Fifteen petabyte network from CERN

Quote from: destined on July 16, 2008, 03:22:43 AM

So where does one get one of these? Well if your doing super-collider research, what kind of network speed do you think you will need to map out the reaction in real time?

Here is a small article on it (its from CERN)

http://news.cnet.com/The-15-petabyte-network-and-the-atom-smasher/2100-1008_3-6243726.html?tag=nefd.top

That is very cool. Reminds me of the quantum cryptography, needing dual fiber links to be able to use quantum crypto on your network.

On the real-time thing, speed has nothing to do with the delay between event and observation, that is determined by what is called latency, which is the time it takes to travel from point a to point b. Say i'm using dialup with a top speed of 24k a second. then jump to broadband at 10mbps. more speed, but the time it takes for the request for the data to travel across the physical distance, and the reply to be sent back is still roughly the same. say the latency is one second, on dialup it will take one second for the data to move from the host computer to my own, although since the host is sending one packet after the other you only really notice the first second. the difference is on dialup, ti would only be able to send back 24k at one time, while broadband it would be 10mb at one time, so the data gets through faster, but the latency is still similar.

Latency cannot be eliminated, currently the lowest-latency link we have is fiber (since light travels at- the speed of light) the only way to have no-latency is some kind of time travelling data. Which, I think is impossible, but quantum entaglement MAY be able to prove me wrong.

Enough nerd-ranting, i'll shut up now.