Amazon (AMZN), Google (GOOG), and a host of startups have done a great job promoting the public cloud, just as VMware (VMW), Cisco (CSCO) and OpenStack have promoted private cloud. Yet public and private clouds are stepping stones to something even bigger: the hybrid cloud. Last week at the Gartner Data Center Conference 2012, the dominant theme was neither public cloud nor private cloud, but rather hybrid cloud.
The hybrid cloud is a form of cloud computing whereby applications and services can run across multiple clouds, colocation and data centers seamlessly, as a single hybrid cloud. The idea of a hybrid cloud has been discussed for years (see this article from InformationWeek in 2010 on the now defunct infrastructure 2.0 working group), the technical and business barriers have kept it, until recently, confined to slides and drawings and tactical tools.
See my recent high velocity cloud blog for more recent background.
This theme is very relevant to virtually all of today's technology companies, including Cisco, HP (HPQ), Oracle (ORCL) and others, because it represents the ultimate decoupling of an application, OS and service from a specific piece of hardware or even location. Hybrid cloud will do to the tether between apps and services and clouds what VMware did to the once powerful bond between apps, operating systems and individual servers.
Such a shift could shift the balance of market power to the likes of Microsoft (MSFT), VMware, Solarwinds (SWI) and other companies with software or service driven product models. Traditional relocation players who do not embrace hybrid cloud could lose out to those who do. Companies like RackSpace (RAX) and Equinix (EQIX) are well positioned as are carriers, like Verizon (VZ), who have made smart acquisitions.
Hybrid Cloud Will Ultimately Win
Because hybrid cloud would allow enterprises to seamlessly optimize their apps and services across clouds and facilities without modification, it would allow IT teams to be even more nimble, efficient and strategic to business units. Until recently, hybrid clouds have been very difficult to build.
Simply moving a multi-tier app and the required services (to allow that app to communicate with databases and other apps resident in the data center) can take months and isn't a risk-free process. Most cloud deployment tools have been limited to software or services that simply allow an app or server to go from one cloud to another, with minimal service integrity or integration with the data center, mission-critical databases, etc.
Other tools have been more network-centric, simply connecting clouds to data centers with minimal app and connectivity intelligence. Public cloud deployment, therefore, can actually increase the IT costs for an enterprise, versus the vendor hype, at least for existing apps that weren't created specifically for the cloud.
For broad enterprise adoption, it will take holistic hybrid cloud solutions developed from expertise in networking, apps, security and storage to cross the hybrid cloud process barrier, and most companies are either architected around a single approach or are architected upon layers of software stacked on (you guessed it) specialized hardware.
Migrating from the private cloud to the hybrid cloud could be as challenging as tasking Intel (INTL) with shifting from PC to wireless-centric semiconductors. Ultimately, a new architecture is often required to make such a leap. The very code and design legacy that is a barrier to entry (for competitors) in the PC chip market could a barrier to entry (for the likes of Intel) in another.
VMware, on the other hand, has been developing more robust capabilities, and has been strategically spinning off and organizing units that could deliver on the hybrid cloud model. Thus far, however, one must standardize on VMware to get to a lesser form of hybrid cloud (requiring VMware virtualized to VMware vCloud, for example), which is antithetical to the notion of hybrid cloud's freedom of cloud. Recent announcements, including joining OpenStack with EMC, signals that VMware is thinking bigger than Amazon and clearly moving toward hybrid cloud.
Microsoft has been a relative latecomer to the cloud market, but is in a promising position to drive the hybrid cloud because of its legacy expertise in applications and operating systems and its potential to leapfrog the public and private battle with a very compelling hybrid cloud offering. Experts think that Microsoft could enter in a big way in 2013, extending upon its cloud storage momentum as a result of its StorSimple acquisition.
Hybrid cloud could also have a revolutionary impact on a host of hardware-bound solution markets, especially disaster recovery. A new hybrid cloud solution (cloud failover for example), could allow enterprises of all sizes to leverage hybrid cloud operating models for substantial costs savings, agility and even disaster avoidance. Think about the potential for an IT team to move an app from one region to another as a storm approaches.
There are no doubts a new generation of hybrid cloud solutions that will come to market in coming years, including cloud migration and cloning, cloud failover and ultimately cloud bursting. These new solutions will represent fundamental shifts in how IT solutions will be delivered and shift the market momentum from the hardware-bound to the software and services-bound. We will no doubt see new generations of lock-in (like a vendor requiring virtualization for hybrid cloud) and yet newer generations of solutions that break the locks. Hybrid is a whole new cloud.
Disclosure: I am long MSFT.