Last week, there were several articles linking Sony’s security failure (hackers stole 77 million custome credit card record; Sony had to shut the Playstation’s network down for a week; data that should have been encrypted was stored in plain text) and Amazon’s major outage (parts of Amazon’s AWS – their public cloud – were offline for two days; parts of AWS supposedly isolated from one another nonetheless were engulfed in a cascading failure).
How do these – and other – failures create opportunities – especially in enterprise IT?
There is a common assumption, which I broadly share, that many enterprise IT activities will migrate into the cloud. Cloud providers can build at a scale, and with a level of expertise, that is beyond the capabilities of even the largest enterprise, achieving superior cost efficiency, capabilities, and reliability. That’s the theory.
How is it working out?
So far, a handful of services are delivered from the cloud – salesforce.com is the most prominent. This may not seem like “real” public cloud, in that it is a service, not a general purpose utility cloud; but it might point the way, with cloud APIs, web-hooks and other forms of extensibility enabling packaged services to become as flexible as traditionally-delivered IT.
Google has made some inroads with Google Apps and Google-Mail-for-business, but largely in small tech-oriented businesses. Google Apps is a long way from being mainstream. Larger enterprises find the limited functionality of the applications troublesome, and are not convinced by service-reliability guarantees nor data security. The online and realtime collaboration features are nice, but have in any case been imitated to a significant extent by Microsoft’s incumbent products. Switching costs can be quite high. For most mid-sized or large enterprises, the incentive to move from Microsoft’s Office suite to Google Apps in the cloud is not strong enough.
Amazon’s AWS, together with Rackspace’s competing service, is getting meaningful traction as a public cloud computing platform. The traction, though, is mainly coming from tech companies that are using it to deliver their own packaged cloud services – often, though not always, consumer services. Zynga (online games), Netflix (online video), Engine Yard (Ruby on Rails automation), Foursquare (location-based social service), Hootsuite (social media monitoring and management), Heroku (Ruby cloud), Quora (Q&A site), and Reddit (news links) all use AWS. Some startups use AWS for “internal” IT, and a few large companies are trying it for certain projects. Yet, the mid-sized and large enterprise is still a long way from embracing public cloud as the means of delivering IT.
Such is the appealing-in-theory-but-resisted-in-practice nature of a move to cloud for mid-sized and large enterprise IT, many vendors are trying to reposition their on-customer-premises products as “private cloud.” There may be more to it than positioning in some cases – by bringing standardized architectures, deployment and management inside the customer, some of the benefits of public cloud (efficiency, predictability) may be transferred also – but in general “private cloud” is no more nor less than the next evolution of the corporate data center. As such, it may offer many opportunities for startups; indeed, in many ways it is probably easier to build a business by being evolutionary, rather than pressing for a public cloud revolution.
Yet, in the long term, shouldn’t the benefits of really-public-cloud overwhelm all the resistance? Is that where a mega-business could be built?
For public-cloud to take over, a whole new infrastructure needs to be created – problems that are solved in private data centers today have to be solved differently for public cloud. Here are some key ones:
- Management and monitoring
- Flexibility
- Security / privacy
- Data separability
- Service separability
- Redundancy, backup, reliability
My Perfect Startup 