Storage Roundtable: addressing the issues today - Storage Automation

Computer Technology Review, Jan, 2002

When you think of mass storage, you tend to think of Silicon Valley, Route 128, or Colorado. The fact is that one of the nation's active storage centers is in the Twin Cities: Minneapolis/St. Paul. Recently, Computer Technology Review, in cooperation with Imation, assembled a regional storage forum to explore pressing storage networking issues. The transcript follows.

The first part of this article appeared in the October issue (Q3 2001) of Storage Inc., the second part in the November issue of CTR, and the third part in the December issue of CTR. The final part appears here.

Roundtable Participants:

Doug Ingraham is the Manager of Product Management for the Marketing for the Storage Router Business Unit of Cisco Systems, Inc.

Nick Williams is the Senior SAN Consultant at CNT.

Jim Ellis is the Director of Strategic Marketing at Imation.

Mike Hogan is the Storage Consulting Practice Manager at Imation.

Bill Peldzus is the Storage Consulting Marketing Mana-ger at Imation.

Rusty Rosenberger is the Director of Marketing and Business Development at Imation.

Jeff Tetzlaff is the Director of Marketing at I-TECH.

Carla Kennedy is the Vice President of Marketing for the Switch Product Group at QLogic.

Dave Sass is the Vice President of Marketing and Alliances at Sistina Software.

Brad Stamas is the Director of Storage Domain Management for StorageTek.

Erik Norlander is the Director of Product Marketing at Tricord Systems.

Bill Webster is a Senior Product Manager with Veritas.

Dan McCormick is the Director of Product Marketing Strategy at XIOtech.

Mark Ferelli: There's the very realistic pain of negotiating service level agreements (SLAs). What is acceptable performance? What is not? How much is the SSP willing to commit? How much is the user willing to trust? And what kinds of data are they willing to trust and outsource?

Doug Ingraham: Mark, the issue of SLAs is not new to the IT organizations. So one of the things you may be asking is if people are willing to accept an SLA from an outside organization, not about the issue of the delivery of their storage. It may be more of an issue of economics, but also can I trust the agreement. If I go outside, I may ask for things a little bit more detailed than I maybe assumed that my internal IS organization is going to provide. The SSPs like to think they are offering you a utility service within the IS organization today. So the question might not be what's it take to get to the utility, but what's it going to take to get to the utility that maybe is outside of the organization today that's shareable across multiple organizations.

Mark Ferelli: I'm sure the confidence issue ... it's a big one. What makes the negotiating of the SLA so vitally important is the fact that it's sort of the physical manifestation of the level and direction of the trust involved. Let me ask you folks now a question that's come up once or twice back in my offices. We talk about clustering as a strategy to help manage storage costs. What does clustering actually bring to the storage market? I think it's a network-based concept, but what does it bring to the mass storage space?

Eric Norlander: Certainly you're speaking to the heart and soul of our value proposition. I think fundamentally the concept of an appliance is a concept that's been proven. Cisco Systems certainly is the best example of that, as they took the basic bridging and routing technologies from the late 80s and early 90s and turned it into a very successful organization effectively delivering networking appliances. We are now at a time where we are able to create appliance solutions for storage problems and that provides really significant benefits in terms of the deployment and ongoing management. But, typically the problem with the appliance model had been how do you grow those environments without doubling, tripling, quadrupling the management burden that goes along with it.

Clustering technology allows you to do that. It allows you to go from a hundred gigabytes to multiple terabytes without increasing the management burden fundamentally. Tricord's clustering technology, which is fundamentally based on file level awareness, is a new technology introduced to the market in an appliance format that we believe brings together the best of all worlds for specific application solutions and point solutions. And we'll continue to see these types of requirements demanded by the end user community whether it's general file services or specific solutions around exchanged environments, web server environments, or data base environments. As long as we continue to face the complex interoperability issues that we see and have discussed already this morning, there will be a need for appliance level solutions that can scale without putting a new burden on the management and that's where clustering technology comes into play.

Mark Ferelli: Dave, file system awareness is very much a part of what Sistina is involved with. How do you see it as it applies to cluster environments?