AWS cloud migration strategies: a deep dive

What is AWS migration?

AWS migration is moving a company’s databases, applications, servers and other IT workloads from the on-premise legacy infrastructure to the AWS public cloud. Cloud migration can be performed for selected workloads or for the whole infrastructure at once.

7R’s of AWS migration

Back in 2011, Gartner outlined the five most effective cloud migration strategies (‘5Rs’), which then were updated to 6Rs in 2016, and 7Rs in 2017 by the current AWS Vice President Stephen Orban. This gradual evolution of the best cloud migration strategies can be largely attributed to the periodic introduction of innovative AWS cloud migration services like AWS Outposts and AWS Application Migration Service.

Let’s figure out what the 7R’s are and how companies can apply them to streamline cloud migration. The following migration strategies are sorted based on the level of effort needed for implementation (from lowest to highest):

Retire

Retiring means removing applications that are no longer needed in your IT portfolio. 

 Pros

• This is a strategy that requires the least amount of time and resources. 
• Retired applications no longer need IT resources, which minimizes performance costs. 
• Resources no longer used by retired applications can be redistributed to improve the performance of other workloads. 

Cons

Retiring applications can be complex due to insufficient legacy documentation and quite risky if the field experts are no longer employed by the organization. 

When to retire

If an application has proven to be useless in the AWS cloud environment, it can simply be removed. In the 2016 version of 6R’s, Stephen Orban claims that up to 20% of applications in an average IT portfolio are no longer useful in a cloud environment and can be eliminated. 

Retiring tips 

Seemingly rational, it’s not uncommon for organizations to postpone retiring applications until other essential workloads are migrated. However, this has proven to be risky as organizations often realize that they actually need to migrate applications that initially were deemed no longer useful in the new environment. As a result, little time and resources are left at the very end of the project. Therefore, it’s highly advised to retire applications at the beginning of the project because it also frees migration teams from certain server maintenance tasks.. 

Retain

Retaining means keeping applications in your existing environment.

Pros 

• Keeping some applications on-premise reduces cloud waste and allows for redistribution of resources for other important workloads. 
• Retaining allows businesses to leverage the benefits of the cloud while leaving some of their workloads in legacy environments.  

Cons

Applications that remain on-premise often drain valuable IT resources. 

When to retain

It may be reasonable to retain applications that need significant resources for refactoring. These are usually the workloads that can’t be run in the cloud due to strict compliance requirements, compatibility issues, security concerns, unacceptable latency times, or insufficient ROI. In the majority of cases, retaining is applied for large-scale hybrid cloud projects that can take several years to complete. 

Relocate

Relocate is the 7th R, which was added due to the release of VMware Cloud on AWS in late 2017. In the AWS migration context, relocating means using VMware Cloud to move an IT infrastructure to the cloud without changing its code, investing in hardware, or disrupting workflows. 

Pros

• Developers can use familiar VMware tools, which eliminates the need for employee retraining or modification of existing workflows. 
• Given that apps require little to no adjustments, workloads can be migrated relatively quickly without interrupting existing processes. 
• Companies which use VMware data centers can replicate them in the cloud using vSphere elements, allowing moving servers without causing disruptions. 

Cons 

Doesn’t support some cloud-native features (for example, VIBs on the ESXi host)

When to relocate

Relocate migration strategy is applicable only to VMware vSphere-based workloads and is the best option when migrating these types of apps. 

Rehost

Also known as lift and shift, rehosting is the most popular migration strategy among enterprises. Rehosting implies moving applications to the cloud with little to no modifications. 

Pros

• By quickly increasing system scalability and accessibility, rehosting can bring immediate economic returns. 
• This is the least stressful and demanding cloud migration strategy since it allows both companies and users to gradually adjust to the new environment. 
• Rehosting requires no employee retraining.  

Cons 

• Given that rehosted applications aren’t cloud-optimized, they can fail unexpectedly, start lagging, or perform poorly in general.
• If a company decides to keep rehosted applications intact, cloud costs can rise significantly, making migration economically ineffective in the long term. 
• Rehosted applications that aren’t optimized for the cloud can have a limited number of cloud-native features.  

When to rehost

In a nutshell, rehosting is a viable option for companies that need to migrate as quickly as possible, but are planning to optimize their workloads for the cloud in the future. For example, rehosting is the most popular choice for apps whose lifecycle is coming to an end.

Tips

AWS Application Migration Service allows organizations to streamline rehosting by automatically converting sources to become AWS cloud-native. 

Repurchase

Repurchasing means switching to an alternative software solution, often from an on-premise license to a SaaS model. For example, instead of using a proprietary video recognition system running on-premise, companies can turn to the AWS marketplace and purchase Sensifiai’s cloud-native Automatic Video Recognition API. 

Pros

• Purchasing a new solution is usually less costly than migrating one. 
• Repurchasing is one of the quickest ways to migrate. 
• SaaS models take care of backend operations, allowing businesses to cut maintenance costs. 

Cons

• Switching to a new solution usually calls for additional employee training. 
• Compared to on-premise licensing models, SaaS-based solutions offer limited customizability and control. 

When to repurchase

Repurchasing makes the most sense for applications that are incompatible with the AWS cloud environment or too costly to migrate. In case application customizability is critical for a particular application, it might be better to consider other migration strategies. 

Replatform

Replatforming means optimizing on-premise workloads to run in the cloud while keeping the core architecture as is.

Pros

• Given that replatforming requires only partial modification of code, migration is relatively quick.
• Minimum employee training is required.
• Replatforming allows companies to squeeze the most out of the cloud compared to other migration strategies.

Cons

• It is one of the most error-prone migration strategies as cloud-optimized legacy workloads can still malfunction in the new environment.
• Replatforming calls for extensive and continuous monitoring to minimize the possibility of failures.

When to replatform

Replatforming makes the most sense for companies that need to leverage only some of the cloud-native features. Organizations commonly use replatforming to automatically scale backend workloads and optimize database management (usually with the help of Amazon RDS) by capitalizing on cloud capabilities.

Refactor

Refactoring implies a complete revamp of applications to make them cloud-native. 

Pros

• Realizes full benefits of cloud migration. 
• Enables AWS auto-scaling, allowing services to adjust to server demands in real-time.
• Cloud-native architectures require fewer resources to operate as maintenance is done by AWS. 
• By rearchitecting applications as microservices and distributing databases, companies significantly improve their resilience.
• Overall application performance is dramatically improved. 
• Enables distributed databases. 

Cons

• This is the most resource-intensive migration method, which takes the most amount of time to complete (up to 20 times longer than rehosting or replatforming according to Jonathan Allen, AWS Enterprise Strategist). 
• Provides minimum short-term benefits, with sufficient ROI realized in the long term. 
• Requires extensive planning and often calls for a significant part of the code to be rewritten. 
• Requires employees to learn new cloud skills. 

When to refactor

When it becomes clear that your legacy system architecture has reached its limits in terms of scalability and performance, refactoring should be considered. Refactoring is about ensuring business continuity. Given that cloud providers (AWS in particular) are continuously introducing new features and innovating in the cloud space, refactoring ensures business continuity and agility.