The fine folks at NASA, with notable contributors such as the Internet's baby-daddy Vint Cerf, have been bit twiddling a new communications protocol this month that has been in the works since 1998. The "launch" of Delay/Disruption Tolerant Networking protocol is currently being field tested on the comet-seeking EPOXI spacecraft.
While TCP/IP has generally worked well beyond its initial design requirements as the terrestrial Internet has scaled unimaginably, it doesn't work so well in interplanetary deep space.
From the fine folks at Ars Technica:
This month, NASA began testing a new protocol
for communications in outer space that could extend the reliability and
versatility of the Internet out of the Earth's atmosphere. The new
protocol, called Disruption-Tolerant Networking, or DTN, has been in
the works for ten years, passed a month of testing with a just-launched
spacecraft and nine ground stations, but is still scheduled to undergo
Communicating in interplanetary space is hard. While even the most remote regions of the earth can be reached by lightspeed communications in a modest fraction of a second, and voice conversations from the earth to the moon can be carried out with only a barely noticeable delay, several light minutes separate planets even at their closest approaches. Back-and-forth negotiation isn't feasible, and the cost of starting processes from scratch are high. Furthermore, disruptions of communication are numerous and routine. Satellites and planetary probes have much less power when they're out of the sun, line of sight must be maintained, dishes properly aimed, etc. Solar flares and other environmental factors can shut communications channels unexpectedly.
Under the new DTN protocol, nodes retain data in their own memory until they receive confirmation the data has been received by a suitable target node. This increases the likelihood that data will arrive at its destination with a minimum of back-and-forth, communication even when communication is intermittent or unreliable.
I wonder if it's IPv6 compatible? After we assign a DTN/IP-Address to Internet-enable each celestial body, we'll be out of addresses again!
BTW, I happen to have access to a DTN-enabled uplink which proxy relays my TCP/IP to DTN through the EPOXI spacecraft. Check out the round-trip times on this badboy:
PING jupiter.solarsystem.com (10.0.0.1): 56 data bytes
64 bytes from 10.0.0.1: icmp_seq=0 ttl=252 time=1662.194 lightyears
64 bytes from 10.0.0.1: icmp_seq=1 ttl=252 time=109.738 lightyears
64 bytes from 10.0.0.1: icmp_seq=2 ttl=252 time=109.098 lightyears
64 bytes from 10.0.0.1: icmp_seq=3 ttl=252 time=109.165 lightyears
64 bytes from 10.0.0.1: icmp_seq=4 ttl=252 time=99.230 lightyears
64 bytes from 10.0.0.1: icmp_seq=5 ttl=252 time=101.702 lightyears
--- jupiter.solarsystem.com ping statistics ---
6 packets transmitted, 6 packets received, 0% packet loss
round-trip min/avg/max/stddev = 99.230/365.188/1662.194/580.053 lightyears
I am interested in understanding if there are any additional security mechanisms built into DTN as it would be a shame if an advanced alien race could perform an interplanetary MITM on our transmissions: "...this is not the planet you are looking for..."
For the sake of humanity I hope so. I'm going to go read the drafts.
Botnets? Data leakage? Clickjacking? You think you've got problems, just think of the firewalls needed to protect against solar flares ;)*