I described in my previous blogpost the first five steps as an approach to evolve from a traditional IT environment towards cloud computing. Let me recap in the list below the first five steps:

1. Building the inventory of workloads in your current organization
2. Prioritization of the workloads to determine the best candidates for cloud computing
3. Determine the potential destination of the workloads in the cloud
4. Determine the best migration path for each workload
5. Validation of the ability to exploit cloud storage

You can read the details of the steps in my previous blogposts. Let me now take you through the next two steps.

6. Plan the evolution of your datacenter

As the evolution towards cloud computing becomes inevitable, it will have a seriously impact on the infrastructure of your datacenter and on the operations in your datacenter.

Moving workloads from a traditional delivery environment in your datacenter to a private, a private hosted, a shared, or a public environment will impact the needs of your physical datacenter, being it in size, power, cooling, etc. Therefore, this impact must be taken in account in the planning of the future of your datacenter.

Below six reflections you can use when planning the evolution of your datacenter:

1. When all workloads can and will be moved to a private hosted, a shared, or a public cloud environment then the own datacenter becomes obsolete. In such case you can consider to prepare the phase-out of your datacenter, taking into account the current contracts in place for the real estate, infrastructure, services, network, etc.
2. When a considerable amount of workloads will be moved to a private cloud then the own datacenter might have insufficient power and cooling capacity for a large high density configuration. The power and cooling issue will be critical during the transition phase. The total power consumption should decrease once the move from the traditional environment to a cloud environment is completed. The occupation of the datacenter should be lowered, releasing space for new projects.
3. When your current datacenter becomes saturated then consider moving new workloads into a private hosted, a shared, or a public cloud environment. This will allow you to keep datacenter capacity in case you need to extend the environment for your current workloads.
4. Running operations across multiple datacenters is a common practice, in particular to implement a disaster recovery environment. If you have an active-passive configuration then you could consider moving the passive environment into a private hosted, a shared, or a public cloud environment. The passive environment in the cloud can be much smaller in scale as you can scale-up using cloud elasticity of your cloud provider when the environment becomes active. Phasing out a passive datacenter can bring considerable savings.
5. Development and test environments are frequently a headache in a datacenter, caused be continuously changing demands, short term needs, and unexpected needs. A shared development and test cloud environment to accommodate different development and test projects can be a relief, as developers and testers can exploit the self-service and automation capabilities to fulfill their needs. Consolidation of development and test environments onto a shared cloud environment should reduce the space needed in the datacenter.
6. Outdated datacenters that need renewal or upgrading are ideal candidates are an ideal starting point for cloud computing. These workloads should be considered for a move into a cloud environment. Even if not all workloads can be moved to the cloud, then a smaller datacenter might be sufficient to house the remaining workloads.

7. Select your cloud service provider(s)

In the previous six steps we identified the workloads, determined the cloud needs, and handled all issues needed to complete a cloud roadmap. In this step we move closer to the cloud by selecting a cloud service provider. To select a cloud provider we pay attention to the following:

1. Many companies say they cloud, but actually do ‘cloud painting’. They reposition traditional IT solutions and give these a cloud flavor, as they try to surf the current marketing trend of cloud. Make sure you go for cloud, and don’t step in the pitfall of solutions limited to virtualization or services which are traditional web services.

- Validate that the solution / services relies on standardization and has full automation.
- Validate the scalability of the solution / service, as cloud services should scale according to the flexibility of your needs.
- Validate the contract flexibility of cloud services, long term commitments with a high baseline indicate traditional delivery models rather then a cloud delivery model.

1. Select a proven provider, which will come with a cost, but ensures a secured and reliable solution. Do not confuse professional cloud solutions with ‘free’ consumer services. The latter are receptive to security incidents and you most probably give up confidentiality and/or privacy.
2. Cloud can scale rapidly, ensure you and your cloud providers have the security mechanisms in place. This typically requires implementation of automation. As cloud services can have a small granularity it might require the implementation of a detailed security policy (access people vs. data vs. service).
3. The contract basics of cloud are comparable to outsourcing, thus pay attention to service levels, security standards, data access, data protection, and auditability of the environment.
4. Avoid a vendor lock-in, thus rely on open standards and ensure you have appropriate exit tools / possibilities.
5. The contract termination must be from both sides, to avoid an unexpected end of service which could endanger your business.
6. Don’t allow click accept buttons or in-application purchases, as this would loosen your control on employee spending.
7. Finally, hoping you don’t need them, make sure you have the escalation procedures in place!

In my next blog I will further elaborate on the last steps towards cloud. Come back soon!