Forgot your password?
typodupeerror
Space The Internet Networking Wireless Networking Hardware

Vint Cerf Preps Interplanetary Internet Protocol 177

Posted by timothy
from the would-be-helpful-at-the-jupiter-hotel-too dept.
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.'"
This discussion has been archived. No new comments can be posted.

Vint Cerf Preps Interplanetary Internet Protocol

Comments Filter:
  • Cool stuff. The caching mechanisms to make information even remotely useful would be great back here on Earth. I hate even a few hundred milliseconds of delay when flipping from page to page. Google of course has a rediculously low latency seemingly in both transmission and server-side processing. I'm going to try and download the huge Wikipedia database and see if I can get it working completely locally. Click, click, click. No waiting, no flipping pages. At some point you need to start bundling large amou
    • Re: (Score:3, Interesting)

      by dk90406 (797452)
      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.
      • by Anonymous Coward on Thursday July 24, 2008 @04:08PM (#24324659)
        aw, shit. now goatcx will be trolled into outer space, giving a new meaning to the term black hole.
      • by jandrese (485) <kensama@vt.edu> on Thursday July 24, 2008 @04:09PM (#24324665) Homepage Journal
        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 [dtnrg.org].
    • 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.)

      • by smoker2 (750216)
        Does this mean I get my NeptuneCam ?
        What kind of bandwidth have they got ?
      • by SEWilco (27983)
        Or put the technology in Earth cars and pass data around between vehicles moving around. Send traffic conditions back to the cars behind you, via both the cars behind you and the cars going the other way.
        • by Teancum (67324)

          Or put the technology in Earth cars and pass data around between vehicles moving around. Send traffic conditions back to the cars behind you, via both the cars behind you and the cars going the other way.

          This already exists on the Earth. A simple 802.11 wireless router can easily transmit this sort of information between cars and send traffic conditions up and down the highway. All that would be needed to finish the idea is a good application data protocol to identify what information you want to share,

      • by Minwee (522556)

        such a network would also be useful for astronauts on another planet or meteor.

        Actually, astronauts on a meteor might appreciate a really good heat shield more than a reliable interplanetary data network.

        They might have also appreciated getting _off_ of the meteoroid before it entered the Earth's atmosphere and became a meteor, but it's probably too late for that.

        • by Teancum (67324)

          I think that an astronaut might really appreciate a good reliable interplanetary data network far more than a good heat shield.

          Currently, the "Deep Space Network" that has been the communication system for dealing with space-based communications requires "tasking" parts of the network for each mission that is "out there" working.

          One of the problems with the Viking spacecraft and currently the Voyager spacecraft was that it required paying salaries of individuals to orient the ground-based equipment receivin

          • I do believe the parent to your post was poking fun at my (rather silly) mistake of referring to an Asteroid as a Meteor. While not necessarily an incorrect usage, it does tend to refer to those rocks that hit the atmosphere and burn up. :-)

            • by Minwee (522556)

              Actually, it is incorrect usage. Keep that in mind when your house is destroyed and you try to collect on your meteor insurance policy.

              Would I make something like this up?

          • by Minwee (522556)

            "meteor" (meaning a rock small enough not to even be considered an asteroid)

            This would all become much clearer if you would look that word up. It doesn't mean what you think it means.

      • by Teancum (67324)

        From what I've read, Mars is in fact the target for the first implementation of this new communication standard. Basically, there is going to be soon so much hardware from so many different countries on Mars that frequency bandwidth is starting to be a significant issue. Assuming that this is a trend of things to come and not a momentary fluke of current planetary exploration at the moment, it is quite reasonable to presume that interplanetary communication between the Earth and Mars is only going to beco

        • Indeed. I personally think it would be amazing if we could develop a set of standard comsats to spread across the solar system. Want to explore Mars? No problem. We already have a communications infrastructure for you. Explore Europa? We've got sats around Jupiter, too. Throw in some sats traveling along the Interplanetary Superhighway and you've got better coverage than your average cell phone.

          I think people tend to underestimate the shear enabling factor of such a communications network. Not only would su

  • by Chris Burke (6130) on Thursday July 24, 2008 @04:02PM (#24324533) Homepage

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

    • by Teancum (67324)

      As funny as it sounds, the IP-over Avian RFPs actually give some interesting insight into the IPN in terms of significant ping times and packet loss.

      While the RFCs about avian packet carriers were largely tongue-in-cheek jokes, some of the concepts really did play a role in developing these new practical standards for very long range communications. At the very least, by thinking in terms of avian packet carriers, it is certainly the kind of "out of the box" thinking that allows you to try and pare down th

      • by Chris Burke (6130)

        What is more concerning is the actual implementation of avian packets.

        According to the wiki page for it, some LUG actually did it and sent nine "ping" packets and received responses. Pretty hilarious.

