Veritas Software Corp. has been strengthening its utility computing offerings through the acquisitions in the last several months of three software developers: Ejasent Inc., Precise Software Solutions Ltd. and Jareva Technologies. Veritas CEO Gary Bloom believes his company, with its established leadership in storage management and independence from any particular hardware platform, is well-positioned to help customers better exploit their IT resources by centrally managing distributed computing resources. eWEEK Department Editor John S. McCright and Senior Writer Brian Fonseca sat down with Bloom at Veritas headquarters in Mountain View, Calif., recently and discussed the companys present and future.
Veritas has done several acquisitions. Where are you putting in the most time these days?
We have tried not to move away from our strong markets of storage, data recovery and high availability, and use that as a platform to extend into what we think is a pretty practical approach to utility computing.
There are two forms of utility computing that have generally been emerging. One is the completely integrated vertical stack where you buy it all from one vendor. The alternative strategy is one IBM is supporting: Buy portions of the stack and then use services in a big way to integrate them and put them all together. We have a little bit of a contrarian view, which is, “Lets provide the building blocks necessary to enable it and let the customers use as much of what they have today to move in the direction of the layers of utility computing,” which is pretty simple: availability, performance and shared infrastructure at a low cost.
Why do you consider storage a key building block to enabling utility computing?
If you look at storage as an industry, its probably one of the few things in an IT shop that is managed today as a utility. In other words, how much of storage is highly distributed anymore around enterprise management? How much of backup-recovery availability of storage is managed in a distributed fashion? Essentially, very little. Its pretty centrally managed, the hardware is commoditized, and customers are now starting to try to get better value out of storage theyve already installed. Its generally pretty available; its recoverable; and its run on a shared, automated infrastructure; so as a building block, its probably the closest thing the industry even has today toward [what] a utilitys going to be in the future, being much more broadly deployed.
Focus on Performance
One aspect of utility computing is enabling an IT department to charge business units for the computing resources they use. Has Veritas seen much of that?
Very few people I talk to want to get in the habit of sending a bill out to the end users. What they do want to do, if you talk to CIOs, is give those users visibility into what they are consuming and what the cost is. When we talk about data recovery and data protection, I want you to decide whether you need 30-second recovery or whether three-day recovery is OK.
As an independent software vendor, how dependent are you on APIs? Do you have to play catch-up in what you need to manage?
On this API area in particular, youre making an assumption that the vendors determine the competitive landscape there, and the reality is the customers determine that landscape. And so, if we were left to our own devices—with IBM, Veritas, EMC, and others all playing in the same sandbox friendly with APIs—well, probably not. But when you add the customer pressure that says the customer wants interoperability, the customer expects this part of cross-vendor solutions, that problem goes away. And right now, given our market share, and our markets, and given our success … Yeah, any time we have a new API challenge the customer has stepped in. [Customers say to EMC, for] example, “If you want us to use your storage, youre going to have to support Veritas.” That changes the dynamics dramatically.
Any particular emerging standards efforts that you need to focus on?
There are multiple standards efforts in storage. As Im sure you know, standards efforts tend to be promoted by those who believe they have the best competitive advantage out of them. Are we a fan of standards? We absolutely are. We think it neutralizes any of this API debate completely. And then we have very strong confidence we can out-engineer the competition—bring better, more-useful products to the market for our customers. The reality when you bring all that together, though, is what does the customer want? Well, one thing the customer doesnt want is hardware-vendor lock in. They dont want to return to a scenario under which they belong to a hardware vendor. And our technology helps insulate them from that, so its an insurance policy against vendor lock in; its not about strategy.
Veritas Vendor Glue
Is Veritas becoming more involved with the hardware pool of out of necessity?
First of all, we dont ship any hardware. Were going to have partnerships with all the key suppliers of servers, key suppliers of networks, key suppliers of storage–we live off of those partnerships; our customers expect them. I dont know any other way to provide high degrees of interoperability between different technologies. [In] some respects, were the glue that holds all of those hardware vendors together.
Where are you with utility-computing deliverables? Any surprises with either customer acceptance or lack of it?
You have to separate between where were at and where customers are at. We never believed that this is going to be a trend thats going to happen overnight. We think we have the vast majority of the key building blocks—especially with our recent acquisition of Ejasent—to give us that one last piece, which is the ability to dynamically move applications around without having outages. If you cant automatically provision storage, automatically provision servers, and move applications around, its difficult to achieve utility computing because it suggests you have to shut things down to do migrations. We like where were at in the curve. We think its going to be a move toward utility computing, not a race. Its going to probably a bit more of a marathon than a sprint.
How do you decide which technology is the best to acquire to push your business ahead in new areas?
When you have a company thats doing well, you want to keep your competitive advantage and use acquisitions smartly; you also dont want to be too disruptive. As I look and counterbalance EMC—trying now to integrate three large acquisitions simultaneously, run all three of them completely independently with separate sales forces—I think its going to be a real challenge for them. We dont think thats the way to do acquisitions. We think its all about integration, putting them together and making them aligned with each other.
Practical Utility Computing
You talk about utility computing, and we think of grid technologies. How do you see grids fitting into utility computing, and what do you think of Oracle Corp.s approach to grid computing with its 10g technology?
Oracles probably been the most confused vendor of all about what grid computing is. Typically grid is that high-end scientific computing area, much less than what utility computing is, which is a much more practical approach to running business computing.
Theres nothing wrong with [Oracles strategy] if youre 100 percent on Oracle. So if all the data youre running for a given application is in the Oracle database, and theres no data in DB2, no data in SQL Server, no data in the file system and your entire environment is dedicated to Oracle and you use [Oracle] Application Server and applications are loaded in their database…Ive just described that portion of Oracles customer base that represents less than 1 percent of [all its customers]. I know [Oracle CEO] Larry [Ellison] has been aggressive about saying, “With 10g you dont need Veritas anymore,” and its wonderful that he has no idea what we actually do.
What is your biggest challenge?
The biggest challenge for us is to manage the unpredictable nature of industry consolidation. Given that at a lot of companies theres a big financial discontinuity between a lot of these players. The more heterogeneous [a companys IT shop], the more vendors [represented in a companys IT shop], the more I like it. With that said, I dont think that whats happening in the customer base today [with users saying], “Im going from hundreds [of vendors] to 20” is bad. We dont want [just] one—and as long as there are three or four vendors in one category, were fine.