Storage vendor EMC has kept its focus on the enterprise. Although other vendors have extolled the value of breaking down corporate computing into departments, EMC has stressed the importance of keeping information -- and therefore its storage architecture -- central. Now, with the emergence of storage area networks (SANs), which centralise data storage onto their LAN, EMC's philosophy is catching on. IDG reporter David Pendery recently sat down with EMC's CEO and president Mike Ruettgers to discuss what's new in the storage marketIDG: What were the most important trends in data storage and retrieval during 1997?
Ruettgers: I think the clear trend was the consolidation of storage. Information that was being distributed all over the enterprise is now starting to come back. The fact is that people want to share this information among various platforms, so this idea of enterprise storage is clearly what we are seeing. I think information has led the consolidation [of enterprise hardware] and not the other way around.
What are the most promising new data communications technologies and protocols?
Certainly from the hardware side, Fibre Channel is big. What Fibre Channel allows you to do is provide better connectivity. But equally important, [it allows] you to have the server at a further distance away from the information. The other side is some of the software today allows you to share information between different platforms, or to have a platform and have it store information both on a mainframe and [on open systems].
What do you think of the Intel/Compaq Virtual Architecture Interface to speed data transfer among clustered servers on a SAN?
I think anything you can do to put standards in place helps. Frankly, Unix would be in better shape today if there were a set of standards provided so that people had a Unix which was the same. The fact that it is still proprietary is making it vulnerable to things like [Windows] NT, because NT is produced by one company and comes with one set of standards. And you can just see the frustration people have with dealing with non-standards hardware.
What is the status of the "next-generation data centre" proposed in October 1997 by EMC, Intel, Sequent, and Oracle?
We've been trying to sponsor companies working together to do these kinds of things. This [project] was designed to allow people to look at how combinations of certain technologies can do things for them that they couldn't do previously. So we've been running a series of seminars that explore this with selected customers, so customers have some sense as to what is possible in the future when you put these technologies together.
Does the SAN figure prominently into EMC's plans? Is it a significant development in storage technology?
Well, I think this is frankly "wordsmithing" around what we have already been doing in enterprise storage. We have seen this view for some time. I think it's a validation of our view that if you organise around information, that is a more useful structure. We believe that's true. So this should help the laggards that don't understand.
To what extent do you think storage products simply need raw capacity, and to what extent do they need to be as "smart" as the data or applications that they are storing?
It's clear that as you put information in the middle, you need to be able to do a bunch of things. You need to manage it, protect it, share it; and dumb storage systems can't do that. Almost everybody I talk to understands that cheap is not the way to go.
How does the enterprise's adoption of NT affect EMC's design and engineering plans?
It's clear that large customers are exploring NT because they believe the predominance of the new application software is going to be written for NT. We were part of the [Scalability Day] that Microsoft did [on May 20, 1997] to show that you can, in fact, have large systems on NT and protect your informationI think it is clear that over time, NT is going to be a substantial operating system and will probably displace the weaker operating systems. I think they are going to take a huge [amount of] share from [Unix].