Future Of The Storage Industry In The E-commerce Era.

Editor's Note: CTR presents the following chat transcript as a reflection of the kinds of questions your customers have about storage. SearchStorage (www. searchstorage.com) moderated the chat.

Moderator: Welcome to our Live Expert Q&A Session with Mark Ferelli, editor-in-chief of Computer Technology Review We are taking questions about his views on storage industry trends. If you have a question, please type it in the bottom of your screen and press Enter.

Q: How far off is true SAN interoperability?

Mark Ferelli: Closer than ever, now that the real hang-up--switch interoperability--has been settled and submitted for standards approval. It also handles the zoning challenges. It took the switch vendors long enough, but they made it.

Q: Do you think serverless backup is truly the "killer app" of the SAN?

Mark Ferelli: At this point, backup is certainly the central application of SAN.

Q: What do you consider the top three breakthrough storage technologies of the '90s and the top three you're predicting for this decade?

Mark Ferelli: The top three breakthrough storage technologies...how long is a piece of string? One would be the evolution of storage architectures, such as SAN and NAS. Another would be the storage application of high speed interfaces, such as Fibre Channel and IP. Finally, I think the concept of serverless backup is tremendously important. Upcoming, look for the improved use of IP in storage connectivity, an improved, post-GMR head for hard disks, and improved tape storage speeds and feeds.

Q: Do you see a trend developing whereby more and more companies are building their IT architecture around their storage product rather than adding it on later?

Mark Ferelli: I think that computing stands on three legs: processing, connectivity, and storage. You're right: storage has been historically an afterthought. Not any more: storage is now the biggest investment component in computing. It's not the center of planning, but it's now an important, well- considered element in planning. Add the fact that vendors are more and more adopting a partnership model for selling, and you have storage properly considered in the mix.

Q: What effect will storage as a utility (the SSP model) have on the industry?

Mark Ferelli: The SSP is a new breed of cat: both a consumer and producer of storage products and services. The availability of storage services will push enterprise IT folks into a "make or buy" decision on elements of their storage vision. The critical question is basically: "How much do I feel comfortable in outsourcing?"

Q: I hear a lot of talk about Fibre Channel. How do you see it affecting future installations of storage systems?

Mark Ferelli: FC has done fantastic service as an enabling technology for storage area networking. The usefulness of the technology in terms of speed and reliability are unquestioned. But as a newer technology, the price is still high...about $800 per port. The future depends on cost and development issues with Ethernet, Gigabit Ethernet, and Infiniband.

Q: How attractive is the segment (e-biz and e-commerce) in this era of Startups? Are they willing to invest in Enterprise solutions? How do you approach them?

Mark Ferelli: E-commerce has become a very hot application for storage. Right now, e-commerce solutions are not really enterprise level...I think that they are more of a departmental solution. Actually, it would behoove the departments to learn a bit from their IT department colleagues down the hall.

Q: What new standards for storage can we expect in the coming months?

Mark Ferelli: Standards are like ice cream: they come in different flavors. De facto standards, standards from half a dozen standards bodies, official and otherwise. Sorry, but your guess would be just as good as mine.

Q: What type of storage going forward will be most desired by e-commerce companies in light of the environment on Wall Street now that they must be profitable?

Mark Ferelli: In the short term, it is certainly Network Attached Storage: NAS. Forecommerce, speed is of the essence. NAS is easy to install, easy to use and a market sleeper that is likely to soar with eagles.

Q: Regarding the question about how far SAN interoperability is - does that mean next year, two years from now or roughly when we will see common solutions available?

Mark Ferelli: Certainly within the next calendar year.. .keep an eye on next Comdex.

Q: What do you predict will be the cost of storage a year from now (GB/MO)?

Mark Ferelli: I'm sorry, but accurately predicting costs is well-nigh impossible...there are just too many variables. Availability of materials, production yields, all of this makes a difference. And what happens if some aggressive marketer sacrifices profitability for market share? Again, your guess is as good as mine.

Q: How long will it take to get to a consensus on standards for SAN, or will there be more than one standard?

Mark Ferelli: It is likely that there will be more than one standard as SAN software matures and SAN interconnectivity settles into its various camps.

