The concept of cloud is now heard in almost every board room, forum or CIO/CTO’s weekly agenda. Whilst cloud makes an impressive proposition in terms of return on investment (ROI) and total cost of ownership (TCO), the main challenge faced by many early adopters is to establish the most effective transition strategy to the cloud.
When transitioning your applications to the cloud there are a number of considerations, and whilst it may be possible to take your entire IT to the cloud in one step it is more likely that there will be a gradual transformation to cloud computing depending on the business applications and their relative importance.
What can be moved?
For those operating entirely in a traditional, non-cloud IT environment, the transition to cloud requires more thought, rationalisation and planning. There is a need to define what services, applications or target environments you are considering for cloud, whether it is production, development or QA/test, etc. Each environment will have specific needs which present challenges in delivering a consistent service in the cloud.
As your business and applications change and develop throughout their lifecycle, increasingly extensive migration planning is required, which can involve the transition from a legacy self-managed model to that of a utility-based managed service cloud.
Migration is not transition
There are common misconceptions about the difference between the requirements of data centre migration and cloud transition. Cloud provides flexibility, scalability and improved speed to market, and cloud providers typically deliver compute, storage and network based on a pay-as-you-go model.
Data centre migration, physically relocating existing infrastructure to a new data centre, does not offer all the benefits of cloud and should be considered a distinct initiative or project from cloud transition. Relocating infrastructure presents some risk, including power failure and end-of-life hardware support, which may impact the availability of your application or service.
Different applications, different demands
Some applications have stringent security requirements, such as data privacy and sovereignty, which present additional challenges in transitioning these applications. Applications may also have geographical limitations where storage of data is constrained by country boundaries for legal or compliance reasons and as a result, gradual transition may not be a suitable option in this case.
Enterprise customers require cloud providers that offer managed cloud services, skills, experience to help take that transforming step whilst maintaining compliance as well as assistance in determining which applications are suitable for cloud.
One example is the bespoke legacy application, created in-house with limited or no support knowledge available, but still vital to the business. This application can still be transitioned to the cloud using an image of the existing environment. Such transitions cannot happen all at once and each application needs due diligence and planning.
Where do you start?
To overcome these and other challenges and ensure the correct strategy is applied, each application first needs to be analysed and mapped to each infrastructure component and given a criticality index based on various factors (i.e., business impact, downtime constraints, uptime or availability requirements, custom SLAs, etc.).
This discovery and analysis maps the infrastructure and application inter-dependencies. Applications are then grouped together based on this analysis. All the applications in one group could then be transitioned at once.
Infrastructure components are built and made available just-in-time for each group, utilising the pay-as-you-go model that cloud Infrastructure-as-a-Service offers, ramping up resources as needed.
Options for transition
To achieve a managed and successful transition the most common approaches are to re-host, re-factor or replace your applications. The reasons to choose any of these strategies will largely depend on several factors around the application lifecycle (i.e., hardware and operating system’s remaining serviceable life and hardware affinity).
Re-host: Application is reinstalled on a new hosted platform in a managed cloud environment. Data replication and mirroring technologies are implemented to keep the new and old data in synchronisation.
Re-factor: Industry standard third party tools are used to convert a current application image into a virtual instance to be hosted within private, public or hybrid cloud.
Replace: Discarding legacy applications and using modern equivalent software delivered as a service. Typically only data requires to be transitioned onto the new Software-as-a-Service environment.
The journey to the cloud largely depends on a business strategy and individual application intricacies. However, gradual transition reduces operational risks immensely. Whatever the requirements of a cloud transition project, success depends on extensive planning combined with the support and expertise of a trusted cloud service provider.
By Joao Crespo, practice manager for Savvis cloud solutions & transformation services, EMEA, Jason Farrow, cloud consultant and Amit Dalvi transition architect.