A leading global bank approached Citihub Consulting to advise on improving its end-to-end processes for provisioning servers. The bank had been running a strategic project to deliver a private cloud service for approximately three years; however, feedback from internal customers suggested the interfaces into the service were not sufficiently agile. Cumbersome, manually intensive processes were holding back adoption of the service. Ultimately, the bank sought to reduce time-to-market, improve quality and drive end-to-end automation of its server provisioning processes.
Have a similar challenge?
How Citihub Consulting Helped
Citihub began by defining a vision of how infrastructure provisioning should work, detailing a self-service, API-driven approach. The team then conducted a thorough review of the existing server provisioning process, analysing 12 months of data to identify common ordering patterns – to ensure the new approach adequately covered all use cases. A new process was developed based on a request, approval and fulfilment structure – representing a simplified approach with clearly defined roles and responsibilities.
To support the request function, Citihub defined a set of front-end ordering screens for self-service provisioning. The screens used external cloud services as a reference point to promote simplicity and familiarity, but were extended to accommodate internal complexities. Central infrastructure reference data was established to underpin the new ordering screens and promote standardisation across infrastructure ecosystems. Existing control points were streamlined, resulting in a new set of lightweight approval processes. These approval processes enabled internal customers to control spend and manage sprawl, while allowing infrastructure design teams to challenge requests that deviated from central standards.
To drive the solution, Citihub worked with internal architecture and engineering teams to define a web-service API. An XML API request format was defined, and underlying technologies chosen to orchestrate and execute the automated fulfilment; these covered capacity & workload placement, virtual machine creation & build, network and storage configuration, CMDB integration, software installation and access control. Build and test activities were co-ordinated across a total of 10 internal teams over a 6-month period to deliver the end-to-end solution.