Most cloud computing users evolve into their cloud strategy rather than build one from scratch. In many cases,...
the cloud is the result of shadow IT within line departments -- individual business units spinning up cloud VMs without the IT department's knowledge or consent. And even when IT professionals formally plan for cloud, it's usually done on a per-application basis without full consideration of costs and risks.
Fortunately, there are many ways an organization can optimize cloud, in terms of performance and cost. To start the optimization process, follow these four steps: Strike a balance between cost and performance; shed any unnecessary applications; know where your users are; and negotiate rather than accept published cloud prices.
Strike a cost-performance balance
Nearly every cloud provider offers a baseline service set. This is the cloud service that the provider represents in its advertised cloud pricing. Even though it's likely the least expensive, most users don't end up with that service. They may pay for premium hosting, elasticity, geographic diversity, high availability and other special features. But while these add-ons can be valuable and cost-effective in the long run, they are often just money pits. For some users, these special features often double their cloud costs.
To validate your cloud feature choices, perform a pilot test of your applications using the provider's standard cloud service. Rather than test everything you run, focus on the application that's responsible for the majority of your cloud usage and costs. To determine the features you need, run the application test for one to three months. Measure the quality of experience (QoE), downtime and application costs in both your test and production environment -- you may be surprised at the results.
When a premium feature's cost is greater than its value, educate cloud planners or business units to avoid this problem with future cloud applications. Make your test framework available as a guideline to cloud service buyers in your organization.
Know which apps are right for cloud
Another cloud optimization trick is to examine the applications you're running. Public cloud services are often less expensive than on-premises IT when applications don't fully use data center or local resources. However, an application that generates a lot of traffic requires a lot of cloud storage and demands regular in-cloud maintenance, which is probably not saving any money in the cloud.
Determine the cost and IT effort associated with each cloud application. Identify the apps that are most expensive to run, and those that require the most work to maintain. Compare user QoE before and after the cloud transition. It's likely that at least a fifth of your apps are costing you more in the cloud. And if premium features aren't causing those costs, pull these apps back into the data center or look for another cloud approach. Try platform as a service or software as a service if internal support costs for infrastructure as a service are too high.
Consider your users' location
Understanding the cloud-to-user connection is critical for cloud optimization. Organizations with the most successful cloud applications ensure users can access those apps across a large geographical footprint. Mobile workers are also good candidates for cloud. But if your users reside in a single office -- or only a few -- the distribution of cloud applications may provide little value. In this case, it would be less costly to host your applications internally.
If your user base spans multiple time zones, or is very distributed, establish hosting points that are local to users or that move across time zones throughout the workday. This improves performance and helps manage costs.
Take a seat at the negotiating table
Negotiations help with cloud optimization. And it's not just a matter of beating up prospective cloud providers on price. Organizations must understand their cloud applications' needs, and get the best services to meet those needs.
A pilot test involving a cloud provider is the best tool for cloud negotiations. Test either your most demanding and expensive application, or a mix of applications. This establishes a baseline for cloud resource usage and user QoE that will then fit into service level agreements (SLAs). Make this the starting point of your discussion with the cloud provider.
There are three primary points to consider when negotiating cloud prices: the provider's guarantees in the SLA, the base price of the cloud service during the service period and the price of incremental services on top of the basic service model.
During negotiations, remember three important points. One is that the provider will adjust prices pro rata, based on the prevailing current price of the cloud services you use -- even within the contract period. Second, you can renegotiate a contract if a named competitor offers a price that's more than 10% below your current provider's price, and your provider doesn't match it. Third, you can renegotiate a contract if the SLA is not met more than twice in any given month.
These points are more important than the strike price for the initial cloud deal, because they offer you the opportunity to get a better price -- or even rethink your choice of provider. With contractual leverage, you can ensure your provider gives you the best cloud service at the best price.
About the author:
Tom Nolle is president of CIMI Corp., a strategic consulting firm specializing in telecommunications and data communications since 1982.
Create an optimization plan to save your budget
Reduce cloud applications costsin three steps
Who should create yourcloud budget?
How the cloud price war is impacting customers
Utilize Amazon Google when building a private cloud