IT departments are reluctant to acknowledge it, but their users now expect the high levels of flexibility and immediacy associated with today’s leading cloud services. While forward-thinking IT managers are moving to meet those expectations and deploy more comprehensive private cloud services, they often encounter a number of unforeseen challenges—in particular building a collaborative service delivery relationship with users—as they try to balance the levels of control they are accustomed to with the many user demands they are trying to meet.
So how does an IT organization begin to offer effective cloud services without starting at square one? The answer is, they must develop a pragmatic approach to service delivery that provides tangible value quickly—even if it is narrowly focused at first.
With that in mind, here are 10 guidelines for building out practical and effective private cloud services within a traditional IT organization:
1. Start small and stay nimble: You don’t need a total IT transformation. Begin the move to cloud by pulling together a cross-functional team of individuals who know the organization but are not locked into a business-as-usual mentality. This team should be empowered to function autonomously outside of organizational stovepipes, so they can help manage the cloud transition without getting caught in needless bureaucratic hassle.
2. Work with a particular need or use case in mind: The number of business workflows and supporting applications in an IT organization can be daunting. Focus on identifying a particular need or use case that is technically well suited to the cloud model (for example, virtualized, distributed and x86-based processes) and can clearly derive benefit from the attributes of the cloud (including rapid provisioning and dynamic expansion/contraction).
3. Enlist willing business partners: A critical success factor to making a private cloud work for the business is the close alignment of cloud services with user expectations. Achieving this requires more of a peer-to-peer relationship between IT and users rather than a traditional hierarchical relationship. IT and non-IT employees need to be upfront and open about their needs and expectations.
4. Clearly define cloud offerings based on specific service attributes: Another key success factor is managing user expectations more effectively. The first step to achieving this is to identify the specific attributes of IT services—performance, availability and cost—in understandable business terms, and to compile them in a comprehensive service catalog available to all employees.
5. Define service level agreements (SLAs) based on reliable service metrics: The other essential piece of managing user expectations is spelling out clearly defined metrics and providing regular reports to users. This way, users understand clearly how IT services are provisioned and delivered, as well as the success factor.
6. Know your true current costs before designing your cloud: Building a cloud service that looks like [insert your favorite cloud vendor here] but without their economies of scale and commodity-based infrastructure is likely to produce sticker shock. It’s necessary to understand your real costs and adjust service capabilities and platform components to align with tolerable and realistic cost expectations.
7. Make sure your infrastructure and software platforms are current: This may be a no-brainer to some, but it’s important to point out that moving legacy environments to the cloud is not a plan for success. Achieving efficiency expectations and ensuring a manageable cloud environment with the cloud OS platforms available requires an up-to-date infrastructure that can handle the transition.
8. Select a cloud management tool that complements (not replaces) your existing infrastructure toolset: Management of cloud services requires an additional set of functions, such as automated provisioning and life cycle management, that go beyond the capabilities of traditional infrastructure management tools. Adding these functions shouldn’t mean replacing existing tools. Look to cloud management platforms that integrate with your existing tools.
9. Determine how high up the stack you want/need to go: We’ve all heard about Infrastructure-as-a-Service, Platform-as-a-Service and Software-as-a-Service. IaaS is a necessary building block for PaaS and SaaS, so it is natural to begin initial cloud efforts there, but it may be important to offer services higher up the stack as well.
10. Communicate, communicate, communicate: Having defined service catalogs, SLA-reporting and a close working relationship with IT all help to ensure good communication and close alignment with business goals. It’s important to have a strong feedback loop and not to be afraid to revise service offerings to accommodate evolving user needs. The key to cloud is agility and flexibility. The cloud team needs to be responsive to this.
It should be clear that the private cloud is much more than a group of technologies integrated together, and it’s more than just another deployment project. It requires a new way of thinking about how to deliver IT services, and takes some time to envision and plan how it will affect a particular organization and impact success. However, change is good, and by evolving to become cloud providers, IT organizations will realize levels of service scalability, flexibility and performance, as well as a closely aligned relationship with their users, that they could never before achieve.
Jim Damoulakis is CTO of GlassHouse Technologies, a leading independent provider of IT Infrastructure Services headquartered in Framingham, MA. Jim brings over 25 years of experience in systems, storage, data protection and recovery, cloud and virtualization, with a strong focus on architecting and implementing highly available enterprise IT solutions.