Thinking of Leaping to the Public Cloud? Watch Your Wallet.
So far, we’ve seen that there are three types of cloud services. The public cloud makes services like email available via the Internet, with the customer paying the cloud provider to make its offerings available. The opposite of that approach is the private cloud, where the customer builds his own cloud in-house. And then there are hybrid clouds, which combine the public and private approaches.
One reason people jump to public cloud options, in particular, is because of the promise of cost savings. Rather than buying expensive servers, routers and other machines, you essentially rent time on those devices, sharing resources with other customers and paying just for what you use, rather than the entire machines.
If you find your company succumbing to that sales pitch, watch your wallet.
More than one enterprise has fallen victim to unexpected costs that were – how shall we put this nicely? – not exactly advertised by the leading public cloud provider.
Unfortunately, some customers have found the public cloud’s cost is greater than they expected when they signed on with large public cloud companies. The bills have driven some companies out of business.
‘Resource Contention’ Causes Creeping Costs
Often these cost concerns arise from so-called “resource contention.” That’s when different cloud customers – remember, in the public cloud, you share resources – battle for capacity on a given piece of hardware. If you happen to share a server with somebody who uses a lot of the machine’s resources, the hardware can stop responding or have other problems.
The apparent solution to the problem is to expand the capacity you use, going from the “small” or “medium” instances you purchased to “large” instances. (Here’s the part where you check for your wallet.)
Companies find themselves acquiring additional capacity, which can cost – by orders of magnitude – far more than you thought you’d spend.
In a TechTarget piece earlier this year, IT consulting firm Cascadeo Corp. said the costs can spiral out of control when companies overprovision themselves and then fail to continuously monitor for opportunities to right-size their instances. So-called “elastic block storage” is something of a trap, with costs that look small when you sign up but inevitably creep.
Sean Perry, CIO for Robert Half International, was quoted as saying the company simulated load tests to determine the time need to load a SharePoint instance running in the public cloud. “And all of a sudden, we’re looking at the bill like, ‘Holy smokes!’” Perry said.
Other Sources of Cost Creep
Most of the ways public cloud costs creep are related to storage. Take for example the common problem of an excessive EBS snapshot policy. “Snapshotting” is a method of ensuring no data loss if a single component fails. Some snapshotting is necessary, but too much wastes money.
Another problem stems from temporary project work. Teams provision clusters of compute and storage capacity for a specific period. Startup can be quick. Turn-down, though, can take a while. That wastes money.
And what of research and development? One benefit of developing in a cloud is that you can test against a precise copy of the production system. That capacity you create tends to hang around longer than you think. That wastes money.
Lastly, there’s human error, specifically our tendency to over-provision our needs. That’s only natural because we want to be safe but – you guessed it – that wastes money.
All of these things help contribute to the complicated task of keeping your public cloud costs under control.
Sorting Monthly Bills Can Be Difficult
Indeed, just sorting through and understanding your monthly bill from the dominant public cloud provider can be a difficult undertaking. The vendor’s bills are so complex that the trade press has done at least one article aiming to help customers decipher how much they’re being charged for various services and why.
This same provider has received tough media coverage for the prices it charges for memory, computing, storage and other technology-based cloud services.
To be sure, public and private cloud providers alike appear to be engaged in a price war. That leading provider we mentioned before slashed its prices in the summer of 2013, marking at least the 37th such price cut it has implemented. Whether you call it a price war or the new reality of cloud computing, though, one question remains: How long can leading public cloud providers continue cutting prices before their services suffer?
We will have more to say about how the cost of private and hybrid cloud can be competitive – and even less expensive – than the public cloud. Stay tuned.