The Case below happens on a large private university that optimized its laboratories, reduced operating costs and delayed infrastructure investments.
In the previous environment, labs uses for classes were mounted on multiple server groups, each group attending a single class. Teachers passed the lab requirements to the operation team so they could set up the environments. A considerable amount of computer resources and operation team man-hours were consumed in the process.
The process of mounting each environment was time consuming and teachers had to request the set up with great advance. After being used, the laboratory was dismantled and computing resource released. If the same environment had to be mounted again, the same effort was implemented.
To solve these problems, the university merged all laboratories in a single pool of computing resources and used OpenStack to virtualize their network and compute structures. This allowed the Labs to be made set up on a self-service manner, through a web portal.
With an pool or resources meeting all lab demands centrally, it became possible to manage resources according to lab requirements, rationalizing the use of infrastructure resources and avoiding the need of keeping a fixed computational pool.
The lab environments are now structured virtually, with all compute, storage and network being mounted within an virtualization environment. In addition to the agility, this model leveraged the support of a greater number of classes per day than the old model.
The virtual environment also allowed the structuring of the labs in script format. This way, each laboratory structure could be stored for future use, allowing the operating staff to concentrate on more noble activities, such as the design of new laboratories. Not to mention the increased on team motivation to be working on more creative activities.
Teachers benefited from the possibility of assembling laboratory through a web interface, no longer needing the operating team support.
But the university ended up being the big winner with the solution. Due to the more rational use of the hardware resources, the university postponed the acquisition of new servers for two years. The reduction in operating staff workload allowed the university to relocate some professionals to DevOps team.
The university’s next step is to unify labs computer and internal systems computing pools, further enhancing the rational use of its hardware. This will allow, for example, the enrollment of new students system, which has its highest peak on vacation, can scale up using lab resources pool, which is not used in this period.