In-Depth
Big Blue's z9-Centered Information Utopia
IBM's vision puts the mainframe in the center of a distributed application ecosystem that consumes data served up by zIIP workload engines
Two months ago, IBM Corp. announced a new addition to its family of mainframe processor engines—the z9 Integrated Information Processor, or zIIP for short.
At the time, IBM officials positioned zIIP as a low-cost workload engine (a la the Java-oriented zSeries Application Assist Processor, or zAAP) that can reduce software costs for some BI, ERP, and CRM workloads.
There was more to it than that, of course. In fact, according to one industry watcher, Big Blue’s new Big Iron processor engine could be a good solution for the data access woes that bedevil most large enterprises—if IBM decides to market it as such, that is.
Richard Ptak, author, long-time industry watcher, and (nowadays) a principal with consultancy Ptak, Noel, & Associates, says it’s not too much of a stretch to imagine a z9-centered Information Utopia, populated by a distributed application ecosystem that consumes data served up by zIIP workload engines. If it sounds like a pipe dream, Ptak argues, it could hardly be worse than other “seamless” or “transparent” data access cure-alls.
“[D]espite all the recent product innovation and [the] creation of [new] ways to share data, from common repositories to keen new ways to network and virtualize data stores, to schemes and techniques to optimize access—problems remain,” Ptak writes.
The good news, he argues, is that business and IT executives know they have a problem. The bad news—or the proverbial rub—is that they feel powerless to deal with it. “[D]espite understanding and [accepting] the critical … role data plays in their businesses, seamless access to that information remains an elusive goal for most organizations.”
Enter zIIP. The idea, IBM claims, is that organizations can tap zIIP to inexpensively expose mainframe data (specifically, DB2 for z/OS version 8 data) to distributed BI, CRM, or ERP applications via SQL calls using DRDA over TCP/IP. In this respect, Big Blue’s newest processor engine is qualitatively different from its predecessors, Ptak says. “[O]ur interest was [piqued] by the role the System z9 Integrated Information Processor plays” in IBM’s data access vision. Ptak notes that the zIIP is not only Big Blue’s fifth overall specialty mainframe processor but also the third (along with zAAP and Big Blue’s Integrated Facility for Linux engines) mainframe processor that’s specifically designed for new workloads. (Another specialty engine is Internal Coupling Facilities, or ICF, which provides extra processing capacity to power the execution of Coupling Facility Control Code (CFCC) in coupling facility (CF) LPARs.)
Ironically, says Ptak, zIIP is a new spin on an old—and once antithetical—idea. “IBM’s strategy updates one used by minicomputer vendors from the late 70’s into the mid-90’s. Back then, the strategy was to ‘surround the mainframe’ with applications optimized to run on minicomputers,” he explains. “This reduced mainframe workloads, [freeing up mainframe systems] to concentrate on ‘bigger’ problems while, not so incidentally, increasing minicomputer sales.”
IBM’s vision of a zIIP-powered, z9-based Information Utopia hoists this strategy on its own petard, Ptak argues. “Coming full circle, the mainframe aims to attract specialized, computer-intensive RAS applications with specialized mainframe ‘engines’ such as the zIIP. These engines, conveniently and economically, also have the potential to reside on the same physical system. Adding to historic advantages, this approach leverages [the] unique architectural and performance advantages of the IBM mainframe in terms of data consolidation, workload management, end-to-end view, and manageability.”
Ptak is quite clear: a zIIP-powered, z9-based information hub isn’t for everyone. But for organizations that already have significant investments in mainframe hardware, it’s a compelling—and potentially tantalizing—vision.
Elsewhere, he’s encouraged by healthy mainframe revenues (which IBM pegs at 5 percent year-over-year) and robust growth in MIPS (28 percent year-over-year), which he says are proof positive that Big Blue’s strategy is working. “IBM is finding [that] mainframe special function appliances appeal to existing clients and [are] attractive to new ones,” he concludes. “IBM will continue this approach as customers see [that] their demands for MIPs [can be made more] accessible without the complexity of processor farms.”
Related Article:
About the Author
Stephen Swoyer is a Nashville, TN-based freelance journalist who writes about technology.