IBM's Storage Tank arrives missing half its ammunition

There are certain moments in the IT industry that require reflection in order to gauge their true significance, and IBM's delivery of Storage Tank is one such moment.

To call the creation of Storage Tank a journey does not do the technology justice. Those taking part in a journey tend to feel some sort of pleasure. Be it a soothing trek around, say, the Grand Canyon or a wrenching trip into one's soul, you end up in a better or more enlightened spot than where you started. This is not the case with Storage Tank - now known as the TotalStorage SAN File system.

It's hard to say for certain exactly when work on Storage Tank began, but sometime in 2000 IBM management decided to make the project public. A group of engineers at IBM's Almaden Research center hopped on their motorbikes (A must see - Ed.) and tried desperately to deliver the product as promised in early 2001.

Even in those first, humble days of the, er, journey, IBM assured customers that Storage Tank would be the file system of the future. It would link servers and storage systems from all vendors, making it possible to view and access a file from any system. While other vendors hammered away on their unsophisticated hetrogenous software packages, IBM was refining something full of grace and elegance.

But here we are with 2003 drawing to a close, and IBM drops the news. Storage Tank will not arrive until November, and when it does, the software will only support IBM's own storage gear, AIX and Windows.

"We've made a lot of announcements about this product and talked about it a lot," said Bruce Hillsberg, director of storage software strategy at IBM. "We think this is the right way to introduce it."

Let us pause, reflect and return to IBM's Storage Tank Web site.

"The goal of the Storage Tank project is to provide a complete storage management solution in a heterogeneous, distributed environment."

Or take a trip back to April of 2002.

"The code is up and running," said Linda Sanford then senior vice president and group executive for IBM's storage systems group. "We have shown Storage Tank handling databases, file serving and file sharing with various applications. We will introduce it when it's ready to perform flawlessly and our customers tell us they have what they need."

Somewhere along the line the terms "flawless performance" and "goal" seem to have been altered.

It turns out that customers will actually have to wait until the middle of 2004 to see support for more versions of Unix, Linux and storage systems from other vendors. Exactly which ones? No one knows. IBM will not say.

"We are not being specific at this point and time," Hillsberg said. "The major reason is that we have fairly rigorous testing and qualification work that we do on the boxes."

Ah.

If you are an IBM customer that does not give a hoot about hetrogenous SANs, then you're in luck. IBM is selling the SAN File System starting at around $80,000. This price includes the file system and a pair of Intel-based servers running Linux, which is where the software sits. The servers look out over the network, keeping track all the files in a SAN.

The "goal" is to make it easier for customers to track and manage data across large networks. IBM installs a virtual file system on all of the servers in the network, which makes it possible to track file metadata.

Like other similar products, IBM promises that the technology will help customers automate many management processes. An admin can set up policies that govern, for example, what type of storage system certain data should run on and when a file system should expand.

Storage Tank also helps with load balancing and fail-over operations. If a server failure occurs, the software will work to make sure that files are offloaded onto the right kinds of storage system and backed up as needed.

IBM sees SAN File System as a complementary product to other software packages offered by Tivoli that handle various management tasks.

All in all, IBM still bills Storage Tank as the next big thing, and who knows, it may be right.

IBM excels at handling complex storage functions such as long distance backups and disaster recovery. It has no doubt built the best of its technology in those areas into Storage Tank along with day-to-day management functions.

The massive delays in rolling Storage Tank are common for any complicated software package. IBM's biggest mistake was touting Storage Tank so early and so often.

IBM is assuring customers that Storage Tank will be crucial to handling huge amounts of data over SANs that continue to increase in size. Customers such as Johns Hopkins University and CERN are already plugging away with the code on massive installations.

Storage Tank will mostly be of interest to large, IBM shops in the near term. For the rest, the journey continues. ®