Q: Serverless backup may be the "holy grail" of SAN, but what about true file sharing, for example GFS - when do you think we will start to see this sort of technology become available and standardized?

Mark Ferelli: Available...it's on its way. Standardized: it is just too new...don't hold your breath.

Q: What role do you see tape playing in both the short and long term?

Mark Ferelli: Tape continues to be a major player in storage in all aspects of computing. Especially in the automation space, tape is important. It is essential to SAN strategies. There's been a lot of nonsense about the end of tape technology. Don't believe it. For backup outside of the e-commerce space, for disaster recovery, for archival applications.. .tape continues to be the cost-effective, time-tested solution.

Q: How long do you think the concept of "buying storage as a utility from a third party" will take to catch on?

Mark Ferelli: The storage services concept you're identifying will catch on only as fast as the IT shops gain a confidence level in the services. The utility concept will only work when IT knows that their data is available reliably when they want it.

Q: What do you think are the great unsolved problems in storage and storage management?

Mark Ferelli: From a user perspective, the greatest unsolved problems in storage revolve around price/performance/availability. The greatest problem with storage management is the manpower requirement. Storage has to be managed when you're talking terabytes or petabytes, and qualified IT manpower is getting scarce. There will be 120,000 new IT jobs coming up available, but the schools will graduate only about 20,000 IT majors.

Q: Do you see tape becoming more or less "involved" in future storage considerations?

Mark Ferelli: Tape solutions will be consistently involved in storage's future. Ask any company who deals with auditors about tape backup. And that's just one example. Tape continues to be viable...one analyst of my acquaintance notes that the combined annual growth rate for tape automation is about 25% per year. You tell me...more or less involved?

Q: With the trend to 24x7 and huge storage facilities, is the backup process as we know it, obsolete?

Mark Ferelli: I don't think the backup process will be obsolete, but it will change. A lot depends on individual IT duty cycles, demands, and so forth. Disk will take over from tape as primary backup in some cases. The IT center needs to be sure, on its own part, not to overspecify.

Q: What comments can you make about the reality of holographic storage systems?

Mark Ferelli: Holographic storage is a work in progress. Lucent and Imation are working together, and I saw their model. There's still a lot of miniaturization to be done there. IBM reports that they are still dissatisfied with media materials. I expect the technology to be developed in due course, but all the problems of commercialization still loom ahead.

Q: Do you believe large disk arrays (e.g. 180GB drives) will become more cost effective than tape libraries?

Mark Ferelli: Cost is less the issue than applications. Drive arrays will have to include value-adds, which drives prices up.

Q: Given the relative complexity and immaturity of SAN, would you expect a near-term resurgence of interest in NAS?

Mark Ferelli: No doubt. NAS's ease of installation and immediate uptime meets a lot of user needs.

Q: Do you think e-commerce startups want to go towards SAN or NAS or both?

Mark Ferelli: NAS first. E-commerce operates on speed...order entry, inventory, fulfillment, billing. SAN later as the volumes increase and data traffic needs to be eased to clear the decks for the applications.

Q: Would you equate the current situation with SAN standards and their evolution with the early state of LANs with Token Ring and Ethernet?

Mark Ferelli: An almost exact correlation, I think. History is certainly repeating itself...let's hope we've learned some lessons along the line.

Q: What do you think will be the next evolution of tape technology?

Mark Ferelli: Be on the lookout for the re-introduction of optical tape to the marketplace. The government/military segment is continuing to invest, and head and media technologies are well in development. Heads in 18-24 months, I think.

Q: Do you see a trend that more and more companies are moving to outsource their storage IT management and assets?

Mark Ferelli: Outsourcing is popular, but needs time to inspire confidence in the various strategies and service providers' offerings.

Q: Some people suggest that enterprise-wide SAN implementations will be hindered because various departments won't want 'their' data out on the SAN, available to others and beyond the department's 'control.' Could you please comment on this?

Mark Ferelli: Control is a significant issue. But that is not the pacing issue with SANs. Now that interoperability problems are fading, SANs need more mature software tools and a serious look needs to be given to all levels of data security, well beyond the storage segment.