I like OpenStack (the Open Source IaaS Cloud Platform intiative), partly because of the model of open innovation and permissive licencing (which reminds me of my time at Eclipse) and partly because even within the existing governance model (which I have criticised) there is the opportunity for different agendas to surface and to drive the project in different directions and this diversity makes an analyst’s life more interesting.
One of the most intriguing names that has hitherto been at the periphery of the OpenStack initiative is Citrix. Up until last week, Citrix’s contribution was to ensure OpenStack ran on XenServer, something which I’m sure Citrix cares about, but perhaps wasn’t top of the list of requirements for the rest of the world. However, this week at it’s Synergy event, Citrix made some more sigificant announcements about Project Olympus, through which it aims to provide (in collaboration with Dell and Rackspace) a route to commercial exploitation of the OpenStack codebase.
For some time I have been perplexed as to what Citrix is doing. Are they genuinely intending to enter this space? Is this the real play or is it a spolier? Is this an opportunity to bind Xen or even XenServer into OpenStack, and then leave it attached to that project with an expectation it will quietly die in due course?
OpenStack always made sense from the Rackspace and Nasa perspective, less so from the software vendor side. The key actors are the service providers not their software suppliers. This is about Rackspace playing the “open” card to try and avert market dominance by Amazon Web Services (AWS). It may also be looking ultimately to gain market share by consolidating all the smaller service players it sucks into the initiative, all of whom (let’s be honest here) are motivated by the free software.
So where does it leave Citrix? You may remember they justified the massive cost of buying Xen and then giving it away because it devalued the VMware product set now known as vSphere. Are they really interested in giving away OpenStack to devalue VMware’s vCloud?- it’s hardly a sustainable business plan, and now the dust is settling on VDI I’m not sure VMware is actually the same threat to Citrix as we all once thought it would be. VMware and Citrix have effectively declared a truce in the sense that they no longer treat each other as mortal enemies. I am not sure that it is in Citrix’s interest to restart open warfare with VMware over yet another attempt to be relevant in server virtualization. Over half of the XenDesktop that goes in, goes in on a VMware back end – and this significantly benefits both companies.
Olympus is also an anti-Red Hat play, because Red Hat isn’t doing OpenStack (Canonical/Ubuntu noticeably is, at the expense of Eucalyptus), but again I don’t see Red Hat as a big threat to Citrix, and anyway they’ll come round to OpenStack in due course, wrapping it in DeltaCloud to ensure their own migration path (and save a little face).
Clearly Citrix can use Olympus to deliver an OpenStack service to compete with AWS (although one might suggest that Dell and Rackspace will make much more money out of the service delivery than Citrix ever will out of the software), and in addition Citrix can play the hybrid cloud-bursting private-to-public game, selling you the software to build the private cloud, then migrating you out. VMware’s obviously playing that game with vCloud, but it doesn’t have quite the same open-ness at the point you go Public. Citrix contemptuously refers to vCloud providers as “vClones”. Amazon has the Eucalyptus thing for Private cloud coming in the other direction, but that only goes so far – they aren’t actually the same codebase and there is always tension between where Eucalyptus wants to add features and where AWS is.
I guess in terms of the revenues that would result from selling OpenStack as software, open source software is a bit like selling bottled water (of the Dasani type). The product is widely available for free but the value is in cleaning it up and packaging it. OpenStack doesn’t have a natural consolidator/packager to provide the supported implementation, and from an enterprise perspective this is always where the money is. In PaaS, VMware is playing that game (and by the way doing a really good job) with SpringSource; Red Hat does it for Linux; Citrix can do it for OpenStack and do a good job because it can offer a self-contained solution (based on XenServer). It is also noticeable that the Open Source consolidator often moves to a monopoly or near-monopoly position from which it can benefit significantly, so all-in-all Citrix is well-placed to become the place the Enterprise buys OpenStack as software, however that isn’t yet a market that exists.
Then there is the effect on its other revenues. Citrix should be (are?) focused on creating a product set that offers an organisation a set of tools to do desktop/application virtualisation and in principle Citrix would be well placed to sell you Desktops in the cloud with Olympus under the covers for provisioning servers. Extension of the NetScaler line with Cloud Bridge also creates additional revenue opportunities. They created the CSP licensing model for XenApp/XenDesktop 2 years ago which may also play a factor in here. They are betting on the fact that customers will be moving as many hosted apps and desktops up to the cloud as they can, generating an incremental increase of connection licenses.
Arguably this project lets them explore an additional revenue stream that is complementary to what it is that they do – but are they big enough or focused enough to pull it off? The effort to manage and maintain the marketing for that service, and the technical support its going to require will be large. Indeed who on the Citrix board & at a senior tech level is driving this? So why develop a platform that has XenApp/XenDesktop in “the cloud”?
You could also say it helps with validating the hybrid cloud-bursting private-to-public model as well as picking up some SMB Citrix partner customers. But cloud-bursting can massively break best practice for designing a stable XenApp/XenDesktop environment with acceptable performance – because now your XenApp/XenDesktop environment is not at the same location as your back end services. Granted, you might have moved your back-end services to be completely cloud based: but that’s unlikely if we’re being honest for a whole host of legal/security/confidence/apathy/cost reasons. And in a web/cloud based app environment – do you need a virtualised desktop?
So, whilst I now know about Citrix’s intentions for OpenStack, I am still perplexed. Project Olympus may or may not succeed, but I don’t think this is strategic to Citrix.