The Impact On IBM i Of Big Blue’s Acquisition Of Red Hat
October 31, 2018 Timothy Prickett Morgan
Well, we can honestly say that we did not see that coming when IBM and Red Hat announced late last Sunday afternoon that Big Blue would be shelling out $34 billion to acquire the world’s most successful business that peddles support for open source infrastructure software.
Ironically, at the time I happened to be writing about how IBM and Red Hat had just announced that they had brought the OpenShift Container Platform, a mashup of Docker and Kubernetes, to Power Systems machines running Linux, and I was lamenting that it was not trivial to figure out how to integrate this software with the IBM i platform. I had some ideas, as I often do. IBM had a more direct one, and that is to buy Red Hat outright for such a high premium – 62.8 percent above the $20.53 billion market value that Wall Street had valued Red Hat at on the Friday before the news broke – that it probably will not see competitive bidders try to swoop in and steal its new hat.
This is a huge amount of money for IBM to blow on an acquisition – nearly 10 times what it paid for Lotus Development Corp in 1995 to get the Domino platform, 46 times what it paid for Tivoli System’s Netfinity system management tools, and seven times that it paid for Cognos for its data analytics tools back in 2008. Pricewaterhouse Coopers Consulting, the first big deal that Ginni Rometty, IBM’s current president, chief executive officer, and chairman, took down in 2002, cost $3.5 billion and significantly added to what IBM Global Services could do. Rational Software in 2003 cost IBM another $2.1 billion. If you add all of the big deals that IBM has done since 2000 and probably the small ones, too, it probably does not exceed $34 billion. So if IBM is blowing its cash and borrowing against its credit lines and stopping its share buybacks in 2020 and 2021 to pay for Red Hat, you have to figure it is pretty serious.
Deadly serious, in fact. Rometty needs to do something that is coherent to the IT sector at large and that gives her a lasting legacy on pointing IBM toward a future that give sit a chance to be a bigger player in IT, not one that is gradually enshrinkening itself by divestitures and progressive revenue declines. The acquisition also gives IBM a relatively complete infrastructure platform that can run on its own Power Systems and System z servers as well as the X86 servers that are sold by its many competitors.
Red Hat’s Enterprise Linux is familiar to more than a few IBM i shops, and is by far the dominant Linux distribution in the enterprise. It is probably the most popular Linux in the world, if you include the millions of customers using RHEL as well as its CentOS clone, which the company acquired a few years back. Our guess is that there are around 2 million machines in the world running RHEL and maybe another 2 million running CentOS. If you count up all of the homegrown Linux instances in use by seven of the Super Eight hyperscalers and cloud builders – Google, Amazon, not Microsoft very much, and Facebook in the United States and Alibaba, Baidu, Tencent, and China Mobile in China – there is probably easily another 10 million to 12 million Linux servers in that mix, and adding all of the supercomputing and technical computing that runs on Linux other than RHEL and CentOS plus the portion of corporate datacenters that is also not controlled by Red Hat probably give you something around 1 million machines. That gives Linux about a third of the installed base of 45 million servers in the world, with the remainder mostly running Windows Server. If you add up all the servers in the world running IBM i, AIX, and z/OS plus any other operating system you can think of (there aren’t many alternatives left), that is probably somewhere around 1 million machines. Tops.
There is a reason why Microsoft is still rich. And this is it.
Now, IBM has put itself back into contention with Microsoft, and that is probably a good thing for its long-term financial health. While we are enthusiastic about IBM’s efforts to sell more Linux on Power Systems for modern workloads – all the new and interesting simulation, modeling, machine learning, and analytics software is always developed on Linux platforms, and so are container platforms for microservices style applications – there is no way that IBM could ever have built up the kind of business Red Hat has done by pushing the Linuxes of others on its iron. I did a detailed analysis of the financials of the deal over at The Next Platform, so if you want to see that, just hit the link. What you need to ponder here in IBM i Land is what the acquisition means for Power Systems in general and IBM i in particular, and as I said last week, how IBM will do a better job integrating Red Hat’s stacks with IBM – without disrupting things.
Despite what Rometty and Jim Whitehurst, Red Hat’s chief executive officer, said during their conference call on Wall Street going over the deal, where they both proclaimed there is no overlap in products from the two companies, this is strictly not true and even if there is not overlap, there certainly is competition. If even IBM doesn’t knock out its own products with those from Red Hat, it might mean that development on the native IBM products slows or even stalls out.
Let’s take a few cases right here and now so you see my point:
- Red Hat has its own OpenStack distribution, called OpenStack Platform; and IBM has its own called PowerVC.
- Red Hat has a Docker container orchestration platform called OpenShift; and IBM has IBM Cloud Private, as we discussed last week.
- Red Hat has a multicloud management stack that works at the virtual machine level and can be extended to containers with OpenShift on top of it called Cloud Forms; IBM has its new IBM Multicloud Manager for containers and has a slew of virtualization management tools that it has sold over the years.
- IBM sells AIX and IBM i on Power Systems and z/OS and other operating systems on its System z mainframes; Linux accounts for the majority of capacity sold on mainframes these days and has for years, and Linux is growing on Power Systems as a percentage of capacity sold. This is indirect competition, mind you, but there you have it.
- Red Hat sells JBoss Java application middleware, and it makes lots of dough; IBM has WebSphere Java application middleware and it makes lots of dough.
- Red Hat has the Gluster parallel file system and the Ceph object and block storage system; IBM has Spectrum Scale (formerly the General Parallel File System or GPFS) and Cloud Object Storage out on the IBM Cloud.
- Red Hat does not have much in the way of relational database management systems; IBM has Db2 in its various guises. Both peddle open source databases and datastores from the open source community.
- IBM sells SUSE Linux Enterprise Server and Canonical Ubuntu Server on Power Systems and System z mainframes.
- Red Hat sells JBoss Developer Studio; IBM sells Rational Developer.
I don’t claim that this is an exhaustive list, but it hits the big conflicts and overlaps and it most certainly is a non-zero number.
IBM has committed to leaving Red Hat more or less alone, in a “distinct” division where Whitehurst reports directly to Rometty, and parts of Red Hat will be reported in its Hybrid Cloud division and other parts in Global Services, apparently. So there goes any visibility into what Red Hat is actually doing once this deal closes. But the real issue is not so much that IBM will be tempted to move companies from its own versions of infrastructure software to the Red Hat versions, or even wholesale off IBM i and AIX to Linux on Power with Db2 behind it. We don’t think IBM will do that, at least not for a few years after Red Hat’s business has been integrated and maybe big chunks of its infrastructure software, too. Our concern is more that IBM will no longer fund development in its own wares, and that AIX and IBM i shops will see software tools that they use whither from neglect.
This is not anything any of us need to worry about for now, especially with the deal probably not closing for the better part of a year. But it is important to keep an eye on what IBM does with Red Hat and what it doesn’t do with its own tools. And that is what we will do. It is why you keep us around, after all.