In-Depth

IBM Revamps Flagship Storage Resource Management Suite

Support for z/OS is still gestating, however, and extending TotalStorage Productivity Center to OS/400 will be difficult, too.

• By Stephen Swoyer
• 09/06/2005

Last week IBM refreshed its TotalStorage Productivity Center suite, announcing new features and functionality.

What, you ask, is the TotalStorage Productivity Center (TPC)? Good question.

Big Blue inaugurated the TPC line last December when it consolidated several of its former Tivoli Storage Manager and IBM storage products into a single suite. Last week’s updates (released as part of IBM’s TPC version 2.3 refresh) were, in fact, the first to be released under the new TPC brand.

IBM officials say the TotalStorage rebranding was part of Big Blue’s ongoing infrastructure simplification initiative, which proposes to deliver management tools and other software to help abstract the complexity and heterogeneity of enterprise IT environments. That may well be the case, but many IT environments include mainframe and midrange systems – such as zSeries mainframes or iSeries minicomputers – and, for the time being, TotalStorage remains a (mostly) distributed systems-friendly play.

Not that IBM isn’t aware of this shortcoming. “TPC is basically the infrastructure simplification management package for all storage. So this is our command and control for configuring, monitoring, capacity planning, performance management, alerts, meshing with SAN fabric, [that] type of package,” says Scott Drummond, programming director for storage networking with the IBM Systems Group.

At the same time, Drummond concedes, TPC doesn’t yet support all of IBM’s eServer platforms. “We will allow for replication management from z/OS volumes to z/OS volumes [using TPC for Replication], but right now, we will support—we have very limited support for [z/OS] and OS/400.”

IBM has had several requests from customers to broaden support for z/OS and iSeries in its fledgling TPC suite, Drummond says. To the extent that it’s possible, he asserts, Big Blue expects to do just that. At the same time, Drummond cautions, it’s far from a straightforward problem. “Have we had customer requests about how to … manage the non-OS/400 open systems and z/OS, yes? The z9 platform has both [proprietary and] open systems—if you think about Linux and z/OS—so we’re working on adding support for the zLinux portion of z/OS, and we have thoughts on how to deal with the z/OS side of the z9.”

Drummond won’t commit to any time table for doing so, however. With respect to iSeries, there are several potential show-stopping bulwarks to TPC support. “On the iSeries product line, there’s two different worlds, there’s more than that, actually. About half the data is in the native OS/400 format, which is single-level storage, but they’ve also introduced AIX binaries inside of there, so there’s a lot of data that’s in AIX format,” he explains. “We could in the future relatively easy manage that [AIX] side of it, but the single-level storage, it’s hard. If [the iSeries] OS [group] gives us some sort of interface into their catalog, then we can do something, [but] currently right now I don’t see any way of doing it. At least we can do part of the user space on the iSeries.”

But What’s New?

The TPC suite consists of four products, three of which are based on former Tivoli- or IBM-branded offerings. All four products are still sold separately, Drummond says, but the quartet is also highly integrated. The four TPC-branded products are TPC for Data—the “brains” of the TPC stack, and the former Tivoli Storage Resource Manager by another name; TPC for Fabric (once known as Tivoli Storage Area Network Manager); TPC for Disk (formerly Multiple Device Manager); and a new product, TPC for Replication, which isn’t based on a Tivoli- or IBM-branded antecedent. IBM’s TPC Standard Edition 2.3 includes the first three products; the last is sold separately.

“TPC for Data is the eyes and reporting of the storage system. What happens is it will gather up information from all of the different hardware that’s attached and it has more than 300 [canned] reports that you can run [against that information],” Drummond explains. It’s in this respect, too, that customers who opt for the complete TPC suite can benefit from the integration between products.

After all, Drummond says, IBM’s TPC for Disk is designed to help manage disk capacity utilization, but it ships with only a limited subset of the reports that are available with the baseline TPC for Data product. “All of our [Enterprise Storage Server] Shark customers out there use this or earlier variants of it to provide performance management. The reason why you need to do this is because the storage control unit can be divided up betweeen HP, SUN, AIX, [and] Windows systems. Each one of those partitions, you can do an operating system[-specific] view of performance of that system, but you wouldn’t see a total box perspective of performance capacity,” he points out. “It’s also the case that a lot of the reports [for TPC for Disk] come from TPC for Data. Customers get a certain set of reports, a subset of reports, but we prefer them to buy TPC Standard.”

In the new release, IBM is adding expanded storage hardware support for its DS family and Storage Management Initiative Specification-certified (SMI-S) third-party systems. SMI-S, for the record, is a fledgling standard designed to promote management interoperability among heterogeneous storage devices. It’s backed by the Storage Networking Industry Alliance (SNIA), and is endorsed by all of the major tier-one storage vendors, including Hitachi Data Systems (HDS) and EMC Corp. Other new enhancements in TPC 2.3 include performance thresholding, cross-platform disk capacity management, and assignment reports.

“TPC for Disk has two different components. One of which is management and configuration of storage control units, particularly IBM ones, but more and more as people are compliant with SMI-S, you can interoperate with other vendors, as well,” Drummond explains. “[SMI-S] 1.02 is pretty much just reporting, and the more advanced management [features] will come in later releases. But right now, we’re one of the most compliant. HDS is doing pretty well, but we’re pretty much ahead of the curve.”

For SAN users, Drummond says, the dream of heterogeneous management from a single console—IBM’s TPC for Fabric—is effectively a reality. “If you add a SAN, a multi-vendor SAN, you have some Brocades and some McDatas, or some McDatas and some Cisco, this would be able to manage all of those heterogeneously. We have a lot of problem resolution mechanisms, or technically fault isolation routines, so you can tell whether there’s a break in a link.”

The last and newest TPC deliverable, TPC for Replication, more or less does what it’s name implies. Drummond positions it as a prerequisite for high-volume replication between distributed sites. “Imagine you’ve got a bunch of Sharks in two different locations and you’re replicating maybe 500 volumes back and forth synchronously between the two locations. Managing that by hand would be a pain, if not impossible.”