Creating a Successful Cloud Migration
If you’ve been a part of the growth of cloud computing technology, you know that creating a successful cloud migration goes far beyond what can be covered in a short essay. However, this article will communicate guidelines or best practices that will greatly improve the success of your migration project. A successful cloud migration will include at least these three stages: planning, design, and execution. Each phase builds on the previous one and no step should be ignored or downplayed. The business cloud migration requires an expert, internal or external to the organization, to manage the process.
Planning: what type of cloud works best?
When we speak of a cloud migration we are referring to a business’s transition to Infrastructure as a Service (IaaS). Migrating to IaaS is the process of converting your on-site IT infrastructure to a cloud service provider and initiating an OpEx financial model for the business. When approaching this migration the business will investigate three provider solution types: hyperscaler, national cloud service provider and a hybrid of a cloud provider with a portion of the infrastructure remaining on-premises.
The largest public cloud providers, AWS, Azure, and Google are often referred to as hyperscalers. This name is appropriate as it is what they do best, allow customers to scale or expand very quickly. This scaling is served up by a self-service model via the provider's web portal which can be very attractive large organizations. Small and medium sized businesses (SMB) have a harder time adjusting to this model as there is very little support. Self-service means the customer is on their own to develop and manage the cloud instances. Another drawback of the hyperscaler for the SMB is that is nearly impossible to budget what your cloud infrastructure is going to cost. The hyperscalers transactional charges and billing make costs difficult to predict. The larger enterprise will often take the strategy of building the infrastructure as needed and then scale back to meet or reduce the cost. SMB typically does not have this type of latitude with budget constraints and will opt toward the more predictable national or regional cloud provider.
The regional or national data center is a better fit for SMB because of their ability to conform to the businesses needs. Often SMB will have unique circumstances requiring a customized plan for compliance and security or special network requirements. Also, this type of cloud provider will provide an allowance of internet bandwidth in the monthly charges. This eliminates unpredictable transaction fees the hyperscaler charges. In this way, the business can predict their monthly cloud cost and budget accordingly.
There are times when an application doesn’t work well in the cloud infrastructure, yet it is still required for the business. This is when a hybrid cloud environment can be implemented. Hybrid cloud in this instance is created when some applications move off-site while others stay and are managed separately. The challenge is to integrate, or make seamless, this non-cloud application with the other business processes. Over the long term, the application creating the hybrid environment can be repurposed to fit in the cloud strategy. Options include redeveloping the existing software to a cloud native architecture or finding a similar application that works more efficiently in a cloud environment.
Design: a cloud strategy.
A cloud strategy requires not only a strong knowledge of IT infrastructure but also a clear understanding of the business’s operations and processes. It is vital that the customer operations and management teams are involved in the cloud strategy development. Details regarding regular compliance and IT security need to be considered in the initial phases of development rather than later. The technical leader of the project will communicate a common strategy of building a cloud infrastructure wider as opposed to taller. Cloud infrastructure is better suited to have many servers with individual applications (wide) instead of one more powerful server handling many applications (tall).
Once all the critical business operations are considered, a cloud readiness assessment (CRA) can be developed. A CRA will dig deep into the business’s critical and non-critical applications and determine the cloud infrastructure needed to support them. In this stage, each application can be considered for its appropriate migration type. A “lift and shift” migration will move the application off-site as is, however some type of cloud customization may be completed before it is migrated. Connectivity also needs to be considered at this stage. This includes the bandwidth required for the business and its customers to connect with the cloud applications. Many times, an additional private and secure connection is required for access by IT managers or software developers through a VPN that will be restricted and have very limited access. IP addresses may need to be changed to a supplier issued IP block to accommodate the migration. This can create temporary Domain Name System (DNS) issues that require preparation. Finally, data backups and disaster recovery (DR) need to be considered. Many believe migrating to the cloud inherently assures backup and disaster recovery and it does not! Both backups and DR objectives need to be uncovered and planned out carefully.
Execution and day 2 cloud.
Now that the best cloud provider and the application migration timeline have been determined, the project is ready for the execution phase. The migration team should have performed tests on the applications as a proof of concept (POC) to assure everything will work as planned. After the tests are complete, the data will then be migrated to the provider via an internet connection or a physical disk delivered to the provider. The business’s IT infrastructure has now been moved to the cloud, but the work is not over. The business’s IT infrastructure is in a place called cloud day 2.
The two services that deliver and assure success in your cloud going forward are monitoring and support. These can be handled internally, or they can be provided by the cloud supplier or another third party. When purchasing the professional services from the cloud provider, it is important to understand their helpdesk operations and have expectations for response times. Make sure you discuss service level agreements (SLAs) for response both during business hours and after. The service provider should be monitoring the health or “state” of all VMs and network edge devices; security falls under these ongoing services. Many security-minded organizations prefer a more security focused third-party provider than the cloud provider itself. It is critical to understand the data backup services that have been included with your cloud instances. Don’t assume there is an off-site backup included in the cloud service, many data center providers have additional charges for off-site backup. DR goes well beyond backups and creates data replication with aggressive SLAs to restore service during an outage. An often-overlooked part of DR strategy is the “fallback” to your primary service location once the primary site has been restored to service.
A migration of IT infrastructure is a complicated process that needs to be performed by a team of experts. Just as important, the team needs to be managed by a seasoned project manager that has your business interests as primary. This is accomplished when the project manager is not a part of the cloud provider’s team. Having the right manager and team can assure your business can migrate to the cloud without a disruption to your business. Two Ears One Mouth IT Consulting can be the partner that guarantees a successful cloud migration.
If you would like to talk more about cloud migration strategies contact us at:
Jim Conwell (513) 227-4131 [email protected]
we listen first…