Cloud Thunder: The Biggest Bang for the Buck, Part 1
May 5, 2011 9:47 AM PT
Cloud deployment and delivery architecture offers many opportunities to reduce costs and improve service levels -- sometimes too many for the beleaguered IT team. So how can your IT team identify what cloud components might be considered early opportunities for your infrastructure?
Cloud deployment strategies provide on-demand services from a virtualized environment, making the underlying technology transparent to the consumer. Some examples of services in the cloud model include Infrastructure as a Service (IaaS), Software as a Service (SaaS) and Platform as a Service (PaaS). These services may be provided by internal or external service providers, or both.
Typically, Web-based portals provide automation to quickly deploy resources as services from a virtualized pool on a pay-per-use basis. The internal cloud provider will typically have implemented the ITIL service provider model. Let's look at each of the cloud strategies, and consider benefits and speed of adoption possibilities.
Infrastructure as a Service, or IaaS (Internal)
Virtualization of the in-house x86 platform and associated storage is an indicator of readiness for transition to an internal service provider model, offering IaaS solutions to the consumer. If the journey has not yet begun or is still in process, IaaS will not provide early opportunities visible to the end consumer. The essence of internally provided IaaS offerings is the adoption of the ITIL-based service provider model. This includes the ability of IT to provide a catalog of services and deliver them under a signed service level agreement (SLA) covering defined service attributes.
The service attributes often include availability, performance, security and mean time to provision. In addition, the IaaS service must be capable of rapid provisioning -- and equally rapid deprovisioning when the service is no longer required. Last but not least, the IaaS model requires that these services can be sufficiently identified to support a billing mechanism, and that means that IT must have a capability of determining the unit cost of the deployable resource, typically GB of storage and GHz of compute power.
The positioning of IT as an internal provider of IaaS offerings is desirable for several reasons. It provides significant savings from a more cost-effective and efficient infrastructure administration by improving service levels through automation and often self-provisioning, and through cost-conscious resource usage. While these benefits are valuable, the attainment of them is not a trivial short-term task.
These days, it is easy enough to virtualize, and many organizations are already well on the way to transitioning their x86 platforms to this model. Indeed, virtualization of the x86 platform is on the edge of being considered a best practice for the effective administration and protection under both operational and disaster modes for the x86 environment.
However, virtualization is only half the story. IaaS requires the implementation of governance under the ITIL best practice of a service provider model (SPM). The SPM also needs to be supported by automated self-provisioning if it is to provide effective IaaS offerings, and these offerings need to be billed out appropriately. Billing requires development of a costing capability that supports determination of a fully loaded unit cost of the deployable resource. This is inevitably a major change in internal financial policies to support charge-back.
There are many benefits to being an internal provider of IaaS offerings, but the attainment of this capability is certainly not "low hanging fruit." This does not mean such projects should be shelved or suspended; there are significant benefits along the way at each point in the plan. There are immediate benefits to virtualization in more cost-effective (and reduced) IT administration.
Additional benefits can be experienced in IT/business alignment that comes from development and publication of the services IT provides in a SPM. The SLA signature demonstrates IT's public commitment to defined services, and the costing model makes it possible to report usage in the common language of finance -- even without charge-back.
The biggest impediment to the introduction of IaaS under IT is that the service provider is the requirement for some form of portal/Web-based self-provisioning capability. Nonetheless, this journey needs to be started and aggressively progressed with budget, focus and urgency, but it is not prudent to expect highly visible "big bang" benefits in the short term.
Infrastructure as a Service, or IaaS (External)
Using an external provider for IaaS may make more sense. However, it is important to understand the limitations that may accompany this service once outsourced. The one limitation that concerns many IT teams is the issue of security. The data is off-site, at another organization, on infrastructure that is shared with perhaps hundreds of other organizations, or on hardware that may be dynamically managed with data moved around as policies dictate.
Understanding the security of your data in these circumstances may require considerable due diligence. However, if your data requires security, then you should avoid storing it outside your internal controls.
The other major issue to consider is your back-out strategy. Should the vendor fail to meet your IT needs, how you get your data back, and more importantly, how do you ensure no trace of it is left behind? It is also important to look at availability guarantees and what the guarantee means. Refunds don't do you a lot of good, so including a penalty clause is a good incentive to ensure the vendor has sufficient motivation to publish realistic goals and develop the capabilities to support them.
Finally, it is important to get scalability issues under control. What scalability does your contract support? If you exceed it, is there a premium? Can the vendor provision higher-level scaling within your required time frame? If you are satisfied that security, stability, availability, performance, scalability, back-out capabilities and pricing are all acceptable, you are well positioned to outsource IaaS.
In addition, there is a hybrid model available for the more sophisticated IT team. When IT has already deployed a successful internal IaaS, they may choose to look at outsourcing these services for overload conditions. A single organization has less flexibility in supply-and-demand management and in funding for demand projections than the external service provider. These providers service hundreds, if not thousands, of customers and can fund and build larger "spare" capacity.
Using the external IaaS as a buffer for overflow creates a hybrid provider model ensuring that IT will never be in the embarrassing position of running out of resources. However, it is essential that IT considers the same acceptance criteria that apply to selecting an outsourced-only model.
To Be Continued...
In the second part of this feature, I will take a look at two additional examples of services in the cloud model -- Software as a Service (SaaS) and Platform as a Service (PaaS).