I have a use case to migrate 200+ EC2 VMs (all similar machines running WP, OS is Debian) from AWS to GCP and evaluating the below 2 options from both a cost perspective and complexity in replatform
1. Rehost:- Lift and Shift to GCE
2. Replatform:- Use Migrate to containers service to setup them up as containers in GCP with Anthos/GKE. Want to hear pros and cons in terms of costs and migration complexity.
Solved! Go to Solution.
Hello @dheerajpanyam ,Welcome on Google Cloud Community.
===Rehost==
Always lift&shift will be faster solutions. More about VM migration here: https://cloud.google.com/migrate/virtual-machines/docs/5.0/migrate/create-an-aws-source.
Cons:
- Fast migration
- Easy migration ( step-by-step tutorial )
- Easy management ( Managing GCE is similar to EC2 on AWS, so you don't have to deal with upskilling people )
Pros:
- Basically you are not improving your infra thus application will also stay as it is. More work with improving in feature
In term of costs, use Pricing Calculator : https://cloud.google.com/products/calculator?hl=en
==Replatform==
Cons:
- You can utilize GKE Enterprise to manage you fleet ( paid option ).
- You can utilize autoscaling, so add/remove VM depends on load
- It's move and improve so basically more advantages in terms of migration
Pros:
- If you don't have skilled GKE / K8S engineers, it might be more challenging use case, as in fact you will migrate load ( because of step-by-step instructions) but managing clusters fleet will be more challenging that with standard VMs
- Depends which application is on your EC2, you must prepare target VM for handle this application.
Migrate to container service is nice if you have EC2 split as microservices ( 4 VMs for SQL, 5VMs for tomcat etc) because it would be better in logical way to migrate app by app. However the complexity of this approach is much more high that simple VM migration
Basically imho approach here depends on AWS EC2 complexity. If you have 200+ EC2 VMs, i think that better would be move them with rehost approach as it might be possible that you will forget something regarding configuration in replatform option thus troubleshooting will be more challenging.
3rd option :
Migrate load as it is, and then plan improving GCE load with GKE / MIGs approach ( split improvements with waves ). It will ensure that you will have more control over migration and improvement as well as workload work continuity will be achieved.
--
cheers,
DamianS
LinkedIn medium.com Cloudskillsboost