Don't let uncertainty stand between you and cloud
Virtualization has taken enterprise data centers by storm—and you'd think that private cloud implementation, as the next logical step, wouldn't be far behind. At many organizations, however, a number of barriers stand in the way of that happening.
Some are organizational, of course. Common examples in this category include skepticism about cloud hype, uncertainty about the security or compliance ramifications of such a unified architecture, and simple resistance to change.
Technical or procedural obstacles are very common as well. If an organization has already virtualized its IT infrastructure extensively, what kinds of solutions will be needed to migrate from that to a full-blown cloud? What will the impact on business processes be—will things get more complex, less complex, or remain largely the same? And which vendors have the proven history of successful cloud engagements and command of best practices needed to help?
For organizations asking such questions, IBM is in an excellent position to help. IBM has worked with hundreds of organizations of different sizes, in different industries, to achieve the best possible outcome from private cloud deployments. And that extensive experience has resulted in a clear understanding of both the hidden pitfalls and the most promising opportunities.
Furthermore, because IBM is also a leading cloud solution provider, IBM solutions have been developed specifically to leverage that insight—helping organizations make the transition process to cloud as smooth and optimized as the cloud they'll eventually get.
Start small—and make the most of what you've already got
"SmartCloud Provisioning delivers—in spades. It not only provisions virtual servers accurately, but does so with incredible speed—literally thousands of new virtual servers per hour, if need be, all based on predefined business policies."
One of the first key points to consider is this: it's not necessary, or even ideal, to make the cloud transition in a single step. Instead, a private cloud is often better implemented in logical stages.
By developing core capabilities first and ensuring that they work as intended, the organization can gradually develop the cloud over time, leveraging it in new ways, assigning it new workloads, and adding new capabilities to it in parallel with business goals and strategies.
If you think of a cloud not as an abstract IT infrastructure but a set of concentric circles, this concept becomes simpler to understand. The goal should be to develop the innermost circle first—the fundamental set of capabilities that makes a cloud a cloud—and verify that it performs as specified. Subsequently, the outer circles can be added. Each of those outer circles will represent new capabilities that either add value to, or directly utilize, the core capabilities.
Such a strategy is much easier to manage than creating a full-featured private cloud from scratch. It also helps the organization control costs and more easily align the cloud to its priorities as they change over time.
It's also typically not necessary to pursue a cloud in a rip-and-replace fashion. That is, the existing virtualized infrastructure should be leveraged as much as possible before significant new hardware or software investment is made.
If the organization already has dozens of physical hosts that run hundreds of virtual servers, why not use some percentage of them for the new cloud?
One of the most common implementations along these lines is x86-based hosts running VMware virtualization solutions. If you're interested in leveraging this type of infrastructure to create a private cloud, IBM can collaborate with you to add new capabilities as necessary to achieve exactly that goal. You’ll quickly dial up the business value you get, while also minimizing the hardware investment you might have expected to make—not to mention the management learning curve your IT team might otherwise have faced.
IBM SmartCloud Provisioning provides the core functionality you need to move from virtualization to cloud
Of course, hardware is only part of the cloud story; it's really software that governs the logical actions the cloud takes. To move from an everyday virtualized environment to a true private cloud, you'll need to add software-driven capabilities—that innermost circle of core functionality mentioned above.
Toward that end, one particularly apt solution you might consider is IBM SmartCloud Provisioning. As you might imagine from the name, it was created specifically to help organizations implement cloud capabilities in the smartest possible way—and it's directly on point in helping them evolve from basic virtualization to the next level.
How? The story begins with its discovery/analytics features. These address a very common problem in virtualized infrastructures: image sprawl. As organizations have leveraged virtualization, they've typically created a large and unwieldy library—or multiple libraries—of virtual images. And over time, as those images proliferate and are provisioned, it becomes harder and harder for IT to understand exactly which images (or virtual servers) contain which software elements. So the total number of images tends to spiral out of control, and the whole infrastructure becomes less efficient and more difficult to manage for best results.
SmartCloud Provisioning solves this problem very neatly. Thanks to its discovery capabilities, it can find images and virtual servers throughout a virtualized infrastructure, then establish what each of them contain in a clear and comprehensible way. It manages all this in an agent-free manner, too, which speeds the total process.
The result is that it becomes far easier to eliminate redundant images, consolidate libraries, and verify which images and virtual servers need to be updated or patched—and which don't. What had been a complex, ungainly pile of images becomes instead a lean and mean engine of virtual server creation.
Of course, creating those virtual servers in the cloud involves much more than just managing images. It also involves actual provisioning, to fill newly generated servers with logical contents—the complete software stack drawn from the appropriate image as determined by the intended business functionality of each server.
Here, too, SmartCloud Provisioning delivers—in spades. It not only provisions virtual servers accurately, but does so with incredible speed—literally thousands of new virtual servers per hour, if need be, all based on predefined business policies. That translates into a private cloud that can much more rapidly scale to meet unexpected workload spikes—the difference between a successful implementation and a failed implementation, in many cases.
Naturally, over time, this private cloud will have to change in ways large and small. For instance, operating systems vendors will release security patches; these will have to be rolled out, both to the relevant images and the relevant virtual servers currently executing in the cloud.
At this point, the discovery/analytics features will come in very handy a second time. They'll provide clear insight into exactly which virtual servers and images are indeed relevant, minimizing the work the cloud has to do in provisioning the patches and helping to keep the cloud as continually secure as possible. Security, regulation compliance, and software licensing worries—all very common for organizations considering a cloud—will all become much smaller as a result.
All of these capabilities, as suggested earlier, are directly applicable to an x86-based VMware-driven virtualized infrastructure. SmartCloud Provisioning, however, offers a much larger value proposition than that.
Because IBM designed it to be hypervisor-agnostic, it will work not just with VMware, but with other virtualization environments altogether—for instance, Microsoft's Hyper-V or Xen. In fact, SmartCloud Provisioning can even be used with completely different hardware platforms, such as IBM Power Systems, which leverage IBM's POWER 7 processor architecture and a fundamentally different virtualization platform.
This design has the effect of logically pulling together what had been separate environments, and empowering the organization to manage them all in a much more centralized, flexible, and scalable manner for more business value. And if that description sounds familiar, it should—it's often the idealized way private clouds are described.
Add new power to your cloud based on business goals and priorities
Once these initial, core cloud capabilities have been deployed, you can then augment them with more capabilities at will—making the cloud more powerful in the ways you need, at whatever point in time makes the most business sense.
And IBM, as a leader in cloud computing, offers a variety of solutions to help you accomplish that. Among others, these include:
The goal should be to make the private cloud as automated, and yet performance-optimized, as possible, across the complete lifecycle of all of its hosted services. And by first deploying core cloud capabilities, and then augmenting them as needed, organizations can make exactly that happen—in surprisingly little time.
Discover a bright new day for your application infrastructure
Register now for the IBM SmartCloud briefings in Boston, Austin or Washington D.C. and in just one morning you’ll get the most information for improving your infrastructure with techniques and solutions that help you better connect across existing IT and Operations silos.