Marginal gains theory is again in focus following the Rio Olympics. In sport where the margins between glory and disappointment are so slender, coaches look for opportunities to optimize athletes natural ability through small cumulative changes.
For example in preparation for the 2016 games when Team GB coaches noticed that women track cyclists sometimes missed training due to saddle sores, experts were consulted, the athletes were given advice on waxing and shaving of pubic hair, and no further training days were lost due to saddle soars.
In another example athletes sleep patterns were analyzed, new beds and the use of hypo-allergenic sheets were brought into the boxing training camp, resulting in an additional 24 minutes sleep per athlete.
The hockey team, who won gold in a penalty shootout, had for the previous year been put through a weekly exercise that consisted of requiring athletes to make decisions in an extremely fatigued state.
So what does this mindset of analyzing details to obtain a slight gain have to do with technology?
I recently sat through a partner webcast on VMware Optimization Assessment, one comment leapt out at me, “if customers realized the savings they could make within there own environment they wouldn’t consider going to the cloud”.
That is a message that needs to be heard loud and clear. Cloud providers such as AWS state “you pay only for the services you need…” go crazy and request stuff you don’t need and your OPEX will soar. Administrators are encouraged to think in terms of cost and cost optimization, not just SLAs.
With a cloud provider there is a clear requirement to manage what you consume, they provide tools to track and monitor resources, it’s the customer’s responsibility to stay on top of your how much they are spending. Get careless in the cloud and it could cost you your job when the bill arrives at the end of the month.
Don’t get me wrong I admire the Amazon approach, one of the simple steps they recommend to control your spend, is Right-size your services to meet capacity needs… That’s great, but why don’t we just do that with the on-premises solution we have!
VMware are offering Enhanced Evaluation of vSphere Optimization Assessment (VOA), in other words sign up for the trial, deploy and wait for the reminders to run the reports. The capacity report after 30 days is going to give lists of clusters and vms that are underutilized, idle and powered off.
If you hear the counter argument that we have already paid for the on-premises equipment, so there is no savings, start thinking like a cloud admin. Management are probably already wondering if they can save money by going to the public cloud, so find a way to put those savings into financial terms, hey I can increase the capacity by 5%, and reduce our tariff to other departments etc.. Or we can put an additional $xxx worth of vms into the existing infrastructure with no extra cost.
One of the big moans I hear against vSphere is that it’s expensive, maybe it’s expensive because it’s easy to set-up and let the sprawl begin, whereas cloud providers hold you directly responsible for usage.
If you are going to be at VMWorld check out the sessions and workshops for VOA, such as;
Generate Revenue with vSphere Optimization Assessment (VOA) for Service Providers [HBC7661-QT]
Be the IT Hero of Your Company and Discover How to Save Thousands of Dollars by Reclaiming Underutilized CPU, Memory, and Disk Resources [MGT8751-QT] US Only
The Million Dollar VOA [PAR3710] EU / [PAR9795] US
How to Successfully Sell Intelligent Operations – & How the vSphere Optimization Assessment can help [PAR3726]
Generate Revenue with vSphere Optimization Assessment (VOA) for Service Providers. [PAR3829]
vSphere Optimization Assessment Workshop [EPW-1701-USE-1]