Let us follow the trail of positive disruption and move along to a consumption ‘as a Service’ model
The first step is to move our virtualised workloads and longer term storage (backups and such) into Infrastructure as a Service (IaaS).
Unmanaged IaaS is commodity compute and storage, meaning you pay for what you use as you use it. Most IaaS vendors standardize on a single hypervisor on compute and storage shared across customers with virtual data centres created per customers. Security separation is provided at the hypervisor configuration level. The hypervisor and hardware is less important than the cost to SLA analysis.
IaaS Cloud Providers should be compared based on price, feature, SLAs, and availability of services in the customer region as necessary.
For example, AWS S3 is ~$0.03USD per GB, for the 1st TB per month and has 99.99% availability and 99.99999999% durability of objects stored in S3. *Note: from an early price check
This price and availability can be compared to other public hyperscale IaaS providers, local IaaS providers, and running an on-premise IaaS to look for a ‘best fit’ option.
The ultimate IaaS cloud can be a public, private and/or hybrid cloud as necessary. Often the hard applications ‘20%’, that were identified in the virtualisation efforts in preparation to move, forces this decision into a hybrid Cloud model - mixing public IaaS and on-premise Cloud services until the hard applications can be decommissioned or upgraded to a point where they can be virtualised on to the public Cloud.
At this stage, it is all about moving expensive long-term storage to commodity storage and virtualised compute and decommissioning as much of the owned/leased hardware and data centres as possible. This is a race within change, control as much as can be needs to be moved across rapidly to achieve results and desired outcomes.
The end result becomes a commodity service that is supported by the IaaS vendor. This allows the organisation to focus on the applications and using the Cloud to the best of their ability.
The Cloud Provider can provide storage migration methods. Dev/Test can be moved immediately, the existing infrastructure should be frozen, except for bug fixes and BAU tasks, and all development can be done on the Cloud Services providing a better go forward with the ability to turn development services off when no one is working on them.
This will result in a bump in operating costs initially until the organisation understands how to manage their usage.
This is not an uncommon problem when moving from a Capital Expense model to an Operating Expense model for any goods or functions.