The path to cloud computing isn’t particularly clear, and IT transformation can’t happen overnight. This sort of change requires an incremental, stepwise progression that yields benefits along the way. This is the explicit goal of the Cloud Computing Adoption Model: a graduated approach for adoption of cloud technologies that allows benefits to be realized incrementally.
The reality is that change of this magnitude always creates new risk. The Cloud Computing Adoption Model helps to lay out a clear path to cloud without putting projects, budgets and even careers at risk. Central to this adoption model is the virtualized application, a self-contained application image that includes all of the operating components and systems software it requires to run in production. This is the vehicle that makes it possible for applications to become portable across platforms, scalable to meet dynamic demand, and available on-demand.
By packaging applications as virtual applications (also known as “virtual appliances”), organizations ensure that the application will remain manageable and controllable throughout its lifecycle. This article will explore a five-step framework for cloud computing adoption that begins where many organizations are today-at virtualization-and ends with true cloud actualization, which is the state most organizations ultimately want to achieve.
Level No. 1: Virtualization
Level No. 1: Virtualization
Virtualization of applications and infrastructure is the foundation for cloud computing. Today, much of the visibility around virtualization relates to the hypervisor, which allows server infrastructure to be carved up and more effectively utilized by application workloads.
The other side of virtualization is about delivering applications as coordinated sets of virtual images. These self-contained units consist of the application, database, middleware and other components, packaged together with necessary operating system bits-known as just enough operating system (JeOS). Virtual images enable applications to be deployed on any hypervisor and moved seamlessly from one virtual server to another.
This approach essentially eliminates the barriers to application deployment, including manual configuration, and tuning and certification of application workloads with underlying systems. Maintaining and updating deployed applications also become easier and faster. All this speed and flexibility adds up to dramatically faster deployments, which allow organizations to more rapidly realize application value and increase business responsiveness.
Level No. 2: Cloud experimentation
While cloud computing is the next logical step beyond virtualization, it can’t happen without a plan based on real experience. Smaller steps are needed to be taken in order to build knowledge, gain understanding and amass experience. So the next step of the cloud adoption model involves usage, experimentation and laying the groundwork for further cloud initiatives.
Amazon Elastic Compute Cloud (EC2) is a good example of a mature and stable cloud infrastructure that offers a low barrier to usage-swipe a credit card and you’ve entered the cloud. At this stage in cloud adoption, we recommend establishing an EC2 account and running a non-mission-critical application in the cloud. The team needs to build a base of experience in cloud infrastructure, processes and operations.
It’s also a good idea to get application architects involved in thinking about the reference architecture for cloud computing so that future applications are designed with cloud deployment in mind. Strong executive sponsorship needs to come online at this point to ensure proper funding is available and a realistic timeline is established.
Level No. 3: Cloud Foundations
Level No. 3: Cloud foundations
At this stage, you’ll need to begin establishing the foundation for an application architecture that will support what is likely to become an explosion in volume and scale. The reality is that virtualization lowers the barriers to application deployment. By reducing deployment friction, levels of demand and volume of application workloads can go through the roof.
At this stage, you should have a platform in place for managing the life cycle management of your virtualized applications; ensuring the consistency, repeatability and maintainability of virtualized application images, and providing a foundation for managing a vast increase in virtual machines and application images that must be managed.
Don’t skip this stage. While you might be tempted to go right from experimentation to full-scale application deployment, you could create new problems if you don’t have a solid foundation of procedures, policies and tools. Take your experimental applications and roll them out to a broader group of pilot users in the organization. Work through the process of requisitioning, provisioning and decommissioning applications on demand. Use your life cycle management system to create, configure and maintain virtualized applications. Get comfortable with all this before going on to Level 4.
Level No. 4: Cloud advancement
In Level 4, you either need to build your own internal cloud or commit to a commercially-available cloud. You’re now ready for full-scale cloud deployment, albeit without some of the automation you’ll see in Level 5.
Your application infrastructure should be solid at this stage and implemented across the organization. Get your virtualized applications into production, along with the processes, policies and procedures that you established in Level 3. Fine-tune and tweak them as you go along but, if your preparation work was good, these changes should be minimal.
Take a moment to sit back and congratulate yourself and your team for your accomplishments to date. Enjoy the view from the clouds!
Level No. 5: Cloud Actualization: Hypercloud
Level No. 5: Cloud actualization: Hypercloud
Level 5 represents complete “actualization” of cloud computing. This level is certainly still on the horizon for most every enterprise today. At this level, you’ll see a fully-dynamic and autonomic compute environment that allows application workloads to become something of a currency, shifting dynamically across compute environments based on availability in services, capacity and even comparative cost advantages-allowing organizations to begin arbitraging compute capacity. Although the technology for these functions is not mature, we believe it is on the horizon and it will become fundamental to full cloud actualization.
By Levels 4 and 5, you’ll have seen the transformational effects of cloud computing on your users, IT infrastructure and enterprise profitability. By starting with virtualization and following this model, you’ll be laying the foundation for cloud computing, while taking advantage of the nearer-term benefits virtualization can provide.
As with any process model, the steps toward cloud computing adoption build on each other. Don’t jump ahead or you’ll undermine your foundation. Work the program, and keep your head firmly on your shoulders while also having it in the clouds.
Before Mercury, Jake led marketing for two WebSphere products at IBM Software Group, which he joined through the acquisition of Venetica. Prior to Venetica, Jake was director of product marketing with Documentum, Inc. (now part of EMC), which he joined through the acquisition of eRoom Technology.
Jake has a B.A. in English and political science from University of New Hampshire, and an MBA from the McCallum Graduate School of Business at Bentley College, where he was an American Marketing Association George Hay Brown Scholar. Visit his blog at http://blogs.rpath.com/wpmu/closing-the-gap/ or he can be reached at [email protected].