April 18, 2017
Cloud migration: Which apps make the cut—and which don’t? It matters.
By: Miha Kralj

Cloud migration can be hard—there’s no way around it.

Doing it right is critical to capitalize on the real value of the cloud and to gain performance potential. One of the keys to success is determining which apps you should move to a cloud platform and which should stay put.

Packing Your Bags for Cloud Migration? Take These Three App Types With You @AccentureTech

Three types of apps should be your top targets and go to the head of the queue:

  1. Apps that have a variable load.

    These are perhaps the best candidates to move to the head of the queue. Examples includes:

    • Environments used only for development and testing, typically nine hours a day, five days a week. When they are in the cloud, you can turn them off at night and over the weekend, so you do not pay when they are idle.

    • Apps that deal with seasonality or bursts of demand—such as an e-commerce holiday site, which during the height of the season could require 60 servers, but only two the rest of the year.

    • Apps that will experience planned, rapid growth. A cloud infrastructure is better suited to support a fast-expanding business than the traditional approach: pre-purchasing apps for the data center and then having to work magic with capacity planning when things take off.

  2. Public-facing apps with a global reach.

    The web or mobile apps are a clear choice, and so are any apps that require a Content Delivery Network—when you need to push some form of data (for example, video or web pages) as close to your international audience as possible. The cloud’s virtualized nature is a much better option than a typical one- or two-site data center.

  3. Apps you expect to modernize soon.

    Such as when you need to invest in refactoring to reduce costs, strengthening security or boosting horizontal scalability to support growth. By moving those apps to a cloud infrastructure, you can achieve all of that and more—faster and less expensively than doing it in the data center.

The following apps you should not move:

  1. Avoid apps that are too hard to migrate.

    These may need custom hardware the cloud platform cannot provide, or require super-low latency, or depend on COTS components or middleware that isn’t cloud friendly. Applications that are older than 15 years always require careful analysis of migration ability.

  2. Apps that are too risky to migrate.

    These may lack sufficient documentation, be based on the old or brittle code, have high-availability requirements without an option to refactor or be part of a “spaghetti” dependency chain. Be on the lookout for old in-house applications created in arcane coding languages—they are often not documented at all and will be very hard and risky to move.

  3. Apps that simply don’t offer enough benefits.

    Some apps are not worth the effort and money to move because the payback will never cover the costs. For example:

    • An app scheduled for decommissioning soon. It simply wouldn’t be in the cloud long enough to generate enough return on investment.

    • An app that went through a modernization in the past six or seven years; it was likely highly tuned to run well in the data center, but moving to a cloud platform would generate only little performance improvement, and you’d still have to invest in the migration.

Regardless of which apps you end up moving, a cloud factory approach could help increase both the speed and quality of the effort. A robust cloud migration effort includes four key features:

Learn more about Accenture Cloud Factory.

No doubt, the cloud offers many significant potential benefits—but only if the apps are a good match for cloud architecture; some simply aren’t.

Save yourself time and money; do your homework and make sure the apps you hope to move can give the bang for the buck you expect.

Popular Tags