume of new applications has picked up.
So, if we're right, if every business truly has to reinvent themselves and become digital, they're going to build a new set of digital apps to run their business on. At that point, things like object stores, I think, reach critical mass. Today, it's the folks on the bleeding edge. But typically, the folks on the bleeding edge are the technology ...The big Web-scale providers and things like that, they are the big opportunity. But over time, I think it becomes everyone.
EMC engineer discusses flash storage technology
Michael Dell unveils Dell Technologies at EMC World
Where DSSD high-performance flash fits in at EMC