On Tuesday we explored where cloud deployments can go wrong at the business level, which, while a major impactor of success, don’t account for the full range of what can go wrong.
The other side of the cloud coin of course is what can go wrong from the technical perspective at the organizational level. Business and technology perspectives have a lot of overlap where cloud is concerned, since the technology is so closely tied to how a business actually functions and therefore collects revenue.
Anyway, here’s our list. Let us know on Twitter @CloudGathering if you have others that you’ve seen really messing up cloud deployments:
Poor measuring of performance: All too often, the efficacy of a cloud is not measured in the right way. To understand the full extent of performance issues or success, measurement has to occur at peak times. While this perhaps sounds ridiculous, there are countless examples of an IT consultant looking at 90 days’ worth of data outside of the holiday rush for a seasonal business that does 90% of its sales during holidays or similar asymmetrical patterns within the calendar year. Understanding when performance is going to be most important is key to appreciating how to evaluate your cloud.
Compatibility is not considered: The cloud works so well for many businesses because it can help circumvent many of the compatibility issues that plagued the IT departments of yester-year. However, many of those departments are currently looking to or are already integrating cloud into their overall strategy. While an app will run on Linux, security features or that other legacy app that it depends on may not. These are issues that if left unaddressed can seriously undermine the success of a cloud deployment.
The question of scalability: The point of moving to the cloud is chiefly for scalability, though there are other well-documented benefits as well. However, if your app hasn’t been designed to scale, the cloud computing isn’t going to make it function better. If adding 20 physical servers didn’t help, then adding 100 cloud servers isn’t going to help. If your app is lacking in this way, address those issues before moving it to a cloud environment.
Failing to reverse engineer: Turn-over in the job market is now the reality of the American working world, but the IT working world as well. This, however, means that most apps were written by different folks than are planning to migrate them. This presents a bit of a Catch 22 for many IT departments – they must reverse engineer everything prior to considering moving it, otherwise it will almost certainly break. Take the time to understand the implications of what the app will require in the cloud. A managed service provider can often be an effective consultant to aid this process.
By Jake Gardner