In the previous blogs and whitepaper, we discussed the reasons that an eventual move to cloud is inevitable and also helped explain some of the terms used in cloud. In this blog we’ll explain more about how you setup for success in migration, what you need to prepare and some of the decisions you’ll need to take.
The first and most important point is to take action and create a roadmap. The migration roadmap has several key components.
The most crucial and strategic pre-migration activity. There are multiple conversations that need to be facilitated including data stewards, data scientists and IT leadership. Create a cross-functional decision-making committee. The key phrase is decision making, the committee must be empowered to make and take action on decisions.
Supplement the team with functional and technical specialists to jump start the process and ensure gaps and vulnerabilities are addressed up front.
The cloud committee must be able to define success. What does success look for the first deliverables? For example, it might be proving that you can have a cloud footprint by taking some existing workloads online, it might be showing that data science can do things they’ve never been able to. Defining success helps with the next step, technical approach.
Are you planning to rework existing workloads or use cloud compute processing to run existing loads more effectively:
• The Lift-and-shift method takes existing on-premises workloads and migrates them to the cloud “as-is,” with minimal or no changes to the application architecture or other mechanisms.
• The Re-platform method calls for redesigning on-premises workloads to a cloud-optimsed architecture to fully leverage the native capabilities of the cloud environment.
The decisions here will help with IaaS vs PaaS discussions. Spoiler – use IaaS for lift and shift and PaaS for re-platform.
A readiness assessment process will reveal applications and datasets that need to be rehosted, re-structured, or even retired. Remember reduce, re-use and recycle from the previous blog?
The output of the assessment should be a plan for time bound migration, what to migrate, what to leave alone, the data prep, the bill of quantities and more than likely a revised set of priorities and scope.
At some point during the technical approach and readiness you are likely to identify a skills gap. Even if you don’t think you’ve got a skill gap its worth doing a skills assessment to check you have key skills covered, for example cloud certified architects, security and privacy experts and data integration specialists.
A critical area not to overlook is the skills assessment of end users. In the context of marketing data specifically, end-user training becomes crucial to enable marketers and business analysts to access data from cloud environments easily and independently.
Make sure that all relevant datasets are available during the migration. Projects that see delays in data availability or run into hurdles to secure access can lead to slippage in migration project plans.
You’ve now done the prep work, you’ve got the team, the data, the infrastructure patterns so it’s time to migrate the first workloads.
Start with a small proof-of-concept / even smaller pilot runs on limited business functionality or test data sets in a beta environment. Not only will this help validate future state architecture, but it will help simulate the impact of the migration at scale and gather information on dependencies that may impact outcomes.
Now that you’ve got the first cloud workloads running then use the same principles for the next use case. Some point that helps drive this are:
Build a cloud centre of excellence (CoE) consisting of professionals experienced in cloud migration. This is an investment that can encourage people to skill-up and earn cloud certifications and accreditations. Your CoE should ideally be led by a key senior leader to champion cloud.
Based on the success of your first cloud migration, cloud more strategically and holistically for adoption in wider business functions, such as enterprise IT.
Re-examine current processes and policies to assess what needs to change to streamline cloud projects. For example, agile development methodology with continuous integration/continuous deployment models can be the most productive and efficient. However, executing such a methodology requires significant changes to legacy policies and procedures.
In summary, moving to the cloud for marketing is inevitable and by following the steps above it helps move existing workloads to the cloud and set up the required components for further success.