Agenda

  • Architecture
  • AWS to AZURE - Thoughts
  • AWS to AZURE - Migration Steps

Pondlogs V2 Architecture

Data Management & ML

AWS To Azure - Thoughts

Given the vast amount of AWS services that we are currently using and most of the developers are already good with these services, it will be quite tough to move away from AWS.

AWS To Azure - Thoughts

As we have already written a good amount of code using the AWS SDK to integrate these services, migration to Azure will require our developers to sweat out during the migration which will impact our priorities & current work plan.

AWS To Azure - Thoughts

In addition to the current services, our team has also spent good amount of time in picking up the knowledge related to the services planned for the future, especially Big Data & ML.

AWS To Azure - Thoughts

Given that AWS is ahead of other service providers, especially when it comes to performance in Serverless Domain, Big Data & ML, we think a full migration to Azure will not be a right choice, at least at this stage.

AWS To Azure - Thoughts

No matter what combination of vendors and products we go for, taking our operations into the cloud itself is a complex technical process, which we have already started long back. But leaving one vendor to get on board with a new provider can be even harder, especially from a technical perspective.

AWS To Azure - Thoughts

Even if movement between Azure and AWS becomes freer commercially, which may not be the case with all the services currently, that doesn’t guarantee parity across application programming interfaces (APIs)

AWS To Azure - Thoughts

Coming to the savings, it would be wise not to compare just the prices at which these services are being offered by the Cloud providers. Instead, the final cost will be dictated primarily by Service Cost & the Developers Time and Efforts. And given the knowledge of our tech team with AWS, migrating away from AWS is not an easy decision

AWS To Azure - Migration Steps

  1. Complete the full product
  2. Compare the AWS services (being used by our products) with those offered by Azure
  3. Analyse the Services Costs
  4. Analyse the Code & Workflow processes
  5. Analyse the migrate costs & efforts
  6. Figure out the timelines & migration process
Given that our new product is still in under heavy development & our huge plans to expand our product features, we are looking at a minimum of 6-8 months for our product to reach the shape that we desire. And once this happens, we can definitely analyse Azure services and see if it is a decision that will add value to our efforts.