The Value Of Things That Don’t Change–And Those That Must
March 28, 2016 Timothy Prickett Morgan
The wide expectation is that IBM will announce the next iteration of the IBM i operating system, up until now known as iNext and now thought to be IBM i 7.3, at its TechU conference in Atlanta, Georgia, on April 12. That is a week ahead of when COMMON is hosting a webinar with IBM i chief architect Steve Will to talk about new features in the platform, so the cat is out of the bag.
The word we hear on the street is that there will be significant enhancements to the database and to the security systems of the platform, neither of which come as a big surprise given the kind of transaction processing work that is still the main job for the IBM i platform. I also suspect that IBM i 7.3 has been tweaked to support the Power8+ processors that are coming out later this year–perhaps in early April at the OpenPower Summit, perhaps later in the fall concurrent with their use in the prototype supercomputing systems that IBM is building for the US Department of Energy mixing Power8+ chips and future “Pascal” Tesla GPU coprocessors from Nvidia. If I was a betting man, I would reckon that IBM will talk a bit about Power8+ and Power9, which will be used in the production Summit and Sierra supercomputers due at the end of 2017, at the OpenPower Summit and maybe lift the veil on Power8+ at the Hot Chips conference in late August. But, with Intel getting ready to launch “Broadwell” Xeon E5 processors for two-socket servers, IBM may let more be known about the Power8+ chips or may have even accelerated the launch.
We don’t know, and we don’t know anyone who does know and will tell, which is a bit unusual at this stage of the game. IBM is holding its cards a little closer to its vest these days, and that is because the stakes are higher.
With a new release of IBM i on the horizon, we are preparing to do a thorough analysis of the features and functions that Big Blue is adding to the platform, as always. But we also have a few other things that we are pondering.
The amazing thing to me is how much doesn’t change about the OS/400 and IBM i platform over its long lifetime, and you can stretch it back to the late 1970s with CPF on the System/38 if you want to get technical about it. In many parts of the IT industry, change is explosive and disruptive, and much too fast for most enterprises. They simply do not have the same skills and capital and operational budgets to absorb so much change. I would argue that the ability to have continuous change be the state of the business is what is separating profitable companies from unprofitable ones over the long term, and it certainly looks that way if you compare the hyperscalers like Google and Facebook to the IT suppliers like IBM, Hewlett-Packard Enterprise, and the soon-to-be Dell/EMC behemoth. This also holds true for their big customers in manufacturing, retail, banking, healthcare, and so forth to a large extent. (That is a big generalization.)
My point is that the profit pools are larger elsewhere, but that does not mean regular enterprises that don’t have an army of Stanford University PhDs can’t still do good and honorable work, and use systems like the IBM i platform to keep the books and interface with the outside world to serve their customers. And one of the key assets that they have–which works in their favor–is a platform they know well, that works well, and that pays for itself in lack of grief. We spend a lot of time talking about how new features needed to be added to the system and how the software has to make better use of the hardware, but it is useful sometimes to take a step back and review just how much doesn’t change in the IBM i platform because it doesn’t have to. It is already the kind of low-maintenance platform that the Googles and the Facebooks of the world have spent a decade or more building from scratch–albeit at a much larger scale.
The new stuff like PHP, MySQL, Ruby, and Node.js could have been added to the platform faster to be more current with the market at large, but IBM is not serving the market at large, and I think some days we all need to get a little more used to that idea. The days when the IBM i platform had a chance to take on Windows and Unix are over, and that is a hard thing to say as I sit here in my 27th year of writing this newsletter. I am not saying that IBM and its ISV partners are not getting competitive wins here and there, or that the platform is not providing value to customers that justifies what I think are prices that are too high in some cases. I am just calling it like I see it, and very likely how you see it, too.
Given this, I have a few observations. The first is a simple one. Take a look at this table:
Assuming that IBM i 7.3 comes out on April 12, it could ship a week later on April 19, and I think that is highly likely if there is not, in fact, any new Power8+ iron immediately available. There is no reason to wait, and it would keep IBM on a fairly consistent cadence of shipping the major releases in late April or early May in recent years. One other thing should be pretty obvious: It won’t be too long before IBM will withdraw IBM i 7.1, which is now six years old, from marketing. IBM likes to have two current releases, not three, over the long haul with only a little overlap. The operating systems, as you can see, are supported for a much longer timeframe, so it is not like IBM 7.1 shops have to do anything but be in Software Maintenance contracts, and if history is any guide, IBM i 7.1 will be supported through September 2017 with extended support lasting out to September 2021. Yeah, I just typed 2021, and it felt plenty weird, too.
Another observation for you. It has also been a long time since IBM raised prices on IBM i licenses and support costs. IBM tried to raise support prices back in 2013, and then it backed off after much weeping and gnashing of teeth. I still think that IBM will want to use a carrot and stick approach to get customers current on new iron and systems software, and I think a Software Maintenance price hike on old releases of the operating system plus hardware maintenance increases on older hardware are probably a matter of course. Call it a feeling in my bones, and I hope it doesn’t come to pass for your sake, but there it is. My hunch. If there is a recession looming and it hits, as some economists expect, then I think you can count on support price hikes for both hardware and software. The IBM i shops that rely on the platform are hooked, and IBM knows that. The pain of changing platforms is much higher than the pain of paying higher support costs.
If it were me running Power Systems, I would go in the other direction and absolutely drop Power Systems prices and IBM i prices and support costs for both for customers who are on older iron and who want to get current before a recession hits and all the budget dries up. Heck, such a mass upgrade might even help forestall a recession. . . .
I also think the imminent launch of a new operating system and a new hardware platform gives IBM the opportunity to radically change the pricing model for both hardware and software. Rather than cook up pricing especially for service providers building hosting businesses based on the IBM i platform (they are not clouds, not really, by any strict definition), maybe IBM should have utility pricing for the hardware (CPW, memory, disk, flash, and network I/O) and software (base operating system and database) from the get-go and let customers decide if they want to host their gear or have someone else do it.
Crazier ideas have been tried.