Citi''s virtualization successes to date are based on conquering the hidden challenges of virtualization. We gather and track proper data, demonstrate the cost benefits to the businesses, and are changing the technical culture from a physical world to a virtual one.
We first ensured the accuracy of data by partnering with our capacity planning and asset management teams to gather data needed to identify servers and their relevant utilization data as well as financial data (chargeback costs, etc.). With that data we then built an internal chargeback model for virtualization that displays clear savings for the businesses over physical servers. The model is unique in that it incents the businesses to virtualize more; Businesses pay each month for the physical infrastructure of their virtualization environment, and the more virtual instances they create, the less per instance they pay. The business savings have been 40-60% less than the cost of physical servers, and concessions were implemented to insure the businesses never paid more than the cost of a physical server during the early stages of implementation. We gave custom presentations to each business CIO demonstrating the opportunity for savings. Based on the high savings opportunities, the CIO''s then pushed the message to virtualize down through their organizations.
This virtualization initiative further distinguished itself by proactively reaching out to server and application owners in the businesses, rather than rely on top-down communications. We built a team with an automated workflow process to manage opportunities from start to finish, from standard communication scripts for the initial contact with the server or application owners, to P2V migrations and port migration follow up activities. As a result, we were able to discuss with each business the details of their servers, confirm we could virtualize, and answer any questions or concerns from the server and application owners themselves. Further, we set directives companywide that all new server requests will be planned for virtualization unless otherwise justified and created an automated workflow for the justification process when a physical server is truly needed.
Finally, from a design standpoint, we isolated the network, server and storage components for our virtual environments to avoid mixing them with the physical server infrastructure. This allows us to help manage the cultural changes with physical vs. virtual servers. We created a central operations team that manages all backend virtual host infrastructure and simply replaced the data center team that racks, installs and cables physical servers with a centralized team that does the same function in creating virtual instances and handing them to the server administrators. By not changing any processes for the server administrators as to how they build and manage a physical vs. virtual server, we have avoided unneeded complexity and confusion in the server administration teams.
We have designed our virtualization initiative as a proactive one, working with each business and pushing them to address their low utilization servers. We have ensured consistency across teams through standard workflows and avoided unnecessary complexity while capturing significant savings for the enterprise.