The question for modern organizations is no longer “Do we move to cloud?”; instead it’s “When and how do we move to cloud?” Some IT departments get top down imperatives to move everything to the cloud, while others seize the opportunities offered by impending license renewals or the possible loss of vendor support. Regardless of the reason, critical decisions must be made to ensure a successful transition. In this second part of our cloud migration series, we’ll cover the “Brand New Start” approach to cloud migration.
Brand New Start in the Cloud
Creating a data architecture has never been cheaper, faster, or easier. Even though new systems can be started in the cloud, they can still integrate with your existing systems through external connections.
Cloud Migration Roadmap
Every successful cloud migration begins with the same two steps. Invest the proper diligence in these decisions and you will maximize the value of your transition.
Determine Who’s Going and What to Pack
One of the first steps in any trip is to identify who’s going and determine what they need to pack; successful migrations are similar. What systems are migrating to the cloud? What aspects need to move? How soon do they need to get there? What baggage should they bring with them? Work with the business stakeholders of those systems to establish priorities for moving them to the cloud. Think about what they need to bring with them, especially with regard to data. At this early stage it is important to define the current state for business processes and consider your plan for migrating data or ensuring seamless interoperability between cloud data and legacy on-premise storage.
Pick the Appropriate Vehicle and Route
Once you have a destination in mind, you can pick a vehicle and plan your route. Stakeholder priorities, objectives and concerns can help you determine which cloud vendor to use and whether a public or private cloud (or both) is most appropriate. The major cloud providers all have similar feature-sets, so picking a target environment is usually a combination of price (each vendor provides a tool to gauge pricing), matching technology stacks with IT expertise and leveraging existing vendor relationships.
The more difficult—and more important—choice is choosing a method for the transition. You have three roads to choose from on your journey; this is the second option.
Steps to Migrate:
-
- Gather and prioritize a backlog of requirements; include any refactoring or new features that are desirable to include with the migration.
- Work with stakeholders to determine a data migration plan.
- Include a data quality initiative so that you can verify that data is moved correctly.
- Start cleaning the data in your source systems, or at least ascertaining the kind of work required.
- Identify all the teams using the data, especially those accessing it for analytical purposes (data warehousing, reporting, etc.), and gather their requirements so that their connections to the new infrastructure will work correctly.
- Identify the best ways to redesign and migrate functionality of the existing system. This includes, but is not limited to:
- Using microservices to replace existing features that need to be highly available.
- Replacing current (and still relevant) software and custom features with open source software or SaaS offerings.
- Using Functions as a Service (FaaS) to handle regularly recurring and triggered events (e.g. logging, moving files, backups, etc.).
- Implementing automation to provide rapid feedback and validation where possible (i.e. CI/CD, Testing, etc.).
- Test your data migration approach regularly in parallel with development. This will ensure you have a solid batch of production test data to develop against, test the migration strategy on a regular basis, and help confirm that the necessary architecture is in place.
- Conduct regular check-ins with stakeholders to ensure that their vision and needs are being met. Show regular progress and use it to help refine their understanding of the requirements; this will minimize rework.
- Determine a cutover plan and prepare for final deployment, data migration, and retirement of the old systems.
- Deploy and enjoy.
When to Use This Method
Brand New Start works best under the following circumstances:
-
- The organization has personnel experienced in creating balanced, well-designed architectures.
- There is a set of clear infrastructure principles and a long-term commitment to implementing them.
- A strategic vision exists to align infrastructure investments with business goals.
- The systems being migrated are relatively small and there is no urgent need to move them.
- Parts of the legacy system are very old and cannot be recreated on modern infrastructure.
Pitfalls to Watch out for:
-
- Don’t commit to unorthodox and untested architectural designs without thorough research.
- Avoid solutions that may adversely limit future extensibility.
- Plan and coordinate with stakeholders; keep them well-informed, especially about costs.
- Limit the number of options for architecture; otherwise, you can be mired in “analysis paralysis”.
- Stay focused; scope creep can extend the rewrite indefinitely.
- Invest for success; trying to maintain the current system on a skeleton team can cause them to distract the rewrite and put its success at risk.
Conclusion
Moving to the cloud is inevitable for most organizations and there are many strategies for the transition. Understanding your current system architecture, reviewing the available cloud services and assessing the pros and cons of each strategy can save you time and money. Whether you choose to replicate your systems directly in the cloud, rewrite an entirely new cloud-native system or revise the architecture over time, make sure to do the research before choosing a strategy.
Check out other installments in the “Three Approaches to Cloud Migration” blog series: