Mark Lewis, CDO, EMC

"We're not into buying competitors. We're buying products that extend our capabilities into new and addressable markets."

February 13, 2007

12 Min Read
Network Computing logo

While wearing a variety of hats since jumping from Hewlett-Packard to EMC in 2002, Mark Lewis has been the unofficial point man for EMC's flurry of acquisitions. (See Lewis Quits HP for EMC and EMC Fortifies the Bench.)

Now, as chief development officer, Lewis is helping EMC put all the pieces together as part of its "One EMC" campaign to integrate more than 20 acquisitions in three years. (See Tucci Aims for 'One EMC'.)

We recently talked to Lewis about how he sees the pieces fitting together, and what EMC's product landscape will look like as a result. While the integration is well under way in many areas, Lewis cautions not to expect everything to be interlocked so soon.

"There's kind of a misperception about rapid integration that speed is always going to be the better way to go in acquisitions," Lewis says. "I don't believe that. When you buy companies, those companies have customers, and they bought into the company because of its values as a separate entity. We've done a lot of acquisitions. They led us to hold separate pieces of functionalities and sales forces. In the case of VMware, we are keeping it very separate." (See VMware to Spin Out.)

Following 22 major acquisitions, EMC execs have made it clear 2007 will be a year of integration. That is something investors and customers have been calling for. But Lewis says the integration work began long ago."We do hear loud and clear from customers that we want to drive better integration, more simplicity in products and more true integration across what we do," Lewis says, echoing themes that other EMC executives made at this week's RSA conference in San Francisco. (See So Long, Security Silos .) "That's something that's been happening for years behind the scenes, and some you see in front of the scenes now but a lot of it is in the core underpinnings behind EMC."

Lewis also says not to expect a vastly different EMC after integration is complete. "The product lineup surprisingly will not look a lot different," he says. "The differences will be in how it operates and how the pieces fit together."

We did get Lewis to talk a bit about how some of EMC's product segments will change, most notably backup and archiving, content management, virtualization, and security.

Contents:

Dave Raffo, News Editor, Byte and SwitchNext Page: Integration Man

Byte and Switch: Are customers demanding you integrate all your technologies into what you’re calling "One EMC"?

Lewis: We do hear loud and clear from customers that we want to drive better integration, more simplicity in products and more true integration across what we do. That's something that's been happening for years behind the scenes, and some you see in front of the scenes now but a lot of it is in the core underpinnings behind EMC.

Just because you see a certain degree of separation in acquisitions, I can assure you behind the scenes we are working diligently internally across EMC on integration.

Byte and Switch: It seems like integration always takes longer than expected in any storage acquisitions. Do you find that the case?Lewis: There's kind of a misperception about rapid integration that speed is always going to be the better way to go in acquisitions. I don't believe that. When you buy companies, those companies have customers, and they bought into the company because of its values as a separate entity. We've done a lot of acquisitions. They led us to hold separate pieces of functionalities and sales forces. In the case of VMware, we are keeping it very separate.

Since we started out, rule No. 1 has been to take our time and make sure we understood the value proposition of what we bought. I call it the "Don't break what you bought" rule.

Byte and Switch: Is there anything that sets EMC's acquisition strategy apart from other storage and technology companies?

Lewis: We're not into buying competitors. We're buying products that extend our capabilities into new and addressable markets. We don't buy competitors.

Byte and Switch: Can you give some examples of your product integration so far?Lewis: We're starting to do more cross-organization and cross-product things in EMC. First, we're making the right connections between products -- connecting Documentum and Centera for archiving, and Documentum and NetWorker for more efficient backups. We've even done new products like Infoscape, that's a brand new product built from our technologies out of Smarts, Documentum, VisualSRM from Astrum, even some components came out of [Legato] NetWorker. (See EMC Vows More for Infoscape.)

We're doing componentization across software platforms. We're taking the same exact components used, say, for management and security, and rolling them into common components we can use across all of our products. This is a very long term strategy to componentization. It will take years and not months.

Byte and Switch: How much different will EMC's product lineup look after all the pieces are integrated?

Lewis: The product lineup surprisingly will not look a lot different. The differences will be in how it operates and how the pieces fit together. We'll have pretty much all the products you see today, but a few key things will be different. There will be a much greater emphasis on solutions, from what you see on our Websites and what our sales force talks about. Interaction with customers will be incredibly solution focused.

Byte and Switch: You mentioned Infoscape before. Where are you with that product? It seems like it was a work in progress when you first announced it last year.Lewis: I would say as a product Infoscape is beating the startups. You have startups getting into data classification. We took pieces we had and built a brand new product line more effectively than the startups.

Infoscoape was 100 percent developed from five different product technologies that we put together. That was what made it great. We'll continue to add more robust policies. We have to identify use cases and solution sets for different levels. It becomes a framework to which we can base different levels of compliances -- SOX, FASB [Financial Accounting Foundation Board of Trustees], and other regulations that dictate compliance. It looks at workflows and policies that need to be set up. Then we'll connect it with more or our products, like products that do data movement.

Infoscape will be the policy engine by which you do more things for ILM. You have startups trying to run around building the same type of thing. Internal innovation is still around at EMC.

Next Page: BURA, BURA

Byte and Switch: What role with the Avamar and Kashya products play in your backup offerings? (See EMC Picks Up Avamar and EMC Coughs Up for Kashya.) Will they be part of Legato, work outside Legato, replace Legato?Lewis: All of those products coming together form our next generation backup or BURA -- backup, recover, and archive -- strategy. They all fit together, but we believe backup as done today and archiving as done today will fundamentally change over the next 12 to 18 months. The change will be driven by new technologies. The key ones are high-capacity disk drives, data de-duplication, and more and more need for archival information to be retained on disk. We're seeing 1-Tbyte disk drives, and the cost per disk drives further dropping. Look at those driving forces, there's a sea change in that market.

Everybody says backup is a static market, it's not growing fast. We believe it will transform from predominantly tape based to disk based, and the enabling features will be things like data de-dupe but also Centera archiving and Infoscape classification, characterizing and indexing data for discovery purposes.

The Kashya and Avamar acquisitions all tie to the fact we saw this change coming.

Byte and Switch: How are your acquisitions driving storage virtualization?

Lewis: EMC is heavily into virtualization. We believe in it, we believe in the technology. We believe in the use case of virtualization -- efficiency, dynamic data centers, moving resources and storage pools from one place to another.We've made bets and have technology plays across the board. We have network storage with Rainfinity and Invista for virtualization all across storage, for NAS and file systems with Rainfinity and block storage with Invista. (See EMC to Buy Rainfinity.) If you think about Centera and object store, it's already virtualization. What we're going to do more and more over time is integrate virtual objects from a management perspective. We don't care if you want to use blocks or files, you have a way to virtually use resources.

This is very much a strategy for us. Rainifinty remains a key piece of that strategy. Invista is going into many, many development environments. We expected it to go into development environments first, and over time people will use it in production.

We already have object, file, and block virtualization; the next move is helping to tie it together with management. That's the next connection point to tie it all together.

Byte and Switch: So Invista's not a bust? (See EMC Unveils Invista and Invisible Invista.).

Lewis: The reason for acquiring Kashya was as a complement to Invista. It's now sold as a complementary option, and it is being integrated. We want to have heterogeneous replication. Invista is right where we expected it to be, with large customers in large environments, but in the development area of those large environments as customers test the product and work out what changes they want to make in their business.Byte and Switch: Any plans for cluster file systems? You haven't made any acquisitions there.

Lewis: We do a number of things in scaling file systems, most notably Rainfinity, and the idea of creating a global namespace is important. (See EMC Refreshes NAS, SAN.) I would say clustered file systems in general have certain use cases we connect in with those systems.

Next Page: For the Archives

Byte and Switch: How do you see the relationship developing between Documentum/content management and archiving/CAS?

