When Microsoft IT began its cloud migration journey in 2009, it followed a similar process. First, it
cataloged its operating system instances and application workloads. This assessment included both
quantitative data that was mostly retrievable by tools as well as qualitative data that was partially
retrievable by tools and also required examination by both the operations team and the business
liaison team. This latter category of metadata included relationships, dependencies, and integration
Microsoft IT then identified the initial prioritized eligible operating system instances and initial
prioritized eligible workload/applications. Next, these initial migration candidates were reduced by
removing any business-critical systems, which would be moved after more experience had been
gained. Then, this initial list of candidates was prioritized and sequenced with less-complex
applications placed before more-complex applications, and applications running on updated VMs prioritized over those running on physical machines or legacy VMs. Some applications were identified as ineligible for various reasons (most of which limitations no longer exist in 2015) and these were migrated to an optimized on-premises datacenter.
After the initial set of migrations was completed, Microsoft IT completed work to make less-eligible operating system instances (OSIs) and workloads more eligible. For example, OSIs with older
operating systems or database versions were updated, more applications on physical machines were
moved to VMs, and more mission-critical applications were deemed eligible. Applications and
workloads that were identified as requiring a major overhaul were rebuilt as services on Azure.
Here’s the process:
1. Identify eligible hardware (OSIs) per Azure compute, storage, and RAM limits.
2. Identify eligible applications, remove HBI apps, sequence critical and complex apps for later, and
right-size to include more apps.
3. Increase eligible hardware and applications by doing the following:
Virtualizing more servers
Expanding to more regions
Including extranet-facing apps
Including HBI apps
Getting current (OS, SQL)
Increasing Azure VM limits
4. Build new applications as services for SaaS IT.
Source of Information : Microsoft Enterprise Cloud Strategy