HP recently hosted its EMEA Analyst briefings in order to update Quocirca and other industry watchers on its current capabilities, and its current and future strategies. The message was pretty clear: that HP is placing a high importance on blade architecture for hardware; that it wishes to dominate the storage market; that it places a relatively high importance on a “green” message; and that it expects significant revenue to come from its “mainframe alternative” programme. Clearly the EDS acquisition is no small thing in terms of how it will shape the company’s path and modus operandi in the years ahead. However, next to nothing was given up on that topic—not for lack of probing—and it remained the unacknowledged elephant in the room.
Of all these topics, first amongst equals was the Mainframe Alternative Program, which HP went to significant effort to talk up. To ensure that there was no mistaking the level of importance HP is placing on the programme, it went as far as hosting the actual briefing sessions in Madrid so that all assembled could bask in the atmosphere of the mainframe alternative “Centre of Excellence”, which is located just outside the city. HP points to a proven and honed delivery methodology, executed through this centre, as being the reason for its successful conversion project track record. So perhaps it is fair to say that those staffing it deserve their day in the spotlight.
HP labels the programme as offering a low cost, “open” systems alternative to what the vendor labels IBM’s “closed” and “legacy” mainframe platform and claims a flawlessly delivered record of forty customer conversions in EMEA. Sometimes briefing sessions are an exercise in the “tell them enough times and they’ll eventually believe us” style of persuasion. Hence the emphasis here of the words “closed”, “legacy”, and “open”, as these labels were oft repeated. So much so that they deserve examination, for behind those labels everything isn’t necessarily all that it seems.
The main targets of the conversion efforts are IBM zOS and VSE-based workloads, though other vendors such as Unisys and Bull are also on the hit list. HP does have a point when it labels vendors such as Unisys and Bull as “legacy” and it would be a rare CIO who would argue that those vendors are committed to supporting a viable and innovative processor strategy. Meanwhile, since the demise of Amdahl and the withdrawal of Hitachi from the IBM clone market, it is true that IBM is the only stable in town for those wishing to run VSE or MVS derived operating systems. Whether that is a good thing or not for the average CIO is another matter; what is relevant here is that when HP uses the word “closed” in referring to IBM it really means “single vendor”.
The use of “closed” as a synonym for “single vendor” is an important one, because HP also describes it's mainframe swap-out programme as “unique”. It regards that uniqueness as a market differentiator that provides key competitive advantage over other open systems vendors. Therein lies the problem. You can’t have it both ways and label the choice of a single vendor as being closed in one breath, and open the next. CIOs who strip their jockeys off the horse supplied by the IBM stable and take them over to the unique vendor “open” alternative will quickly find themselves saddled to a similarly narrow field of hardware and software choices.
Of course, every IBM mainframe is also running a wide assortment of ISV software too and that will need to be swapped out for an “open systems” alternative. While Oracle and Microsoft get a look in as replacements for DB2, and Oracle/BEA’s Tuxedo is the HP recommendation for replacing CICS/IMS; HP software stands first in line for everything else. HP recommends HP software as the replacement for any Compuware, BMC, CA or other systems management software you might currently run, and backs that up with the expected price breaks.
The notable exception to that selection strategy will be an open systems alternative to IBM’s RACF, or CA’s ACF2 or TopSecret, because with the end-of-lifing of HP’s Openview Select Access product, HP doesn’t offer a host-based access control solution. Both the assembled set of HP Mainframe Alternative experts, and the reference customer who took us through his experiences, assured us that “open systems don’t need anything like RACF”. According to HP, none of the forty customers who have already swapped out their legacy mainframe for some new HP kit even asked for an alternative. In Quocirca’s opinion HP Hubris v1.0 should not be solely relied upon to be a functional equivalent of RACF or its ISV alternatives. In fact, there are a number of open system RACF alternatives that might be considered: CA’s Access Control product and Tivoli Access Manager For Operating Systems being the two closest contenders for equivalent functionality. There are others too, but at least considering those products gives the advantage of being able to compare like with like, given that you’ll be running either vendor’s security software already on your IBM mainframe. Quocirca recommends that security considerations be placed at the forefront of any mainframe workload migration consideration and proposed architecture.
Looking at the financial case for the move is also enlightening. HP uses a spreadsheet-based ROI calculator to support its claim that HP’s hardware and software costs will be substantially less than your current mainframe costs. Even factoring in the conversion costs themselves, ROI is claimed in the third year. If nothing else, these numbers are worth looking at and then waved in the direction of your mainframe ISV and HW provider to see if the wheels of negotiation are greased. But again, there is more to it than first meets the eye.
HP claims that almost all mainframe workload can be migrated to its platform. Easily and fully transportable workload includes off-the-shelf packages such as SAP, which is of course available on either platform. CICS or IMS transaction workloads move fairly easily to Tuxedo. In order to reduce end-user retraining efforts you can even retain the 3270 look and feel, including a recommended ISPF emulator. Batch can also be fairly easily moved to shell scripts, while JCL gets converted to a syntactically similar open systems alternative. COBOL is similarly supported these days on UNIX. The further you go down that stack though, the more difficult, expensive, time consuming and risky the conversion effort will become. Any undocumented COBOL programs will probably need to be scrapped and their functionality rewritten. That also applies to any Assembler programs you have today, none of which can be flipped to the new platform.
The HP alternative might indeed be cheaper then; if you can manage to actually decommission the mainframe completely that is. If you can’t, and you probably can’t unless all you run on your mainframe is a fully transportable workload, then you’re going to end up running both systems in parallel. Unfortunately one potential pesky side effect of having to run both the legacy mainframe and the new HP kit is that your energy bill, and therefore your data centre emissions, will go up. Indeed HP’s own calculator showed that such an outcome was indeed to be expected. It goes almost without saying that this is not a good strategy.
In short, the staff of HP’s Mainframe Alternative Centre of Excellence deserve an “A” for managing to deliver forty successful mainframe swap-outs. However the vendor gets a “D” for grammar due to its confused use of the word “open” in describing the benefits of the alternative. It also gets a “E+” for driving up a data centre’s energy usage. Lastly, it gets an off-the-scale “G” for its dismissive stance on security, mainly because the word “gobsmacked” comes to mind when thinking about it.
Migrating some of your current mainframe workload to the HP alternative may make sense for you, if for no other reason than to avoid an impending MIPS expansion by freeing up headroom through offloading some work. Some CIOs find themselves with mainframe skills shortages too. Although at an industry level, it is worth asking whether the scope of that problem is somewhat overstated by those with a vested interest. In a world of ever changing technology alternatives, Quocirca recommends that CIOs continue to measure the performance of the status quo. However, any potential changes should be given thorough consideration that goes beyond vendor positioning.