Navigating the Turbulent Clouds for your Applications

cloud.gif   I don’t know about you, but I find it hard to sort through all of the noise around Cloud Management to get to the real heart of the matter. Between cheesy slogans like “To the Cloud” and impenetrable marketing, it is very hard to peel away the nonsense and fluff to understand the real solutions. The situation isn’t much better in the world of Application Performance Management (APM) – a little less hype, but just as much marketing mumbo jumbo. So, if you are looking at leveraging cloud resources for your applications, how do you navigate the turbulent waters to find the right way forward. To help in your quest, I would posit a few core tenets that should help us navigate.

1. 6 weeks is the new 6 months

     If you are considering moving, or already have moved, some of your application infrastructure to the cloud, you are likely doing so because of immense pressure to stay competitive and reduce costs. In that kind of world, how does a multi-month implementation schedule even make sense? Your APM solution should provide value within days, not weeks or months, and it shouldn’t require you make deep changes to your applications (like code insertion). And when you do get it up, it should be easy to use. So, just say no to slow and complicated solutions for APM in the cloud. You deserve better.

2. End-User performance is king

     The vast majority of cloud-hosted applications are web-based. Even if you are using the application on a mobile device as an “app”, it is likely based on good ol’ HTTP on the back-end. This is great for developers, because you can ditch the old client-server complexities. However, this also means that your users expect more from your applications. They expect loads of less than a couple of seconds, and companies like Google have found that users will visit a site less often if it is only a 1/4 second slower than competitors . In this kind of environment, it is amazing that only 1 out of 4 companies know their end-users’ performance on a continuous basis. The conclusion is clear – if you want to be competitive, you need to keep your finger on the pulse of your users, and react quickly when you see a problem.

3. Global is normal

     Successful internet sites today can expect a worldwide audience. That means thinking about complex distributed architectures and application delivery models like Content Delivery Networks (like Akamai). In this kind of model, you need to understand where you users are coming and what path they are taking. In particular, you need visibility all the way from your cloud to the user, even if that path crosses a cloud network like Akamai’s. At the end of the day, your users don’t care how hard it is for you to do this. The web site is slow. They can take their business elsewhere.

4. Sorry isn’t good enough

     There will always been errors in any new release to your applications. Murphy’s Law won’t allow you to avoid them. What you can control is how you react to the problems. And I can guarantee you that waiting until your users call and complain and then taking slow action isn’t the right answer. The reality is that most companies only see 1% of the user errors that happen on their site. That isn’t a  recipe for success. As Wayfair grew from a small company to a serious online retailer, they found themselves unable to keep up (Watch the video). Wayfair decided that they need to have that pulse on the user we discussed in #2, to make sure that their releases were successful. And now, watching the end-user is a non-negotiable part of their release process.

We’re really only scratching the surface here, but we can already see that the Cloud is a catalyst for turning what might have been manageable problems into nightmares without the right tools. Bottom line, in the wild world of the cloud, you either prepare well, and do it quickly, and your cloud venture will go to website heaven to mourn the old days with Pets.com and WebVan.

Image from http://www.mymomisms.com/

These postings are my own and do not necessarily represent BMC's position, strategies, or opinion.

Share This Post