Well, it’s been a while since my last post but we’ve been working hard on our Ajax project and I think I can share some of the challenges we will be tackling along the way. Here is a summary of the challenges with a brief explanation of each.
- When analyzing traffic flowing into an Ajax application you generally see some page request and then you see a series of Ajax calls with the page as the referrer some time later. So the unit of work that used to be a page is no longer very useful for these app types.
- When you see an Ajax hit you don’t know what was done by the user to initiate the call so if there is an error, performance problem, or general need to understand the context of the hit you can’t do it. This was very straight forward when the web was synchronous.
- In order to get the full picture for Ajax or other asynchronous web applications you need to understand three things instead of 1, like in web 1.0, those 3 things are:
a. What did the user do to initiate the call (mouse over, Click, drag and drop, etc..,)
b. What did the asynchronous call return as a response to the request
c. What information from the response(s) is useful to report on and/or replay
As you can imagine there are hundreds of ways to approach these challenges. What I wrote in my last post just started to scratch the surface for number 1 above. I would imagine many of you have even started to look at ways to solve these challenges yourselves out of necessity as you delve into this new web era. I’d be interested in hearing how some of are working differently today to address your performance and availability challenges with Ajax applications and also would love to hear comments on the problem definition I have above if you have any.
In my next post I will talk about some of the hacks, workarounds, and real solutions that are starting to show up today and then work towards the ideal solution that we all hope to have when it all washes out.