4 ms·
A little over 10 years ago I was on an engineering internship and worked on a NASA project that was part of the overall goal of converting ad hoc links to being
by phaedrus 7y ago
A little over 10 years ago I was on an engineering internship and worked on a NASA project that was part of the overall goal of converting ad hoc links to being IP-based. I wrote a test tool which allowed simulating long distance and/or unreliable links. You could specify chains of filters on the command line for both uplink and downlink, as well as introduce taps. The reason I couldn't use an off-the-shelf tool was that by "long distance" I mean "interplanetary". It's one thing to introduce a delay/drop, but my tool simulated the possibility of gigabytes of data being "in flight" along the link.
- rootbear 7y agoWas this part of Vince Cerf's "Interplanetary Internet" project? He's spoken about that at NASA Goddard a few times. I once toyed with the idea of using "in flight" data as mass storage, where data would be kept in a ring the size of a planetary orbit, with satellites stationed along the orbit to receive and re-transmit the data. You'd need a lot of redundancy...
- sitkack 7y agoThis for others following along more than you, but the work that Luigi Rizzo did on IPFW [1] and FEC (forward error correcting codes) for internet scale video delivery is amazing. I used ipfw (my router was running FreeBSD) to give ICMP traffic 300ms of jitter and 5% packet loss. Enough to make my DSL endpoint sour enough that the hack attempts went to zero. They would scan with ping, decide that my link was garbage and move on. [1] https://github.com/luigirizzo/dummynet https://github.com/luigirizzo/dummynet [2] http://www.iet.unipi.it/l.rizzo/research.html http://www.iet.unipi.it/l.rizzo/research.html