Q&A: EMC's Mark Lewis, EVP of Open Software
To say mergers and acquisitions (M&A) and strategic activity in the storage space are heating up is a gross understatement, especially when the conversation involves EMC
The Hopkinton, Mass.-based company has unleashed a flurry of news recently regarding its new push for information lifecycle management (ILM), which is perhaps the hottest phrase uttered since Web services lit up the high-tech skyline a few years ago.
Once known as one of the largest proprietors of storage hardware for Fortune 2000 companies, the vendor has drilled down into a different software layer
by first making a serious acquisition in ILM specialist Legato Systems for $1.3 billion and next making a $1.7 billion run at one of the largest enterprise content management software players in Documentum.
Throughout these M&A maneuvers, which also includes the purchase of BMC Software's storage management assets, EMC has been pounding the table for software standards, casting aside its company-centric WideSky initiative to ingratiate itself with the Storage Management Initiative specification (SMI-S), an in-progress schema geared to make storage products from multiple vendors interoperable.
The moves have prompted several analysts to quip that EMC is a proprietary hardware company, but an open software company.
At the center of all the recent action is EMC Executive Vice President of Open Software Operations Mark Lewis. Lewis, who reports directly to President and CEO Joe Tucci as one of a cadre of executive VPs, is responsible for making sure the integration of the large software purchases runs on well-oiled wheels.
Lewis defected from HP last year to become the EMC's CTO. He was elevated to his current position in May.
To say that Lewis is a busy man is also an understatement: on the day EMC closed the Legato deal, he presented an integration path that shows how the company's many layers will work to form a complete information management solution. Less than a week later, he presented a more complete integration strategy at the Fall Storage Network Show in Orlando that factored in the assets of Documentum, assuming its bid will be successful.
Internetnews.com recently caught up with Lewis to chat about adding software to make the ILM formula work, how EMC measures the competition, and the direction of the company.
Q: On May 30, EMC announced you'd be moving into a different position as executive vice president of Open Software Operations, working under Joe
Tucci. How did your previous employment at HP help prepare you for this role?
Moving in as CTO was one of the first strategic and staff roles I've ever had. So, interestingly enough, it was a great way to come into EMC and do something that was actually different for me and fun to do. But coming back into running the software business is really more to my roots and my training, and what I've done with the rest of my career, which is running businesses. I started the software business for Compaq and then ran storage for Compaq. I'd say this is closer to what I've done in the past.
Page 2: Interview with EMC's Mark Lewis (Continued)
Q: EMC has made an awful lot of references in the last couple months to information lifecycle management. Describe what ILM is and why the company is banking so much on the ability to deliver it to customers. Then, describe how the increased focus on government regulations such as HIPAA and Sarbanes-Oxley impacts or influences EMC's innovation in ILM.
ILM for us becomes the next extension to what we believe we successfully delivered with automated network storage itself. And the neat thing about ILM is obviously not only that it's a combination of product, but it really describes a point of solutions and a value proposition to customers. So, more specifically for us, ILM is our ability to help customers dynamically move information to the right platform and the right storage media at every point in a data's lifecycle such that we reduce the customer's TCO of managing that data.
The foundation for that comes in differentiated platforms and saying there is even more expensive storage and less expensive storage. We built up to that with storage networking. We built up to that with automation that we put in. All of those pieces and building blocks are needed for ILM — and the new thing becomes this dynamic data movement: the idea that you move from static backup and even static replication to the idea that data movement becomes very dynamic — something that can be done while the application is running.
Now, the regulatory implications and the regulatory happenings of the world today have accelerated the need for ILM because, up until now, customers really weren't addressing their electronic information the same ways they were their paper information. They really weren't dealing with compliance across their business through all of their records, across all applications. So, it was something that just wasn't dealt with.
So now with HIPAA and the SEC regulations and the DOD regs and all of the different ones, the common thread is that customers have to come in across their applications and ensure that their data is protected and archived and maintained. So last year we introduced Centera — the fixed content storage to help be the record retention placeholder for the information. ILM becomes more about putting the software in place to help us move the information and records onto products like Centera while not bringing down the applications.
Q: In the past year or so you've made a number of software-related acquisitions, including Astrum Software, Prisa Networks, BMC's storage management software, and Legato Systems. And you've recently moved to acquire Documentum. Analysts in the past year have come to refer to EMC as a proprietary hardware company, but an open software company. Is that fair description? Is that too simplistic? How would you characterize EMC's direction?