Lewis: There's full integration today, and we're working to make it more robust in terms of what we do. We're seeing an incredible uptick in growth in Documentum driven by more and more folks looking to use a suite set with content management. Instead of using Documentum for an application, folks are going to use content management across their entire content space. That's a trend we said would happen over the years.While people are working with files and file systems, it's becoming more and more about placing applications in Web services-types of environments, SOA types of things. SOA changes the way we construct applications by changing how application services interact. (See Wall Street Eyes SOAs and Users Send SOA SOS.) Content management is the next layer of value between applications and information.

Byte and Switch: EMC doesn't talk about File Area Networking (FAN) as much as some vendors. What's your take on FAN?

Lewis: As it's being defined in theory? We're doing that. What we're doing with Rainfinity and global namespace, we're completely ahead of this game with file area networks. Like with Invista, we've chosen to do out-of-band technology in a way that lets you create a file area network but doesn't disrupt the architecture between the filer and NAS environment.

Our competitors are just building clustered file systems or in-band boxes that are just NAS boxes -- back to the old days of Hitachi putting a block storage box in front of a block storage box and calling it virtualization. The way I'm going to virtualize my NAS is to have another NAS head.

Byte and Switch: Another product line you’ve already begun integrating is Smarts. How big of a role is Smarts going to play in resource management?Lewis: Smarts and model-based management is becoming pervasive across our management infrastructure. It pervades our thinking how we devise systems, how we look at discovery, provisioning, monitoring, and analytics around root cause. They're all getting in tune with model-based networking.

We've acquired nLayers to add to that. (See EMC Nets nLayers, Scopes Security.) nLayers does discovery by looking at networks and fingerprinting IP traffic. Smarts can connect all that IP information together, put it in a store.

We can discover an application, tell you where it's located and all the interdependencies, all without installing all the agents in the servers. We can say, “This SAN switch just broke, here are the applications using that switch and here are the applications that can be affected by that.”

Next Page: Locking It Down

Byte and Switch: [EMC CEO] Joe Tucci said you would embed security into all your storage devices and appliances. How long will that take?Lewis: It's going to roll out over time. You'll see this quarter on out, products adopting more security capabilities. We'll have common security platform pieces for sign on, authentication, authorization, embedding, encryption and more important global key management. It's something that will roll out in quarters and literally years before you look at ubiquity for encryption in every device and every disk drive and every element.

Ultimately, you'll be able to have encryption in every device that's information related. You will have to encrypt everything from cell phones to thin clients. We're focusing our efforts on how you manage encryption keys -– appliance based, host based, we do RSA application-based encryption. The secret is putting key management out there. You can create a lot of chaos if you don't have key management set up.

Byte and Switch: Are security standards developing fast enough, especially for key management?

Lewis: Standards are evolving. The market has been quick to want to add in encryption. The data security market is a market that a couple of years ago customers weren't talking about, and all of a sudden it became top of mind. It went from "I don't care to this is my No. 1 priority" overnight.

Byte and Switch: Isn't encryption still more important for tape than disk?Lewis: It will play large role across tape because of the physical handling. You're more worried about the physical loss of tape because there are a number of people handling it. Strategically, at the end of the day, once key management is worked out, encryption will become fairly pervasive in what we do.

Byte and Switch: Will software-based encryption ever become prevalent?

Lewis: Absolutely. In the highest performance areas there will be hardware assistance, but servers with multicore designs will enable enough performance to encrypt without hardware assistance.

Byte and Switch: Outside of encryption, what technologies are needed to secure storage?

Lewis: Our focus is on identity and access management, the security of our own infrastructure, and in compliance. Network Intelligence is a compliance-oriented product. It's security information management -- taking in audit logs and trails. (See EMC Pockets Network Intelligence0

SUBSCRIBE TO OUR NEWSLETTER
Stay informed! Sign up to get expert advice and insight delivered direct to your inbox

You May Also Like


More Insights