In 2013, NetApp IT created its first Infrastructure as a Service (IaaS) portal that allowed a small group of superusers to request virtual machines and storage from a service catalog. Within minutes, a fully running infrastructure landscape for compute and storage would be provided, irrespective of the cloud destination. It was our first foray into IaaS, built on automation and a cloud decision framework.
Referred to as nCloud, our first IaaS offering predetermined the cloud placement of the workload ― either public or private ― according to answers to an online service catalog request. Several factors determined cloud placement behind the scenes such as P1 eligibility, SOX compliance, and customer accessibility, as well as how critical the application was. It all sounded great at the time, but it was complex and required extra governance by the catalog team to maintain and automate.
The first-generation IaaS offering was also limited to superusers, a small group of technically sophisticated folks who represented broader user communities. They were knowledgeable about infrastructure and the costs of public cloud versus private cloud selections. Although these superusers appreciated the speed of spinning up compute and storage, they were not allowed much flexibility in terms of post-provision, “day two” operations such as mounting and unmounting volumes, adding volumes, and resizing. For these kinds of operations, they typically had to engage IT Operations directly.
Matt Norton is a senior system administrator for Cloud and Compute Services at NetApp. Matt designs, implements, and supports infrastructure services for the NetApp global enterprise and creates user interfaces for the internal CloudOne platform that includes DevOps, IaaS, PaaS, DBaaS, and CaaS.