In the battle of the public vs private clouds, the lines have been drawn between those that offer infrastructure for use in existing data centres and those that offer platforms and services in their own data centres. The discussions, while technically based, are philosophical, passionate, opinionated and about the market laggards trying to reel in the leaders. Google and Amazon got there first(ish) – building services based on their own existing infrastructure and now other large IT organizations are trying to play catch up. Big players like HP and IBM who have not seen the gathering cloud are now scrambling to find products in their arsenal to put a cloud badge on and in most cases, the easiest products to re-badge are those related to networking, storage and virtualization. With no existing big data centres to call their own the targets for their newly minted offerings are the large data centres of their existing clients. With solutions that can only sold to on-premise customers it is naturally in their best interests to bring out their biggest marketing cannons to fire at the public cloud.
The pro-private arguments, although maybe a bit too loud, are valid and represent real concerns expressed by enterprise customers. There are concerns about security, vendor lock-in and and questions such as “What do we do with our legacy applications that can’t run in the cloud?” or “Why should I rewrite my applications for a platform that can only ever run in XYZ’s data centre?”. Public cloud vendors are keeping quiet on the vendor lock-in question, just in case nobody notices, and they don’t really have answers to how you run your obese ERP system on their cloud and, in most cases, probably don’t care. Google is not going after the enterprise market with their cloud offerings (yet) and is far more supportive of, and appealing to, the hip and cool start-ups that reflect their culture than boring suits asking about security of financial data in Oracle. The lack of response from the public cloud providers adds fuel to the fire of pro-privates and the frenzy is resulting in the prediction of the collapse of the public cloud model, with the argument that enterprises will simply not accept the limitations and risks of the public cloud. Perhaps the market will turn to favour enterprise spend, but also perhaps Google, Amazon, Rackspace and others are playing a longer (and lucrative) game – a game where the applications of the future are not monolithic, obese and only the domain of large enterprises which, at least recently, have demonstrated that ‘too big to fail’ is no longer the case.
Thrown in the middle of this is Microsoft which has, with varying degrees of success, straddled enterprise, SME, home and mobile markets for ages. I don’t know how much of Microsoft’s revenue comes from large enterprises, but I bet it’s a lot – a lot of Windows, Office, Windows Server, Exchange, SQL Server and a whole host of other products run in just about every enterprise. So Microsoft has some enterprise penetration and credentials that allow them to compete with the biggest and meanest of the enterprise vendors and they probably are the biggest software only supplier to the enterprise. Similarly, in the cloud they also straddle the enterprise and start-up markets with the most obvious and widely marketed being their cloud PaaS offering – Windows Azure, but they also keep selling windows licences. So while the pro-privates bang the private drum, Microsoft still makes money, it doesn’t matter to Microsoft if you run your Exchange server on physical hardware or your virtualized private cloud.
Microsoft is not currently marketing Azure at the enterprise, rather going for the hobbyist and start-up with promises of lower initial costs and elasticity as the business grows – a market that doesn’t ask too many difficult questions. Microsoft’s plan for Azure, I am assuming, is fairly long-term and I doubt they expect to get a return on their investment within the first year or two – it is a hedge on a new wave of computing pushed by Web 2.0, ubiquitous broadband, demanding customers and dynamic businesses; not your Mom or Dad’s enterprise software. When announcing Azure last year, one of the most common questions in the community was “So can we run Azure on our own servers?” with the answer, at least at this stage, being an clearly enunciated “No”. The reasons for this response are unknown, I reckon that at least for now, the public cloud is where they should leave Azure – giving it time to mature, be tested, be tooled and to pay for all of those data centres. Perhaps within a few years the private cloud discussion would have faded away anyway, making the idea of private Azure fairly mute and uninteresting.
But if… if Microsoft sees that the market is turning against the private cloud and if their Excel spreadsheet shows that there is more money to be made out of private Azure than public Azure then all that they need to do is package their Azure fabric as a product. Azure runs on Windows on commodity hardware and could (probably) easily be worked into a package that can be marketed to enterprise customers. I am sure that there is some code that will need to be reworked and some wizards added, but Microsoft is reasonably good at writing software, so that shouldn’t be a problem. What Microsoft does have is a distribution channel and a piece of PaaS software that runs on Windows. Google would never be able to (or want to) extricate their PaaS from their commodity hardware, networks and secret sauce and Amazon doesn’t have a distribution channel for enterprise software – leaving Microsoft as a viable cloud provider that could have both a private and public cloud offering. So as competitors, yet again, predict the collapse of Microsoft – this time based on the private vs public cloud – I suggest that they don’t poke the sleeping giant too much.
The ace that Microsoft has up it’s sleeve is to publish ‘Microsoft Windows 20xx Azure Edition’ as a SKU on their price lists for their well established channel to take to market. We just don’t know what the ‘xx’ is.
Simon Munro
@simonmunro
Disclaimer: This is completely my own opinion and does not reflect the opinions or views of my employer or Microsoft. I have no inside information, knowledge of things covered by NDA or any MVPness. I make this up based on experience and publicly accessible information.