Cal Poly State University describes how an On Premise as a Service (OPaaS) approach enabled a new data storage infrastructure with better scalability, manageability and lower cost as it shifted to an IT as a Service (ITaaS) model.
Like many institutions, California Polytechnic State University, a public university with nearly 24,000 faculty, staff and students and a strong engineering and science focus, is shifting to a service provider model for its information technology resources. The goal was for the IT services group that is responsible for the majority of the campus applications and infrastructure to offer Information Technology as a Service (ITaaS) to each of the colleges and departments within the campus.
The IT services group relied on just one large NetApp storage array running mixed workloads for multiple departments. As a 7-year old resource, the NetApp array was coming off warranty and had functional limitations that didn’t fit the IT as a service model, where it needed to scale, be agile and more responsive to changing departmental needs. The multi-departmental demands had recently begun causing an I/O storm that was impacting everyone’s performance. The storage array was also complex and difficult to manage, causing delays in the team’s ability to respond to departmental requests for additional capacity.
The IT Services group knew it needed to make serious changes, but the team wasn’t sure that it had any options.
Learning about On-Premise as a Service
The IT services group learned of a novel concept called On-Premise as a Service – an example of the move toward software-defined data centers, where the infrastructure is virtualized and delivered as a service. The benefits of software-defined approaches have been proven in corporations of every size, but are relatively new in deployments at universities.
In the OPaaS approach, the vendor places the storage hardware and software on the universities premises and behind its firewall – yet the vendor manages the deployment remotely, with minimal assistance from university staff. The university pays for the resource on a per-use basis based only on the storage as used – and out of operating expense (OpEX) budgets, not CapEx.
Buying data storage the typical way – where teams try to predict storage needs for 3 to 5 years in advance, buy significantly more than needed, and pay for it as CapEx –is a process that teams have put up with for decades because there was no other way. When the 3- to 5-year window is up, organizations are forced to rip and replace the infrastructure and do it all again at significant hassle and expense.