Supporting Robust Back Ends, with No Buts….By Nanther Thangarajah
February 13, 2015
Everyone wants their shot at internet fame. There really is no magic formula, well, unless you're Kim Kardashian. Doubly so, if it also happens to be her…unclothed rear.
Marketeers and engineers are both concerned about internet traffic. One group with creating it, and the other group with supporting it. So, things can get quite interesting with us engineers who work in support of marketing.
The biggest challenge is trying to help folks understand the technical aspects of servers and web services, and in general, how everything works together (even if spit and baling wire are frequently used, and mystical incantations are required to make them function properly). Most folks might not feel that they need to know or understand the process, as long as there's others who know what's required to make things happen.
But we've all been in situations that require us to provide someone as much information as they need to make decisions. It takes collection, tabulation, organization and analysis of data, and a presentation of the risks and benefits of one course of action over another, as well as fallback plans and what I call "OCOs" ("Oh, crap" Options).
That's why I love coming across articles like this one from medium.com. It takes an anticipated situation (enhanced traffic to an already well-visited site) and uses it as a case study (for marketeers, engineers and everyone in between) to explain what they had, what was required and how they got from here to there, while keeping things going the whole time. Along the way, we're given plain language explanations of concepts like content distribution, scaling, virtualization and load-balancing.
Now maybe people will understand why I regularly wear belts *and* suspenders. It's not about anything going wrong, but all about being prepared in case something does.