Working with Enterprise ALM: Mobile, cloud and more
A comprehensive collection of articles, videos and more, hand-picked by our editors
Editor's Note: In this first of two parts, contributor Tom Nolle discusses the business side of migrating applications to the cloud. In Part 2, he explores technical considerations for cloud application migrations.
In the early days of cloud computing, senior business leaders were often captivated by the potential to shift all their IT operations to the cloud to achieve major cost reductions. More recently, nearly all have realized that many applications don't belong in the cloud; in fact, many will actually cost more in the cloud than in-house.
So the challenge lies in figuring out what application characteristics best fit the cloud business model. Application architects can help with that process.
Successful cloud migrations requires considering three crucial factors:
- The unit value of labor of workers impacted;
- The economy of resource (not economy of scale) gains possible; and
- The value of agility in application-to-resource mapping.
Bottom line: If there's no benefit in these areas, the application isn't a migration candidate.
Understanding the unit values of labor
IT processes benefit business -- and justify investments made in them -- based on the productivity gains they can bring. It stands to reason that if a given application saves workers' time, the value of the time saved will be determined by the cost of labor -- what industrial analysts call "unit value of labor" for the workers. High unit values of labor mean that productivity gains return more on investment.
While that's a well-known principle, there's a little-known corollary: An application's cost is proportional to the number of workers it has to support. When you combine these two points, you get an important principle of IT empowerment: The best applications to target for anything are those supporting a modest number of workers with high unit labor values.
Application architects can easily determine the number of workers supported and the general job classifications involved. With this information, it's easy to create a profile of the application relative to the company overall. Anytime an application targets workers whose unit value of labor is lower than the top two quintiles -- that is, the top 40% -- of the worker population, it's likely that the application will have minimal return on investment (ROI) potential, making it hard to justify any significant cost in migrating to the cloud. Widespread use of the application will also magnify cloud risks by increasing the impact of transition problems or ongoing differences in quality of experience (QoE).
In contrast, applications targeting the top quintile of workers can tolerate more cost because of the higher productivity value of the workers involved. Give applications in the bottom two quintiles a big minus and those in the top quintile a big plus!
Calculating cloud-migration cost savings
The next business consideration for cloud migration is the actual savings potential. In most cases, cost savings from cloud computing are related to presumptions of economy of scale. The general assumption is that a cloud provider has a bigger data center and thus can run it more efficiently, in both capital and operations terms. In fact, that's probably untrue for large enterprises; their own data centers will approach cloud efficiency levels.
What's important here is the application's use of resources. Server data can provide a measure of the resource use, and thus identify applications that are underutilizing their server resources. Those applications are prime candidates for cloud migration.
To identify underutilization, look for servers that are running a single application, an application that's rarely used or an application only a few people use. A quick check of server resource statistics will establish whether these applications are, in fact, underutilizing their hosting resources. At the same time, a resource inventory can establish the approximate cost for cloud-hosting these applications. Of particular interest would be sets of applications supporting a small number of high-unit-value-of-labor workers; these applications will likely provide the largest ROI in cloud migrations.
Assessing an application's 'agility value'
The final thing to look for in creating business value for cloud migrations is the "agility value" of the applications involved. It's here that software professionals can provide the most valuable aid to business or operations analysts.
Identify applications that are underutilizing their server resources. These are prime candidates for migration to the cloud.
Many applications support a relatively fixed community of workers on a regular schedule with fairly low rates of functional change over time. These applications can almost fall into a "deploy-and-forget" category, so they benefit little from being built to expand or contract resource usage, substitute components, customize worker interactions or support large numbers of workers even across company boundaries. The flexibility and agility of the cloud isn't likely to benefit these applications -- but those with the opposite characteristics are top candidates for migration.
Application lifecycle management (ALM) procedures and audits will normally provide software specialists with an understanding of how often a given application changes, which will be a partial measure of how likely it is to outgrow its current hardware platform.
Sudden changes in server resources, the need to add or remove workers and major changes in utilization through the year based on workload are also indicators that an application is going to easily outgrow its platform. When server assets are written off prematurely, capital and operations costs are higher than in cases in which servers can be maintained through their expected depreciation periods. Businesses report that the largest reason for premature write-down of servers is unexpected growth in application demand.
Identifying the best candidates for cloud migration
If an application meets all of the business conditions outlined above, it's a good candidate for consideration for migration to the cloud. Even meeting two of the three is likely enough to merit further review.
However, applications that fail all three tests may not be worthy of the next steps in cloud migrations: the assessment of the application's technical characteristics to see whether they match cloud capabilities, and a detailed review of cloud deployment and operations impacts on application QoE and cost.
Because that level of evaluation is both expensive and time-consuming, application specialists should work with business operations professionals on the first critical business review of application migration potential. Doing so will avoid wasted effort and improve cloud project quality.