I’ll state the obvious: The last five years have seen incredible changes in IT, including the rise of new cloud operating models and the spread of virtualization into production data centers. A new generation of upstarts has captured the imagination of the IT industry by unleashing new software and service-centric solutions, including private and public clouds, new devices and operating systems, all leading to a larger and grander software-defined IT era.
Now the not so obvious: Software-defined IT is only a stepping stone to a much larger, service-defined IT era, enabled by the rise of hybrid cloud automation.
We are not there yet when it comes to hybrid cloud, despite the plethora of hybrid cloud announcements based on, of all things, virtualization lock-in. Why? The notion of migrating a virtual app and its critical services from a rack of x86 servers only to be stranded within yet another set of racks in a public cloud is not a hybrid cloud operating model. It is simply an early form of cloud migration.
Cloud migration will ultimately evolve into hybrid cloud integration within a true hybrid cloud operating environment. It isn’t the ability to move a virtual app once from one environment to another that is hybrid cloud, but rather the ability to move physical and virtual apps and services on demand, as needed between environments. Hybrid cloud is all about control for both the enterprise and the service provider.
Migrating VMs from cloud to cloud is a tactical payoff; integrating existing and virtual apps with clouds is a strategic game changer: control with agility.
First Gen Cloud Drawbacks
Virtually all of the cloud operating models in use today involve significant tradeoffs, often with control and security or unplanned downtime risks.
Most clouds are not cheaper for existing apps with predictable workloads, except in the case of small and medium-sized businesses with even smaller IT shops. The first generation cloud payoff is greatest with new apps created for the cloud or for existing apps with unpredictable workloads or sporadic seasonal or event-generated usage spikes; or when apps can share a larger and more energy-efficient server with other occasional use apps.
Eventually hybrid clouds will be the most optimum environment for many more apps, even those with predictable workloads. Hybrid clouds are, in essence, physical and virtual data centers apps and services integrated with clouds. They will be infinitely more agile and economical than today’s islands of public and private cloud capability surrounded by rough seas of platform lock-in, manual processes and risks.
Yet today, because of the recently discussed hybrid cloud barrier, virtualization is spreading aggressively in the production data center, driven in part by the promise of private cloud computing and the ability to migrate workloads across larger theaters of operation. That is a significant payoff, especially considering recent acquisitions and added mobility capabilities. But virtualization is not the end game but rather an enabling technology for an important step in data center agility and efficiency.
The Service-Defined Era
Service providers promising hybrid cloud today are tied to this albeit temporary private cloud vision; that is, until they can deploy existing apps without virtualization. When that barrier is broken, virtualization and IaaS vendors will then start competing fiercely on service, SLAs, security and opex advantage. Service providers own more of the relationship with their customers because buyers become service and relationship-driven, versus being bound by various forms or layers of lock-in.
OpenStack has been discussed extensively as an alternative hybrid cloud infrastructure. It is in development. If it is completed it could be a hybrid cloud catalyst, especially for service providers.
OpenStack is important because solutions that break the locks win. We’ve seen this before, over and over again. Proprietary, hardware-bound solutions give way to proprietary software-bound solutions which can then give way to service-based solutions. The service providers break their dependence upon a single vendor or cloud platform, freeing them to deliver superior service and opex gains.
The current state of public + private = hybrid is, therefore, a temporary condition and is not the ultimate answer for the enterprise. Instead the equation is public X private = hybrid cloud, where existing, unmodified (and even multi-tier) apps are cloud-integrated without lock-in, and can operate across premises and clouds seamlessly and without dependence upon a virtualization vendor’s platform.
The next generation of hybrid cloud solutions will enable cloud migration and integration, with cloud failover, cloud devtest (cloud cloning) and cloud bursting. Look for these solutions to enter the market this year and next, and usher in the new age of the cloud service provider.
Stay tuned, the adoption rate for these new solutions could even exceed Datamation’s multi-billion hybrid cloud forecast.
Check out Greg’s articles here and at Archimedius.