Looking Ahead To The IBM i New Year
January 8, 2018 Timothy Prickett Morgan
There is always plenty to worry about in the world; that is what a world is for, after all. Human beings are both blessed and cursed by their frontal lobes, and all of us are engaged in some kind of prediction and prognostication in our daily lives, from little things like what someone is going to say next to big things like what we want to do with the rest of our lives. Running scenarios and simulations is how we make our way in the world, and it is both fun and useful so long as you don’t take it all too seriously.
At this time every year, when most of the days are unminted and the few that have come into being are shiny with the date imprint, most of us start thinking about the future and some of us make resolutions to try to shape our own behavior for the better. The best of these stick and mold our lives.
With this inaugural 2018 edition of The Four Hundred, we make the same resolution that we do every year: to provide the best insight into the goings on in the IBM i marketplace and to be one of the several focal points for the IBM i community to exchange ideas. So that is the easy part.
The prognostications are a bit tougher, of course. But we have a few that we want to make, and I will be doing mine here and others will have their say as they see fit. I would love to hear your IBM i predictions for 2018 as well, so send me an email as feedback if you want to share. Anyway, without further ado, here are my auguries for the coming year as it relates to the IBM i platform.
Major shifts in computing. This is an easy one, and it is a trend that we have been predicting for many years. The difference now is that the shift in computing is really going to happen this year. First and foremost for the IBM i faithful, the first Power9 processor, the “Nimbus” variant with 24 cores that is aimed at machines with one or two sockets, started shipping in HPC-style Linux machines back at the end of 2017, and will eventually make its way into systems that can run IBM i and AIX as well as the big endian versions of Linux on top of the PowerVM hypervisor. Later in the year, perhaps by the COMMON conference and expo, Big Blue will deliver the 12 core “Cumulus” variant of Power9 aimed at NUMA systems with four, eight, twelve, or sixteen sockets. It is hard to gauge what the enthusiasm level for Power9 chips will be for customers running IBM i or AIX workloads, but customers with Power6, Power6+, Power7, and maybe Power7+ iron will be looking to replace the machines that are looking a little long in the tooth.
But Power9 is not the only transition. AMD has credible Epyc X86 server processors and is building customer momentum, and the Spectre and Meltdown bugs affecting Intel Xeon and possibly other processors is going to make everyone re-evaluate their processor choices a little. As we go to press, we are not certain if these two bugs, which became apparent last week but which have been known for some time, affect the Power8, Power8+, or Power9 processors and we will be doing a follow-up once we find out more. Qualcomm and Cavium also have good ARM server chips in the field and are also getting some traction with their respective Centriq 2400 and ThunderX2 chips among the cloud builders and hyperscalers. It remains to be seen what enterprise customers will do, but with most enterprises having lots of X86 iron, it seems a no-brainer to look at Epycs as an alternative to Xeons, and for those who run Linux workloads, ARM is now an alternative with Red Hat officially supporting ARMv8 architecture chips with its Enterprise Linux distribution.
New IBM CEO. Ginni Rometty has been president and chief executive officer of IBM since October 2011, and she turns 60 at the end of July this year. That is the age at which IBM’s top leader has traditionally retired from the job, and there is no reason to believe that Rometty, who has served as CEO during one of IBM’s toughest times, will be any different and hold onto the job longer than her predecessors. It is very likely that IBM will start signaling who is in the running for the CEO position soon, possibly within a few weeks when it announces its fourth quarter financial results for 2017, maybe during the summer. We have our own thoughts about who might take the helm, and how long Rometty might stay on board as chairman of the board once she relinquishes the president and CEO positions.
We will talk about this issue and the possible affects on Big Blue in a separate story, but suffice it to say that with 22 straight quarters of revenue decline, this is a year when revenues should start to rise again, thanks in large part to Power9 and System z14 mainframe rebounds, and that is a good time to do a transition. We are eager to know who is going to be in charge and how a new IBM team might further transform the company, and we are ever hopeful that whoever does get the job and whoever is elevated to positions of power in the divisions of the company that IBM remembers that there is still plenty of margin in hardware and that it would be fun to see Big Blue be a force in systems again. Buying GPU chip maker Nvidia, FPGA maker Xilinx, and networking chip maker Mellanox Technologies, for instance, presents some very interesting possibilities and would make the OpenPower partnership a product line.
Continued consolidation in the IBM i market. So long as cash is cheap, IBM i software vendors have captive markets that bring in healthy profits, and company founders want to cash out and retire, then consolidation is going to continue apace in the IBM i arena. It all comes down to market definition, of course. Mature markets tend to like two or maybe three dominant players within a market or submarket, and then sometimes, if the conditions are right, the dominant players start building a broader conglomerate across product lines and submarkets as a way to leverage what they have to get more of what they want. HelpSystems, Fresche Solutions, Syncsort, and Rocket Software are very quickly building conglomerates that have some or a lot of products aimed at the IBM i market, and this is possible because of the maturity and profitability of their IBM i products. While much can be said about the size of the Windows Server or Linux markets, in terms of the number of customers or licenses or the revenue streams, we suspect that in terms of share of revenue that drops to the bottom line, the IBM i market is holding its ground. There is no other reason why so much private equity would be funding the acquisitions and mergers we have seen in the past decade and will see even more in the coming year and beyond.
While conglomerates are good for the vendors and probably good for investors and customers (sometimes the mergers or acquisitions don’t pan out as expected for anyone), there is still thankfully room in the IBM i market for smaller players with innovative products, and we think given the profit profile and the current stability of the IBM i base that we will see some new companies come into being and do new things as well as old things better.
Investment in the IBM i platform. It has been a lean couple of years in the IBM i product line, with a bunch of layoffs and changes in how products are developed and sold, but Big Blue has at the very least continued to invest in the Power architecture and the IBM i software stack. With the Power9 rebound this year, there is an outside chance that IBM will spend some money to try to convert more of the legacy System i and Power Systems base to more modern machines. But we would not hold our breaths on that one. It seems just as likely that IBM will use any profits to fund development of Power10 processors, which it is committed to bring to market in 2020 or so. That said, we do expect for IBM to continue to support its ISV partners through co-marketing programs, as it has done with varying budgets for as long as we can remember. We don’t expect for big changes here – more steady freddy.