There’s a lot going on in the world of load and performance testing and it seems there’s more to keep up with every day. With that in mind, in this issue we’ve collected some of the most interesting links from around the web. Discover how to combat the claim that “requirements suck,” learn about SPDY and what it means for you, find out how to answer the question, “What would you say… you do here?” and, finally, establish once and for all why extrapolating from small scale tests can put you in performance hell!
1) Johan Zandhuis’ Response to Fiona Charles’: Requirements Suck? Get Over it!
A big debate right is whether or not “requirements suck,” or, to put it a bit less inflammatorily, whether or not they are the best way to get information for testing. In this article, Johan Zandhuis makes the counter point to a video talk by Fiona Charles from EuroSTAR 2012 that was recently published. His point is that, while requirements may not always be perfect, they are the basis for getting the necessary information. That said, requirements gathering is a skill unto itself that needs to be developed. The answer here isn’t that requirements suck, so let’s get rid of them; it’s that testers need to get back to requirement gathering. Check out the article to learn more.
The author starts this piece out by stating that, “this is a huge leap forward for web performance.” And it may be true, as wider adoption of SPDY means less need for hacks like domain sharing. Beyond just enabling for Facebook.com, it looks like the company has even flipped the switch for mobile. What does this mean for Facebook’s performance? And, more importantly, what might it mean for you? Check out the article from Zoompf.com to find out.
It may not be possible to state why this article is important better than the author already has: “Have you ever been in a situation where you had to explain why you can’t test everything or why sometimes bugs escaped your best efforts to find them?” If that person is not a tester, especially if he or she is non-technical, you might have a tough job on your hands. To learn just what to say, check out TestingThoughts.com.
The cartoon example may be funny, but the danger of extrapolation as it applies to application performance is not. Try to extrapolate and you’ll find out that your 2-server system might be able to handle 200 people…but when you scale up to the 5-server system, it may crash well before 500 users. Read on to find out why the real network is way more complex than simple extrapolation can handle and why you need tests to be as realistic as possible.