What we can learn from the mainframe world

Bridging the gap between storage for open systems and mainframe storage has never been a job for the faint of heart.

To some degree this is a curious state of affairs, because so much of what we have become used to in the open systems world has percolated its way to us from mainframes. For example, storage technology for the big iron has given the rest of us hierarchical storage management (HSM) and storage-area networks (SAN). And of course in the old days (when most computers were still running on coal and making a computer's core memory was actually a cottage industry of sorts) our most basic storage technologies, disk and tape, were developed to serve mainframes.

But mainframes, once described as dinosaurs, were supposed to have died a long time ago. It is thus something of an embarrassment to industry pundits that, despite a decade and a half of predictions to the contrary, the mainframes still haven't gone away.

The reason for this is that there are still many environments that seem to operate best on those big boxes: IBM, for example, is making a strong case for its zSeries servers in many corporate e-business environments.

This continued reliance on big iron has occurred because many of the "dinosaurs," like those thunder lizards of old, have evolved. In terms of storage connectivity efficiency, ESCON, the original mainframe SAN technology developed in the 1990s, has given way to FICON. Most IT shops have made a transition to the newer protocol, often assisted by switch makers InRange (www.inrange.com) and McData (www.mcdata.com).

Other areas of improved mainframe efficiency have been provided by the vendors of mainframe storage resource management tools (Tivoli and BMC, for example). And of course, some of those mainframes now run Linux.

As would be expected, the storage required to support those big boxes isn't shrinking (by some estimates, more than half of the world's corporate data still exists on mainframe systems). While the term "on mainframe systems" more likely means "on SANs for mainframes," the salient point here is that a lot of enterprise data, about half, is still stored for use by the big iron.

And the other half of the world's data isn't.

Given those facts, let's make two assumptions. First, assume a 50:50 split between mainframe and open systems storage (I'll be interested to hear from folks at enterprise mainframe shops so that I can verify the accuracy of this estimate).

Second, assume that a more-or-less equal amount of business-critical data is stored in both environments (if you are a mainframer and don't believe this, stop being so naive: just consider the reports, technical data and so forth that are distributed around your company on desktop machines and in research labs).

Most IT staffs that support both environments often resort to having one set of admins and tools for the mainframes, and a second set of both for the Linux, Unix, and Windows machines. Double the pleasure, double the fun? Hardly.

Why do so few storage resource management vendors offer solutions that bridge the gap between storage for open systems and storage for mainframes? The answer is that in most cases, vendors' development teams have been as split as the IT environments they sell to. Most mainframe tech teams lack knowledge about the way open systems run, and most developers of open systems tools know next to nothing about the mainframe world.

Enterprise IT managers however - particularly those of you at Global 2000 companies - live in both worlds, and surely benefit from any opportunity to centrally manage all storage with a single tool set, or lacking that, by managing multiple tools from a single console.

Unfortunately, right now there aren't a lot of vendors providing comprehensive solutions that cover both sides of the house. So, if you are looking for a way to consolidate management of mainframe and distributed systems, you don't have to initiate a very extensive search right now. Key players that bridge the gap are BMC (www.bmc.com), Legato (www.legato.com), TeraCloud (www.teracloud.com) and Tivoli (www.tivoli.com/), most of whom started with mainframe technology and then extended over to open systems through acquisition of existing vendors in that space.

Alternatively, you can look to virtualized storage. Virtualization of mainframe and open systems storage assets will of course offer alternative answers, but while the concept is valid, the jury is still out on the vendors and technologies that will be providing such solutions.

Mike Karp is senior analyst with Enterprise Management Associates, focusing on storage, storage management and the methodology that brings these issues into the marketplace. He has spent more than 20 years in storage, systems management and telecommunications. Mike can be reached via e-mail.

Storage archive
Past newsletters.