Don't migrate apps to the cloud without considering the costs
For many organizations, the public cloud is a way to cut costs -- but some apps may be costing you more in the cloud than you think. Find out which enterprise apps are better left on premises.
Many organizations tout cost efficiency as one of the main benefits of cloud computing. In reality, though, roughly 30% to 40% of enterprise apps cost organizations more in the public cloud than they would on existing, on-premises systems. This is because many enterprises fail to identify which applications will provide the biggest bang for their buck after a move to the cloud.
Before you migrate apps to cloud, there are a number of characteristics to evaluate -- ranging from how software was written to its security requirements -- to determine its cost in the public cloud.
Let's say an enterprise has 2,000 applications under management, and it moved 100 applications to the public cloud this year. Those 100 applications cost $1.5 million to operate on premises per year. When relocated to a public cloud provider, they cost $2 million per year. What went wrong?
The 100 applications likely had attributes that increased costs in the cloud. These can be simple things, such as poorly rewritten applications that eat too many compute cycles or storage resources, and thus draw a larger bill from the cloud provider.
Alternatively, some applications perform poorly in the cloud because the cloud platform moves the application too far away from the data it uses. This causes user productivity to fall and costs to rise.
Three warning signs you shouldn't migrate apps to cloud
The best way to understand if your application uses too many resources is to see what the code tells you.
In general, there are three types of applications, or application requirements, that signal your apps will cost you more to run in the public cloud than they would on premises:
Applications that are poorly built and designed.Moving to the cloud will only highlight the fact that these applications consume too many resources, and thus will generate a larger cloud bill than expected. For instance, some applications make a large number of I/O requests. They read and write to storage systems too often, and, as a result, put stress on the network and storage systems. In the cloud, you have to pay for this kind of inefficiency.
The best way to understand if your application uses too many resources is to see what the code tells you. Good developers understand how to use resources effectively, and spot such issues right away. Another way is to use code analyzers to understand when and where inefficiency exists. If you don't have access to the source code, you can use an application profiler, which monitors application behavior and reports issues, such as too many I/O requests.
When an application consumes an excessive amount of resources, the only path to success is to refactor or rewrite the application to make the most out of the native cloud platform. That, however, adds risk and money.
Applications that are spread too far from the data.Organizations love to keep data on premises, and leverage applications in the cloud, or vice versa. However, network latency needs to be considered, and the cost of poor performance.
Applications that have especially strict security and compliance requirements.Depending upon the industry, moving some types of workloads or data to a public cloud requires creative security and compliance processes and technology. As such, you could end up spending hundreds of thousands of dollars, or more, on these changes, and increase your costs when you migrate apps to cloud.
App developer employment and cloud computing trends
Cloud computing efforts span a spectrum from all-in to indifferent. Take this quiz to see how your organization measures up in cloud computing trends.
For instance, a healthcare application may have a regulation applicable in some countries, where data must be encrypted in a specific way when hosted in a public cloud. As a result, organizations would need to change the application to use that encryption model, and that may introduce performance issues, as well as higher operational costs. Other applications require certain audit information be logged, such as when data moves from on premises to the cloud, or back. Again, it costs a great deal of money to change and redeploy those applications in the cloud.
Anytime we deal with the challenges around platform-to-platform migration -- including on premises to the cloud -- the issues are pretty much the same. Rewriting applications for the new platform, ensuring efficient use of resources and security should be top-of-mind for IT teams.
Take a realistic look at what your new platform can provide, and select the workloads that will offer the most value before you migrate apps to cloud. If you do that, you'll find that the cost savings are there -- if not, you'll face higher costs and, in many cases, confused executives.
