While working on our new Acceleration Reference Architecture, I have been reading a lot of reports documenting the effects of making websites and web applications faster. I’ve spent literally minutes of my life carefully analyzing these reports and meticulously cross referencing their reported results. Yes, I can now confirm readers: faster applications are definitely better. I do hope no one has spent too much time and money generating these reports. While their authors are certainly cleverer than me, and in many cases their organizations bigger and better known, I couldn't help wondering why they had bothered. What was I missing? Surely everyone that lived through the 90’s and 00’s waiting for those bytes to squeeze themselves down the fine capillary tube of our 56k modems know that slow applications are worse than fungal infections.
And yet there are still slow website and web applications. And studies telling us that they are a bad idea. Neither of these things should really exist.
I’m guessing that no one sets out to deliver a painful user experience – but the barriers to fixing application performance can be pretty high. Re-coding applications, using CDN’s or deploying dedicated appliances all bring with them cost, complexity or both.
That’s when it hit me:
The studies are there to justify the time, money and effort you need to spend to accelerate your web application.
So I decided to work with some of my colleagues – in particular Dawn Parzych our acceleration guru to design an architecture to demonstrate that just enabling a few simple features on your F5 BIG-IP could produce some meaningful performance improvement. By using what you already have, and taking advantage of F5’s new simplified licensing model, I think you’ll find that accelerating your applications is easier than you think, and you won’t need to read a ten page study to justify doing it.
I’m hoping to get the architecture launched by the end of the Velocity show, so watch this space.