At ViFX we are in a very fortunate position to have deep engagement with nationally significant consumers of public cloud resources. This deep engagement facilitates great insight into what actually matters when an enterprise or Independent Software Vendor moves to significant public cloud consumption.
Many of our customers moved to the cloud to reap the benefits promised – low touch, pay as you go, potential cost savings by scaling automatically, etc. What they actually experience can be not only those benefits, but also a few unanticipated side-effects such as usage sprawl and considerable bill shock.
Why is that?
In the good old days of the late 00’s, VMware had revolutionised the data centre. Every organisation worth its salt was either virtualised or in the process of it. The massive cost benefit was unbelievable, as well as the agility it facilitated. One of the problems back then however was VM sprawl. VM sprawl is defined as rapid creation of virtual machines, and therefore not only the licensing and operational cost issues that went with it, but also rapid consumption of the platform capacity.
Roll forward seven years and sprawl is an even bigger issue. If an organisation doesn’t have automated (or disciplined) IT service management and provisioning in place, sprawl is a very real issue.
One of the greatest benefits of cloud is the degree of agility facilitated by infinitely and instantly available resources. Now it’s possible to deploy whole systems from templates in just minutes. What this means is that now it’s possible to incur hundreds, if not thousands of dollars a month in cost, at the click of a few buttons by deploying whole systems in just minutes. See what I mean?
In public cloud there are no sunk costs. The concept of spare capacity is absent, so sprawl immediately incurs cost. Loose or no IT lifecycle management can often mean that no-one knows if an object (VM, storage, cloud service, etc.) is actually still needed, so no-one turns it off.
Old-school operational processes don’t work
A regular modus-operandi of on-premise IT admins is to work around application performance issues by throwing horsepower at the problem. For example, an app owner says their app is running slow and has identified it is using high memory. The diligent IT admin looks at the VMware layer and sees that although the VM is using high memory, it’s actually actively using far less, so she declines the request to add more memory, and asks the app owner to look harder. Alternatively (or if the app owner insists), she might just give the VM more memory as she doesn’t need to fill out a PO request - it’s a sunk cost and there for use.
Move that scenario to the public cloud and the IT admin is unable to determine if the VM actually needs more memory using traditional tools. So unless she has Application Performance Management (APM) tooling experience (think New Relic) she has to give more memory to the VM, and voila the monthly bill just went up from now until the end of life of that VM.
People may need to change
As described above, organisations that intend to utilise public cloud at scale need to prepare their IT teams to be successful. There will be many in the IT team that don’t want to re-skill. There may still be high demand for their legacy skills so they will probably move on. IT team members that choose to stay will have to become APM experts if they are to remain valuable. As mentioned in a previous blog, IT infrastructure folk will have to become specialist scripters, and possibly coders, if they are to remain relevant.
Many people won’t make the leap, and this is a significant and tangible risk to an organisations IT transformation.
The reality of public cloud
The promise of cloud is absolutely tangible, and based on business need is most likely useful and more cost effective for many aspects of most organisations IT. Existing IT teams won’t have the capability to ensure on-premise Service Levels, so enabling IT teams to be successful could mean moving some people on, re-skilling some, or bringing in a 3rd party to help your organisation “cross the chasm” from legacy IT.
ViFX is a specialist provider of Managed Services for both on-premise and public cloud infrastructure. Talk to us today about how we can assist you to bridge the gap between legacy and modern IT infrastructure management.