IBM’s Z series mainframes stand alone with the strongest security and reliability in the industry. However, they typically require a dedicated, unique rack, which makes it harder to implement in an existing datacenter that might not have the floor space or in a remote location that doesn’t need a full-sized mainframe but does need local processing power due to unreliable data networking or unacceptable distance-based latency.
Well, this month, IBM addressed this problem with a new rack mount Z16 LinuxOne configuration that should help IBM expand its mainframe footprint while addressing the needs mentioned above.
Let’s talk about that week.
The irony of mainframe edge computing
The mainframe is a centralized resource. Even though it significantly preceded “the cloud,” due to its historical I/O, security, and reliability advantages, it was a better cloud solution than what is currently used there. The problem is that the market has been moving away from fully centralized computing architectures for some time now. Instead, it is embracing a concept the industry calls “edge computing,” where the computing capabilities are moved closer to those using them. The result is designed to better survive adverse conditions like major weather events, outages, wars, and other natural and man-made disasters.
As a monolithic architecture, the mainframe doesn’t lend itself to edge computing in its original iteration. At least it didn’t. The irony is that even though the mainframe (particularly in its LinuxOne form) is uniquely capable of providing the kind of critical, unfailing service the market requires (particularly banking, healthcare, and government), its lack of ability to be distributed not only limited its growth but put the platform at long term risk.
Mainframes are considered a centralized resource supplemented by mid-range computers like the near-legendary AS400 (one was once walled up for a decade in a room with no doors, and it happily continued to function without anyone even being able to touch it). Once the AS400 became obsolete, there wasn’t a great solution to move this kind of massive reliability, security, and I/O capability to remote offices or into existing data centers that were already space constrained.
Now that IBM has developed the rack-mounted version of its latest Z16 mainframe with significant price-performance advantages over its more traditional form factor but with the same reliability, security, and I/O advantages that the mainframe is famous for, IBM now has a viable way to provide mainframe capabilities at the edge and for data centers that don’t have the floor space for a more traditional mainframe configuration.
I expect this advancement will play best with existing IBM accounts that need this edge computing option or need to add capacity but lack the floor space in their data center for it. This is because those that don’t have mainframes are unlikely to buy them despite their advantages for several reasons: because they aren’t already IBM customers, they likely don’t have anyone on staff who understands the mainframe, and IBM isn’t marketing the advantages of the platform in a way that would expand its installed base. (Convincing an enterprise to try out new technology typically requires a massive marketing campaign, something IBM hasn’t done since Sam Palmisano eliminated Lou Gerstner’s impressively capable and well-funded marketing organization). This is a shame because that organization was fully capable of marketing a solution like this effectively.
However, this will reduce competitive pressure to displace the IBM solution and increase sales opportunities for IBM inside existing IBM accounts which, in a down market, would certainly be a win.
Doing a rack-mounted mainframe is amazing and helps address the unmet need in IBM Mainframe shops for a solution that will fit into existing racks or that can be placed affordably into remote offices to address edge computing needs. Given the market downturn, this release could provide an offsetting sales opportunity that might minimize the impact of this downturn on IBM.
Given that IBM is moving from 7 to 9 decimal places for uptime and that companies are increasingly valuing uptime and security, I wonder what might happen if IBM were to roll out a demand generation effort like it once did under Gerstner? Regardless, this announcement is a very positive one for IBM and its customer base.