News for hackers. Stuff that matters.
Re:MIT Researchers Develop Computer-Generated TCP That is Tw... by diroussel on Tuesday July 23, 2013 @10:23AM (#6090195)
Looks very interesting.

So we if we imagine this gains traction, could a mobile device use a different transport protocol for different network situations? Such as a phone going from 3G to 4G to wifi?

I guess the problem is the other end would have to adapt too?

Also if you are going to add new protocols, then would be not be better to add more message oriented protocols, rather than stream oriented? As many applications (twitter, payment instruction, email send) are message oriented.

Re:MIT Researchers Develop Computer-Generated TCP That is Tw... by octo_t on Tuesday July 23, 2013 @05:15AM (#6088979)
Bit of a misleading headline:

>> For example, with eight senders sharing a 15 Mbps link, a >> Remy-generated algorithm achieved more than twice the >> median throughput, with less than half the queueing delay, >> of Compound TCP and TCP NewReno. It achieved 70% more >> throughput than TCP Cubic and more than a threefold reduction >> in queueing delay.

Re:MIT Researchers Develop Computer-Generated TCP That is Tw... by igravious on Tuesday July 23, 2013 @05:23AM (#6089004)
Re:MIT Researchers Develop Computer-Generated TCP That is Tw... by killermonkeys on Tuesday July 23, 2013 @05:36AM (#6089038)
What exactly is misleading given there are several common algorithms used (cubic by Linux and compound by windows) and the gains are at least an aggregate 2x improvement on key performance metrics.

One thing I do see is that it hasn't yet been tested on real networks, only synthetic networks. It will be interesting to see if it provides real gains across the range of conditions, particularly in hyper optimized data centers.

Re:MIT Researchers Develop Computer-Generated TCP That is Tw... by cycnusx on Tuesday July 23, 2013 @05:02AM (#6088940)
Cool.