At the Open Compute Summit V today (January 28, 2014) in San Jose, Microsoft rocked the cloud and hyperscale world and joined the Open Compute Project (OCP). Microsoft has contributed complete server and storage designs, as well as open sourcing the software code it created for the management of hardware operations, such as server diagnostics, power supply and fan control. Microsoft committed to help build an open source software community within OCP as well.
Suddenly OCP has multiple additional hardware specifications covering the general computing worlds of search (Bing), communication (Office 365) and general purpose computing (Windows Azure). Even more important, OCP has an API that will connect the hardware layer management to the software-led infrastructure layer.
The leading current software-led infrastructure contender is OpenStack, and the combination of OPC and OpenStack is a pre-race favorite for service providers. The OCP V conference included a demonstration of provisioning server configurations in real time from an iPhone. This was demonstrated by a Phoenix based company called IO, and their infrastructure service is built on OCP and OpenStack.
This move catapults OCP from a contender to favorite to become the de facto standard for the hyperscale open source hardware market, particular for cloud providers and data center service providers.
So what is Microsoft’s reason for embracing OCP? Amazon and Google are Microsoft’s biggest competitors for cloud computing and the biggest threat to their application and software infrastructure revenues. Microsoft believes that creating open standards will help accelerate the growth of mobile cloud, public cloud and enterprise private cloud. By embracing OCP and including hardware in a total solution, Microsoft have enabled multiple channels for selling and supporting Application and middleware "Single Managed Entities" (SME), both for their own cloud and for other cloud service providers. Microsoft can use this to increase volume, and gain from analytic driven automation of provisioning, operations and break-fix across software and hardware by treating the whole stack as an SME (see Note 1 in footnotes).
Figure 1 shows some of the different ways that systems can be integrated. The hardware layer is where OCP would fit, and interface with the layers above with the new API. Microsoft is in a position to create a Single Managed Entity all the from hardware through to the application, and deliver this in its own cloud, and to cloud service providers.
Bottom line: OCP drives economies of scale.
Action Item: The growth of hyperscale computing will allow globally connected mega data centers to offer traditional outsourcing, private clouds and public clouds for a combination of enterprise IT, cloud service providers and data brokers. The topology of all data being close together, and this data being close to cloud data from data brokers, cloud service providers, mobile, social and the Internet of things will allow the potential to extract the greatest business value from all data sources. In addition it is likely to be the lowest computing cost. This model of computing will provide the greatest agility and lowest cost for many enterprise customers. The combining of OCP and OpenStack ensures that there are alternatives to Amazon and Google.
Bottom Line: Enterprise & Government CIOs should cut investment to zero for the expansion of current enterprise-owned data center space, and develop a data-led topology based on services located in global mega data centers.
Footnotes: note 1
This, of course, assumes that Microsoft will have a CEO and board that understands and fully supports the enterprise cloud, and can stand firm against the windmill charging legacy of Balmer wanting the throw away treasure and focus on mobile devices.