Vint Cerf Preps Interplanetary Internet Protocol

TechFiends32 writes "After years of working with NASA to bring Internet connectivity to deep space, scientists say Vint Cerf's efforts may be nearing completion. To combat the apparent challenges of extending the Internet into space (such as meteors and weighty, high-powered antennas), Cerf and others have made significant efforts, like adjusting satellite-based IP, and working on delay-tolerant networking (DTN) to address pure IP's limitations in space. According to principal engineer at The Mitre Corp., Keith Scott, 'The 2010 goal is designed to bring DTN to a sufficient level of maturity to incorporate it into designs for robotic and human lunar exploration.'"

Comments

More work to be done

[Chris Burke]

I assume then that at some point someone will have to write up a new RFC on "IP Over Space-Avian Carrier"?

KA9Q

[karl.auerbach]

Phil Karn's old KA9Q implementation of TCP (for amateur radio) was designed to accommodate very long delays.

Re:KA9Q

[Jimbookis]

Phil Karn's old KA9Q implementation of TCP (for amateur radio) was designed to accommodate very long delays.

Only because it takes such a long time to tap out IP packets in morse code.

Re:Caching would be great here too

[dk90406]

Yes, caching seems very nice. But the article don't explain how they'll handle the huge latency. It must have a huge floating windows for ACK/retransmits.
On a less serious hand, I hope the satellite IP connections are severed from the Ethernet (like electrical plants are (or should be in some cases), or hacking a satellite will be the next goal.

mooncam

[nblender]

I can't wait for the very first webcam on the moon; to see a live earthrise, etc ...

Re:mooncam

[__aapbzv4610]

actually there is no such thing as an earthrise on the moon, as the moon does not 'rotate' in relation to it's movement around the earth. At any point on the surface of the moon facing the earth, the earth will always be in the same point in the sky, always.

Re:mooncam

[vux984]

actually there is no such thing as an earthrise on the moon, as the moon does not 'rotate' in relation to it's movement around the earth. At any point on the surface of the moon facing the earth, the earth will always be in the same point in the sky, always.

An 'earthrise' is still possible, you just have to put the mooncam on a buggy and drive in the correct direction.

Re:mooncam

[Anonymous Coward]

Wrong.
Libration causes the visible face of the moon to oscillate slightly.
Therefore, you can see an Earthrise from certain points on the moon without being in motion relative to the moon yourself.

http://en.wikipedia.org/wiki/Libration

Re:mooncam

[Kingrames]

"Earthrise" is the name given to the famous picture taken of the earth from the moon. You have most likely seen it, it's the most famous picture of the Earth.

Africa is prominently visible in the picture, if you're curious.

Re:mooncam

[Tweenk]

Your argument seems good at the surface, but it's not true!

The truth is that the Moon librates a bit (a few degrees), so there actually ARE earthrises when you are near the edge of the Earth-observable Moon surface. The Earth just doesn't do a full circle around the sky, it travels along a Lissajous figure.

Even Wikipedia is incorrect on this, at least when you look up "Earthrise".

Re:mooncam

[Tweenk]

Before anyone questions this: The angular span of the Moon's libration is a few times larger than the angular diameter of the Earth as seen from Moon.

Re:mooncam

[Free the Cowards]

Rotation != the only thing that can cause Earthrise.

Re:Caching would be great here too

[Anonymous Coward]

aw, shit. now goatcx will be trolled into outer space, giving a new meaning to the term black hole.

Re:Caching would be great here too

[jandrese]

The trick is that you don't have to use TCP as your transport layer. DTN bundles can be transmitted over UDP, NORM, sneakernet, carrier pigeon, or anything else you can write a convergence layer for. Since DTN abstracts away the lower levels, each hop can use the transport layer that is most appropriate, like TCP on an internet hop, SCPS on a satellite hop, etc...

More information is available on the DTN Research Group's homepage: http://dtnrg.org.

Re:Caching would be great here too

[AKAImBatman]

I think you're missing the point. The general idea here is to have a packet switched communications system throughout the solar system. That way if a probe is in the shadow of, say, Jupiter, it can bounce a signal off a probe orbiting Venus, which will relay the signal back to Earth.

The end result would be a more robust communications system. In the future, interplanetary communications satellites could even end up doing most of the grunt-work, thus allowing probes and manned spacecraft to carry smaller communications packages designed to work with the network rather than broadcasting in as many conditions as possible.

such a network would also be useful for astronauts on another planet or meteor. Rather than setting up a communications station, they can use orbiting satellites to relay their transmissions. (Something which NASA already does on a smaller scale with probes like the Mars rovers.)

Bold (Crazy) Prediction

[religious freak]

This will be in wider use in 30 years than IPv6

Maybe not, but wouldn't it be crazy if it was?

From TFA

[scubamage]

"Reliability in DTN is provided by a mechanism called custody transfer, where nodes in the network can assume responsibility for retransmitting lost messages. This allows for retransmissions from inside the network rather than having to retransmit data from the source, as is the case with TCP." Hmmm, sounds like DoS just got a whole lot easier. Instead of having to get nasty at an endpoint, you could attack a single router and have everything get all kinds of wonky. I understand why they want to do it this way, but the seperation of responsibility was put there for a reason in TCP waaaaay back in the DARPA days so that if any link goes down you have no data loss. What happens if critical data is being transmitted from a source, and the source gets cut off. The retransmitting router gets hit by a meteor and is trashed. Critical data loss. Am I missing something?

Re:From TFA

[644bd346996]

You seem to be extrapolating quite a bit to say that this scheme is much more vulnerable to critical data loss. (And your claim about DoS is pretty irrelevant when you consider that all implementations of this protocol will be owned by NASA and their associates.)

