The cloud got a fillip with Netflix making a public announcement that it will move all of its web applications to Amazon.com. This came as a surprise, given that Netflix and Amazon could be potential competitors as they stream movies to people’s homes.
Given its critical dependence on internet infrastructure, it takes courage, foresight and risk for Netflix to abandon plans to build one’s data center and run that off the cloud operated that too by a potential competitor. It implies that moving to the cloud is cost effective if one doesn’t have a data center already and do not want to manage the data center. So, the case of moving to the public cloud is strong in such cases
While Rackspace, GoGrid and other vendors are emerging as competitors, Amazon Web services is far ahead of the game. Since 2007, Amazon’s S3, the Simple Storage Service, and EC2, the Elastic Compute Cloud, Amazon has offered the general-purpose cloud computing platform with a business model that made it attractive to developers large and small.
Recently Google showcased their Cloud offering in their much publicized Atmosphere event. Google reported that there are about 30 -50K signups per day for the Google applications which essentially reside on the cloud. In some analyst reports, I have seen this as counted as part of SaaS revenue stream and in some cases this gets categorized as Cloud revenues. All said, Google never sold the cloud, but the apps. It so happens that the apps are in the cloud.
Microsoft seems to be taking a leaf from the Salesforce.com’s playbook by building new applications on its Azure platform. Salesforce.com is the only enterprise apps player to move beyond software offered in a SaaS model and create a complete platform – Force.com which has over 100,000 applications.Though it is early days, Microsoft also seeks to do to the cloud what it did to desktop software for over 2 decades – by making applications run on Azure alone and harnessing the entire Microsoft developer community to build applications for the Azure platform. Surprisingly Microsoft’s first application on the Azure platform is an online collaboration suite and advertising tools for political campaign management.
So, the public cloud is beginning to gain some traction. It could very well be that it is the applications that is offered which could determine the future of the infrastructure (cloud) platform, for all players other than Amazon which has restricted itself to just offering development tools in addition to computing and storage.
As one can see, Google’s enterprise apps is a case in point and Microsoft’s strategy reinforces the point.
If applications on the cloud becomes the raison d’etre for cloud adoption, the differentiation amongst cloud providers could very well be the data subsystems that they employ since it is not easy to deliver data at the speeds required by the applications. Storage and computation thus become starting points, not end offerings.
If the public cloud is now gaining traction, the private cloud lags behind despite all the vendor hype. The private cloud for most part is just repackaging data center consolidation, server virtualization, IP storage technologies to offer a more scalable infrastructure.
The private cloud still needs lots of CAPEX and infrastructure still remains a fixed cost, unlike the public cloud. It is still by and large inelastic. Moreover the private cloud will still be run by enterprises. Will large companies which have made significant investments in their own infrastructure move to a vendor cloud? VMWare after earning its spurs in the virtualization market is now trying to make it easy for applications to migrate between cloud providers, and creating an easy interface between private and public clouds. VMWare’s acquisition of Zimbra and Springsource is indicative of its intent to build a developer’s platform much like Azure or Force platform.
Despite all these developments many CIOs are wary of putting their entire software and business operations on a vendor’s data center. Looking beyond the hype, CIOs are likely to test the waters as privacy and compliance considerations weigh heavily on the CIO’s mind, not to speak about interoperability and performance issues. There is also a switching cost involved and if one runs an efficient virtualized infrastructure, there is no compelling reason to move to a vendor cloud. So, next time you hear private cloud, it is just that vendors are selling the same thing with a new approach. A case of selling old wine with a new label. We still have a long way to go.