[RodgerB]

I was just thinking how much increased server load Google would receive after it's implementation of Google Instant, after navigating a page which didn't have a search feature at all and relied only on pagination (note that I'm not talking about the autocomplete feature, I am talking about the page loading search results as you type).

I was thinking I might like to add a search feature to this site for free (it is the site for my prospective university's student clubs/associations list, I'm looking to start a double degree in Bachelor of Engineering (Robotics and Mechatronics)/Bachelor of Science (Computer Science and Software Engineering) next year). Of course, this project wouldn't be fun if I was just adding a stock standard search implementation, I would like to develop an AJAX feature such as Google Instant using jQuery.

Developing such an application I would find relatively easy as I find the jQuery framework very simple to use, but engineering the application is a totally different scenario, where I am able to make it security-minded, but I am unsure how to engineer a web application to perform well, if say, a lot of people were using it at the same time, and how to make sure the infrastructure that is being used would be able to handle the load.

My question is as a web developer, how do you engineer your applications for their server intensity in a real world scenario instead of a sandbox? I'm not looking for cross your fingers and hope for the best answers, nor am I looking for responses advising me that the Google Instant type search functionality for said website will likely work and that I shouldn't worry about it; I just want to know the 'engineering' process for developing such an application.

Also, have there ever been instances where you've strategically needed to use a cloud computing model, or have designated a server(s) to process AJAX requests independent of your main web server?

Thanks in advance, I just thought this might spark some interesting discussion.

[JackOfAllTrades]

I would say the major concern is going to be database performance with all the "LIKE" queries. The DB would need to have well-designed schema and proper indexing, perhaps full-text indexing. It might be valuable to look into an external indexing engine such as Lucene/Solr or elasticsearch.

[RodgerB]

Yes, that's probably more to the point. Sad this hasn't really sparked as much discussion as intended, but that was to be expected from the JS forum .

[e_i_pi]

RodgerB, on 20 December 2011 - 11:45 AM, said:

Yes, that's probably more to the point. Sad this hasn't really sparked as much discussion as intended, but that was to be expected from the JS forum .

Well I was going to say cross your fingers and hope for the best, but you poo-pooed that idea