Simon Burbidge and a team of 1E consultants developed some unique techniques that were used in a Windows XP to Windows 7 operating system deployment (OSD) project that achieved what many would agree is amazing success. The project was for a regional United States bank on a tight deadline with 90,000 clients, many in the bank’s approximately 3700 branches. This was indeed a challenging environment. However the deployment achieved an operating system migration rate of 5,000 computers per week, 3,000 of which were at the branches. Those counts are successful builds that were fully functional in terms of applications installed, with 41,000 happy users already migrated.
Implementing an OSD project can be a bumpy ride, especially considering the plethora of requirements in terms of location, build types and the way organizations like to operate their systems management environment. One of the requirements in this case was for the ability to deploy the majority of build types at all locations. By build type we mean the application profile that identifies the function of the machine within the company. Considering the number of types and the number and dispersal of the locations to consider, utilizing T1 network links or worse, it was clear this really was going to be a challenge.
How did they measure OSD success? The core questions are:
- What are the numbers? Are you migrating clients at a rate sufficient to meet the project deadline?
- Are you attaining the desired rate without ever compromising branch operations?
- Once a user’s workstation has been migrated, do they always have the applications installed that they need to carry-out their function within the business?
Clearly in this case success was impressively achieved. The numbers and rate are large for any organization. The techniques used ensured that branch operations were not compromised and the workstations had the applications required. The unique techniques the consultants developed include:
- How to automatically confirm at deployment time that all the critical content needed is already available locally (precached). Missing applications would mean that for some users the migration was not a success. For this reason it is critical that in engineering the ConfigMgr deployment task sequence, we ensure that once a commitment is made to install the new image, we have every confidence that the deployment will be successful. Non-critical content might not be available locally but must be managed to ensure branch integrity while offering the flexibility of allowing OS migration to proceed where appropriate.
- How to determine which applications are needed by the users and how to provide alternatives in cases where the user needs allow less expensive solutions.
- If there are application installation issues, how to ensure that application installations are a success during the migration by initiating installation retries and closely monitoring the process.
Those techniques are explained in great detail in a new whitepaper that 1E is pleased to offer. Please see this link to download your copy.