Is there a viable opportunity for Red Hat to position themselves against the very OpenStack competitors with whom they’re now nominally working? Red Hat still bundles software with solutions from companies like HP, but continues to market its own competing implementations of OpenStack. Meanwhile, Mirantis is making a lot of noise with words, and even more noise with their results. This leaves Red Hat OpenStack in a place made precarious by a distribution problem that has nothing to do with their technology and everything to do with their channels. Red Hat’s website is reflecting this insecurity: as of January 2015, Red Hat has published a competitive report that is openly hostile to Helion, HP’s OpenStack play. And a simple Google search will cover your screen with the bad blood between Mirantis and Red Hat.
On the surface, it seems that Red Hat would have the dominant position when it comes to OpenStack adoption. They have the strongest presence in the Linux market already, making them a natural choice for most enterprises. IDC data suggests that Red Hat has a 64% share of the paid Linux space. Moves were made in May of 2014 to lock users of Red Hat Enterprise Linux (RHEL) into Red Hat’s distribution of OpenStack, to which HP countered with Helion. This undercut Red Hat’s seeming attempts to close off 64% of the total addressable Linux market from other OpenStack vendors. One would assume that most end users who have already deployed RHEL would naturally look to Red Hat as a source for OpenStack deployments, and that enterprise customers already using RHEL would gravitate toward using Red Hat OpenStack.
Somewhat surprisingly, this isn’t always the case, as witnessed by the rise of Mirantis. Mirantis’ 2014 revenue is more than double what they were in 2013, and the company is booking new business at a rate of $1 million dollars every week. These numbers are perhaps more important than the $100 million they raised late last year. It seems as if Red Hat may not be fully capitalizing on what appears to be a very advantageous market position. Meanwhile, what is a nominal partnership between Red Hat and Mirantis has become increasingly hostile. Mirantis co-founder and CMO Boris Renski recently went so far as to predict that Debian Linux should and will be adopted as the official Linux of OpenStack, and also rubbed sand in Red Hat’s face about HP, who are standardizing Helion on Debian.
DOES RED HAT NEED HARDWARE LOCK-IN TO COMPETE IN OPENSTACK?
Perhaps making it difficult for Red Hat is the fact that OpenStack is now largely being deployed on newly acquired systems—making it fairly likely that an end user will be acquiring hardware from HP or another major vendor. To resist this, Red Hat established a large ecosystem of certified OpenStack partners, including Cisco, Dell, IBM, and Intel. Notice that HP is not among them. Of course the salesperson from HP will be pushing HP’s OpenStack as opposed to Red Hat’s, as they are incentivized to do that. This is the problem Red Hat is facing: when it comes to new system sales it is tougher for them to compete with companies selling both hardware and OpenStack.
There may be an answer to this if Red Hat is willing to continue a strategy that intensely encourages those who have RHEL to deploy OpenStack on RHEL. What else could address this issue? It could benefit Red Hat to work exclusively with hardware vendors that do not have their own distributions. This would be a strategy with great risk, potentially alienating many customers. However, with the push for adoption of Debian as the official Linux of OpenStack, not taking such a risk could leave Red Hat in decline with respect to VMware and others, as was pointed out by Mirantis’ Renski in another recent article. Lenovo is already a partner with Red Hat, and would be worth considering as an exclusive ally. Locking down these partnerships into such exclusive alliances, as warlike as that seems, could substantially improve Red Hat’s standing in the OpenStack world.
The bottom line is that Mirantis is raising hell all over the web, and that HP is attempting to sell turn-key engineered systems, saying basically: “We’ve got the entire stack packaged and built together. OpenStack is just one piece of it. Why would you want to single out OpenStack and try to deploy it? Why wouldn’t you buy our integrated platform?” Thus, as much as Red Hat wants to stand out, they can only be successful in HP hardware acquisitions when customers select the Red Hat distribution of OpenStack, rejecting the turn-key HP option.
Red Hat is facing challenges on several sides: by lacking their own hardware line, from pushes to adopt competitive distributions of Linux as official to OpenStack, and from direct competition with a successful, vocal and hostile company in Mirantis. The issue for Red Hat now becomes staying as competitive as possible. One way to do that against HP and other hardware vendors who have their own flavor of OpenStack is to take lock-in to another level. Anyone who has a combined hardware-software solution that is offering up an OpenStack implementation is going to be a problem for Red Hat, even if they are customers of Red Hat. With so many problems, extreme measures might be needed to save the day. Determining whether hardware and OS lock-in could bring success is research that might determine if it is time for Red Hat to stop playing defense and start playing offense