Do you really think, based on just TFA, that Vint Cerf of all people would design such a flawed protocol? The point of custody transfer is that retransmissions can be handled by the routers that form the network, rather than wasting precious power using a planetside rover that has better things to do.

Re:From TFA

[Mattsson]

You don't simply send your packet and then wipe your cache. You keep the packets at every hop it traverses until you know that it has arrived at its destination, so that you can resend it in case of a transmission error or fatal equipment failure.
Especially since you, with these distances, can have a large chunk of data in transit between two satellites due to the slow pace of light and radio waves.
That, along with the limited transmission speeds, is also one of the reasons why you do not want to resend lost packets all the way from the originator, which is still possible in the worst case scenario where the data is lost in all the routing satellites.
Unless the probe, or Mars lander or whatever also loose the data before it gets through.

Actually, such a system is more secure, from a data loss point of view, since the data can get through even if it is lost at both the originator and at some point in transfer, since it can be resent by any router that it has passed through.

It makes sensitive data vulnerable to interception though... In case some aliens where to abduct a satellite that's caching data before it can be purged. =)
 

Re:Caching would be great here too

[HolyCrapSCOsux]

I'd like to put a few people in space, sans suit.

Re:Caching would be great here too

[jandrese]

Caching is very useful in space. What happens when your satellite orbits around to the other side of Mars? You have several hours of no-communication and have to store everything you were going to send (and people on the other end have to store what they were going to send to you).

Excellent

[silentcoder]

Now we'll be able to send e-mail to Dr Edgar Mitchell's aliens and ask them if they exist !

In related news ...

[PPH]

... Comcast moves to block P2P (planet to planet) traffic.

Interplanetary ICMP

[kwabbles]

~$ traceroute voyager2.heliopause.net
traceroute to voyager2.heliopause.net (207.46.193.254), 30 hops max, 40 byte packets
  1 192.168.0.15 (192.168.0.15) 0.180 ms 0.186 ms 0.205 ms
  2 netblock.dslcarrier.com (66.159.218.1) 14.379 ms 17.076 ms 20.048 ms
  3 satrptr.spacenet.net (66.51.203.33) 36.531 ms 45.014 ms 42.245 ms
  4 mars.spacenet.net (206.223.143.41) 92.229 ms 101.596 ms 99.575 ms
  5 jupiter.spacenet.net (216.239.43.12) 220.073 ms 266.554 ms 254.288 ms
  6 saturn.spacenet.net (209.85.253.178) 880.760 ms 854.294 ms 878.981 ms
  7 uranus.goatse.net (209.85.251.94) 1233.954 ms 1332.028 ms 1315.059 ms
  8 neptune.spacenet.net (74.125.19.104) 1703.205 ms 1721.652 ms 1733.635 ms
  9 pluto.spacenet.net (73.113.43.11) 2301.311 ms 2435.201 ms 2448.221 ms
10 * * *
11 asteriodb3221.microsoft.com (207.46.191.230) 3411.411 ms 3813.153 ms 3761.314 ms
12 voyager2.heliopause.net (207.46.193.254) 7810.134 ms 7956.324 ms 8103.132 ms

~$

Re:Interplanetary ICMP

[Anonymous Coward]

Man I wish I could afford a faster than light connection like you... Here I am stuck on a legacy electromagnetic link:

~$ ping voyager2.heliopause.net
PING voyager2.heliopause.net (207.46.193.254) 56(84) bytes of data.
64 bytes from (207.46.193.254): icmp_seq=1 ttl=54 time=72877083.5 ms
64 bytes from (207.46.193.254): icmp_seq=2 ttl=54 time=72877853 ms
64 bytes from (207.46.193.254): icmp_seq=3 ttl=54 time=72979083.2 ms
64 bytes from (207.46.193.254): icmp_seq=4 ttl=54 time=72877483.6 ms
64 bytes from (207.46.193.254): icmp_seq=5 ttl=54 time=72897053.8 ms

--- voyager2.heliopause.net ping statistics ---
5 packets transmitted, 5 received, 0% packet loss, time 364508557ms
rtt min/avg/max/mdev = 72877083/72901711/72897054/72901711 ms

It took me over 4 days just to run that!

Re:A new area awaiting patentification

[clarkkent09]

Citizens of Earth,

It has come to our attention that an earthling called Vint Cerf is making unauthorized use of our work in violation of GPTO (Galactic Patents and Trademarks Office) patent number 0932984720392837409 for Interplanetary Internet Protocol.

We demand that he must immediately cease the use and distribution of our work and that he forwards all copies and relevant documentation to us by the earliest space courier. Failure to do so will result in a lawsuit to the amount of our estimated damages of 1,008,076,123.09 galactic credits (equal to 0.0008 USD).

Very truly yours,

Aliens

P.S Greetings, Dr Mitchell

Re:A new area awaiting patentification

[jandrese]

They must have had a 1000 years of space-Bush presidency.

in space...

[notgm]

in space, no one can hear you ping.

Also appropriate for station wagon full of tapes?

[Locke2005]

"Never underestimate the bandwidth of a station wagon full of tapes hurtling down the highway." -- Andrew S. Tanenbaem, Computer Networks, 4th Ed. p. 91

Hmmmmmm

[tgd]

Per byte its probably still a lot cheaper than using SMS.

Interstellar networking

[ckthorp]

Interstellar networking: putting the Ether in Ethernet.

Virus Uploads still allowed?

[A440Hz]

Will this extension of IP still allow virus uploads to alien ships via MacBooks?