Well, every product you build is in and of itself somewhat proprietary in the sense that, you know within Windows, Microsoft only allows limited access to what people can put around it. So I believe that with both software and hardware, in the sense of how the world would categorize it, [we] are attempting to be very open. You saw [EMC's executive VP, Storage Platforms Operations ] David Donatelli talk today about SMI-S, and about how we'll have all of our products SMI-S supported. We support heterogeneous environments with disks and tape and other suppliers' software and other suppliers' hardware arrays.
The funny thing I've always found is that I think EMC is the most open hardware player in the market today. If you take an EMC Symmetrix, it will run on an IBM machine and a Sun machine and an HP machine. Go to an IBM Shark and say: "Where will it run outside of IBM?" Go to an HP-XP and say: "Does it run outside of its environment?" Go to Sun. Each of those vendors that says EMC is proprietary is actually proprietary on their own platforms. So it just doesn't fly with me.
Page 3: Interview with EMC's Mark Lewis (Continued)
Q: Further along the rivalry stack, on July 30 you announced significant advances to the DMX hardware line and in addition a couple of interesting software announcements, including snapshot software, SnapCopy, and the new version of the company's replication software that spans longer distances. VERITAS came at me with some interesting things to discuss regarding this. "With these software add-ons" — referring to the snapshot software and replication software — "EMC continues to chase a rapidly evolving storage market and remains well behind established industry leaders. VERITAS Volume Replicator software is the fastest growing replication software among the top 10 storage vendors, according to a recent Gartner report."
They went on to describe how EMC's new software has no file system capabilities and only works on its own high-end arrays. Then there is a quote from storage analyst Arun Taneja: "If EMC does not start to hustle in delivering inexpensive replication solutions, everyone is going to move away and do applications either from the network or some gadget in the middle of the fabric. This is a survival move for EMC. If they don't do it they are going to hurt." Are these fair assessments?
Well, it's funny. Things are both in fair context and not. You know it's usually fair to say that those statements are 100 percent accurate and 100 percent misconstrued in terms of what the true meaning is. My favorite is the stats from Giant Loop, because Giant Loop went out of business within I think a week of publishing that data. So it was pretty interesting that the customer VERITAS chose got no traction. The market will speak on these things.
We have focused our replication tools and what we've done with replication on our enterprise accounts — the global 2000. These companies are looking for deep capabilities in replication to manage business-critical applications, and none of them are using VERITAS. It's not that market. VERITAS has been attempting to build a market in the low end for replication, to which we do agree there is an entry-level market where we've had SnapView and MirrorView in the CLARiiON space and now additional capabilities in Symmetrix, but it's really not about a low-end versus high-end product in Symmetrix. Customers need full copies and replication for business continuance, and the snap copies make sense for efficiency.
So again, it's a high degree of rhetoric. It's "fastest growing" — when you don't have any market share, it's easy to be fastest growing — but that's not where we want to be. We will develop a very complete line of products, and with the acquisition of Legato we acquired a product called RepliStor, which is a host-based replication tool. Now, having said that, post-base replication is of limited use because it does require interaction with the file system, it does require that the host be running and so it cannot tolerate host failures, and it does require tie-in to the file systems, which in some applications is an advantage, but in most apps — again, we want to be able to do replications in a heterogeneous way, and it doesn't work in a heterogeneous way there.
So again, we talked about the right tool for the right job. We're working on building all the right tools, but we don't feel the tools that we have are bad, but in some cases we do need to extend the market. That's what we're doing. But there's no catch-up. Anyone using EMC BCVs would not even be contemplating — that's like someone using software RAID contemplating buying a Symmetrix. I don't know of a single player that is contemplating using the Microsoft RAID V they embed with Windows or buying a Sym. The difficulty is going to be for IBM and HDS, because now Shark and the HDS product line are at a significant competitive disadvantage, and that's where the competition will be.
Q:You've already mentioned ILM as the here and the now, but what other trends do you see unfolding in the storage space over the next five years? What will EMC do to accommodate these demands?
Storage will really separate itself as an infrastructure component, just as PCs did for access and just as networking did, so we believe that over time, that will become fairly complete. We don't believe that anything is going to stop that. So that's kind of the overarching mega-trend, and below that it says — again that we build in ILM as an overarching capability, and below that things like moving intelligence into the network from the server, and those tenets are really what we see, and for customers what we envision is what I really call making storage invisible.
We really see it as — if you look at a network today, while it's highly valuable and highly complex, and it's a great business, Cisco shows that it's invisible to you and me — it's not something you have to interact with. It's very automated and very well run, so there's lots of value, but it's invisible, and I think that's where we want to take storage. Ultimately while we still see value in the products, to the end-user customer that's consuming it, we really want to make it invisible.
Interview courtesy of internetnews.com.
Back to Enterprise Storage Forum