One of the advantages to cloud services over typical managed services, is that the service provider owns, operates and maintains the assets necessary to provide the service. However, many large organizations are adopting cloud in a piecemeal fashion and simply using cloud services to augment and extend their existing on premise provided services. Bolting cloud onto the existing infrastructure makes the infrastructure more complex. Now cloud based applications, databases and services have to communicate with on premise applications, databases and services or other cloud based applications, databases and services. The result is that the infrastructure architecture and the application architecture has not been simplified or collapsed; we have further complicated it.
As well, IT organizations build “Frankenstein clouds” (hybrid clouds) which might be viewed as an on premise “excuse” for not moving directly to the cloud. Alternatively, Hybrid clouds may be a good interim solution for some organizations. They enable an organization to move to cloud in stages, managing risk and simply deferring some applications. Organizations can realize limited cloud benefits while focusing on migrating other more suitable applications. The concern is that these hybrid clouds never go away. Like Mary Shelly’s monster, they never die. They lock the organization into their data centre. The combination of cloud, hybrid/on premise and legacy infrastructure creates an even more complex environment that is harder to manage and administer. In the IT world, complexity often translates into unreliability, lack of agility, higher costs and higher risk.
Treating cloud as another managed service and taking the view that the organization will slowly and naturally evolve into the cloud means that the organization may spend many years in a transition state. A complex combination of cloud, hybrid cloud and legacy. The objective should be to eliminate on premise infrastructure, close or sell data centers and reduce the hardware footprint of the organization as quickly as possible. Fundamentally, there is little intrinsic value for businesses to own servers, storage, and data centre(s).
It is important that enterprises implement an IT Strategy that mandates migration to cloud as quickly as possible. Organizations must build a detailed inventory of all of the applications, technologies and services they have and develop a plan/roadmap. This cloud strategy needs to support an overall digital strategy for the enterprise with cloud being a key enabling component. The roadmap needs to consider what applications and services can be moved to cloud first. What technologies are at end of life or nearing end of life? Wherever possible, no further investment in on premise infrastructure should be made. Instead investment should be focused on migrating the application or service to the cloud. In some cases, organizations will want to sweat an asset. In other cases, the cost, effort and risk of migrating to the cloud is simply too great. To address these issues, organizations may choose to use hybrid cloud (to realize some benefits from cloud) or ring-fence the service in the data centre. However, the longer an organization remains in this dual/transition state, the higher the cost, complexity and risk. Also, the more likely it is that an organization remains locked in a transitionary state for an extended period. Finally, as the number of assets are removed from the data centre(s) the cost per unit remaining rises (fixed costs remain largely the same). The benefits of going to the cloud are eroded by the costs of the legacy infrastructure and services remaining. For all these good reasons, organizations must move quickly and aggressively to the cloud.
Some organizations will implement a “Cloud First IT Strategy”. This too is problematic as it implies that there may be a 2nd or 3rd viable strategy if cloud is not selected. There needs to be a separate Transformation organization or unit with the required architecture, security, risk, compliance, financial and technical skills that is responsible for driving services to the cloud. This organization builds the cloud governance model, certifies cloud providers and enables business units to consume cloud. This organization builds the cloud migration roadmap and the investment and lifecycle profiles for the application portfolio. There is also the need to refocus the IT organization from system integration and operations to service management and business enablement.
The Transformation organization also manages the change across the company and the adoption and consumption of cloud services. This organization needs to be able to monitor the disparate environments looking for cloud sprawl, excessive spending/consumption and shadow cloud use. When signing up for niche cloud services involves a company credit card and a few clicks, the risk of data loss or data leakage rises. Educating the user community on cloud services and risks is therefore another priority. Ultimately, IT needs to be an enabler, working in close cooperation with the business and business application teams.
Moving to cloud, in an accelerated time frame can be difficult. Cloud is not a neatly encapsulated managed service. For large enterprises with complex legacy and infrastructures and thousands of business applications there needs to be a centrally managed and coordinated approach to cloud adoption.
Simon Morris is a Digital Transformation leader at KPMG. When his head’s not in the clouds, he is riding his bike, carving turns on his snowboard, or building water cooled personal computers with his son. He can be reached at firstname.lastname@example.org