SpazBlog

So, Google released a web accelerator. Interfaces with browsers as a local HTTP proxy, but communicates with Google's web servers. Since downloading it tonigt, it claims to have saved me 1.3 minutes of load time out of what would have apparently been 16.7 minutes of page loads.

I think Google can explain what it does better than I can:

• Sending your page requests through Google machines dedicated to handling Google Web Accelerator traffic.
• Storing copies of frequently looked at pages to make them quickly accessible.
• Downloading only the updates if a web page has changed slightly since you last viewed it.
• Prefetching certain pages onto your computer in advance.
• Managing your Internet connection to reduce delays.
• Compressing data before sending it to your computer.

Pretty neat. I wouldn't have thought of some of those, like local/remote caches with diff-based communication between them. Not quite sure what they mean by "Managing your internet connection to reduce delays" means though.

It has been pointed out that it can be used as a semi-anonymous proxy; semi because Google still knows where you go. I would also suspect you could get extra benefit by having multiple machines use the same instance of GWA, but they might only allow localhost to connect. Still, pretty neat stuff!

This is actually in line with something that I've wanted to mess around with recently, but haven't had the time (An NDF type project if I had NDFs). I have long wanted to write software to tunnel TCP connections over slow/saturated/unreliable connections, such as the one we have at the college. Something like taking over management of packetloss handling by encapsulating the TCP data over UDP with some semi-smart redundancy. UDP's connectionless status alone would significantly improve things, if you could manage the packetloss.