Running servers costs money. But what is the cost of not running them? In this blog post, we dive into the fascinating world of speculation, guided by studies to keep us on track. Join us on this exploration on what the cost of poor capacity planning can entail!
How much capacity do we need to deliver our service or cloud application? The true number is of course related to the amount of users we wish to support, constituting our capacity demand. How much we actually have is our capacity availability. Using smart tools, such as the Elastisys Cloud Platform, we can keep these numbers in synch using e.g. predictive analytics. But we are getting ahead of ourselves a bit. For now, let’s just define two terms:
- under-provisioning is when our capacity demand is higher than our capacity availability (we don’t have enough servers), and
- over-provisioning is when our capacity availability is higher than our capacity demand (we have too many servers).
Over-provisioning is often incorrectly thought of as a good way to ensure that we do not run into under-provisioning. As we have blogged about, the only effect one achieves is a false sense of capacity sufficiency (oh, and a higher than needed usage bill). The effects of under-provisioning are a bit harder to define in pure financial terms, but that just makes them more interesting.
The Costs of Poor Capacity Planning and Under-provisioning
So we have too few servers, what’s the big deal? Well, it depends.
- Computer users feel frustrated with slow sites,
- customers spend less money at slow retailer sites,
- lose substantially on costs related to downtime for business-critical services, and
- Google also punishes slow-to-respond web sites in their page ranking algorithms.
In short, the costs of under-provisioning are hard to quantify, but dangerous to under-estimate. As we have blogged about before, merely trying to keep a capacity safety margin relative to current demand is not sufficient.
That is where Elastisys Cloud Platform comes in. It continuously monitors your application, predictively determines the correct capacity, and pro-actively scales your deployment. This ensures you that your cloud application has the capacity it needs, even in the face of unforeseeable events. Use it to make sure that your viral opportunities turn into a success stories, not a missed opportunity to reach a new audience. It’s 2015, we should all be above that now.
How do you deal with capacity changes? Do you have additional insight into the cost associated with poor capacity planning? Let us know in the comments below!