How not to do Cloud Testing: a real life experience

 How not to do CT: a real life experience

This article eventually was not published, as the body for which it was destined deemed it too negative. Well, everyone knows we learn more from mistakes than from successes. In my opinion, it is better to publish a "negative" story once in a while rather than yet another positive platitude that will not offend anyone but won't engage the readers either.

Migration to the Cloud was sold to our client as being the silver bullet to their cost-cutting efforts [Big Mistake #1] and the basis for their continuous business with our firm. A cross-divisional, cross-functional and cross-country team was assembled for this task. I was on the application testing side of things, together with the other co-located developers-cum-testers who had been the steady stewards of this application for many years. Our mission was handed down to us from higher above: make sure the application runs on the Cloud the same way as it runs on its current platform [Big Mistake #2]. An overworked architect from another city, an invisible team from another country and an under-responsive software partner were in charge of provisioning the test environment while my team was dusting off the manual test scripts from the last software migration. While we were ready to start our mission at any time, accessing the Cloud environment had become an exercise in frustration: everything was way more complicated now because it had to go through new global applications to request the required level of access to various cloud components we had little to no understanding of [Big Mistake #3], and had to install a rather finicky two-factor authentication system. The application being a legacy, highly customized, niche-software one, there were issues with installing the required licenses in the Cloud. Part of the issue was technical but a bigger part was political. To complicate things even further, a conflict over resources surfaced between two account managers. Our November deadline passed, so did the Christmas and then the Easter ones, while saber rattling was going on at the upper echelons and head scratching was a daily occurrence in the trenches. The joke on the team, when we were getting the updated plans from management was “Oh, now it’s for October? What year?” Eventually (11 months later to be specific), the environment was finally set up, the system was finally installed in the Cloud, and manual testing was completed successfully. But we managed to learn nothing about Cloud – we did not need to, in order to do the testing, and upgrading our skills was not part of the migration plan. The promised 20% cost reduction became a liability that our company had to pay. Even so, the contract was not renewed. The promised cost reduction morphed into a resource count reduction and two team members were let go. I believe the system successfully runs to this day. Lessons learnt? [1] Do not oversell the benefits of Cloud [2] Do not underestimate the complexity of the Cloud [3] Do provide meaningful Cloud training for your whole team.

As told to TA

Comments

Popular posts from this blog

Mobility of the workforce: how far have we come, how far will we go ? (2003)

Retirement rituals: parting (not partying!) with a gift

Customer support: lessons from IT (2) - the appliance vendor