  • KA9Q (Score:5, Informative)

    by karl.auerbach (157250) on Thursday July 24, 2008 @04:02PM (#24324539) Homepage

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

    • Re: (Score:3, Funny)

      by Jimbookis (517778)

      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.

    • by ShakaUVM (157947)

      Heh, I remember Phil Karn from the days when cable modems were first being deployed in San Diego (he's an engineer at Qualcomm). He wrote the first linux client to allow computers to connect to the cable modem network (this being back in the days when roadrunner required a weird login application), and was also pretty active on the local USENET forum for RR... also wrote some papers on how to make nuclear weapons or something... good times, good times...

      Weird seeing a name you knew briefly 10 years later...

  • mooncam (Score:4, Interesting)

    by nblender (741424) on Thursday July 24, 2008 @04:05PM (#24324607)
    I can't wait for the very first webcam on the moon; to see a live earthrise, etc ...
    • Re:mooncam (Score:5, Informative)

      by SKiRgE (411560) on Thursday July 24, 2008 @04:20PM (#24324895)

      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: (Score:3, Insightful)

        by vux984 (928602)

        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 (Score:5, Informative)

        by Anonymous Coward on Thursday July 24, 2008 @04:57PM (#24325531)

        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 (Score:4, Interesting)

        by Kingrames (858416) on Thursday July 24, 2008 @05:08PM (#24325663)

        "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.

        • by houghi (78078)

          And if you are REALLY curious, you can look at it [google.com]

        • by Dirtside (91468)

          Hm, I'm not sure you're correct. "Earthrise" is this picture [jackkennedy.net].

          The famous picture of Earth with Africa visible is this picture [ning.com]. It wasn't taken from the Moon (I recall this from reading about it in the past, although I can't cite it; however it's easily provable, as the position of the camera is much too far below the Moon's ecliptic; you cannot see the entirety of Antarctica from any point on the Moon).

          • by Kingrames (858416)

            go figure. I have a huge photoshop on my wall, that has that earth picture where the other would be.

            on the other hand it might be that there's more than one picture named "earthrise" since the title of my wallpaper when I got it was called the same thing.

      • Re:mooncam (Score:4, Informative)

        by Tweenk (1274968) on Thursday July 24, 2008 @05:34PM (#24326029)

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

        The truth is that the Moon librates [wikipedia.org] 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".

      • The phenomenon is known as Tidal Locking [wikipedia.org].
      • by Cyberax (705495)

        Earthrise on Moon IS possible. Moon's orbit is not a perfect circle, that's why we have http://en.wikipedia.org/wiki/Librations [wikipedia.org]

        In the best case Earth rises at about 15 degrees.

  • by religious freak (1005821) on Thursday July 24, 2008 @04:10PM (#24324685)
    This will be in wider use in 30 years than IPv6

    Maybe not, but wouldn't it be crazy if it was?
    • by Daimanta (1140543)

      The only way to make ipv6 is to force it down the throats of the ISPs. Too bad force and business don't mix. Unless it's the businesses that force something through(like the DMCA)...

      • by jandrese (485)
        Interestingly enough, the only way I'll use IPv6 at home is if my ISP offers it. I have no interest in trying to set up complicated tunnels on my dynamic home link.
    • by Zeussy (868062)
      IPv6 won't have enough IP's for all of space!?! we need IPv8!
  • From TFA (Score:4, Interesting)

    by scubamage (727538) on Thursday July 24, 2008 @04:12PM (#24324727)
    "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 (Score:4, Interesting)

      by 644bd346996 (1012333) on Thursday July 24, 2008 @05:05PM (#24325637)

      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.

      • by scubamage (727538)
        Given that no other data is given, I really don't have much more to go on. If you do, please share. Yes, the point of custody transfer is that retransmits can be handled by the routers that form the network. But therein lies the problem. When TCP was created the goal was to create a network that could survive a nuclear war. TCP is designed so that if there's no confirmation, the source retransmits. However according to this article the routers handle retransmits. What happens if there is a bit of data that
        • by profplump (309017)

          Or you could just use out-of-band confirmations for critical data, and retransmit at an appropriate interval when you don't see the confirmation. And even that only applies to unsolicited transmissions -- for interactive traffic the remote host already expects a reply, and can simply retransmit the request if it doesn't get one.

          For example, it would be trivial to use this model for loose interactivity without wasting more than a few bytes of bandwidth:
          1. Earth host transmits new instructions with sequence n

        • by Mattsson (105422)

          The question is:
          In interplanetary transmissions, do you save more bandwidth by having less overhead, or by not having to retransmit every lost bit all the way from, say, a probe on Titan to Mission control?
          I don't know, since I haven't studied the problem, nor worked with interplanetary transmissions.
          I'd wager that there's quite a lot of bandwidth overhead lost to error control and correction in such transmissions today in order to have as few retransmissions as possible, though.

      • by naasking (94116)

        Do you really think, based on just TFA, that Vint Cerf of all people would design such a flawed protocol?

        Yes, if he can design it under the assumption that all nodes are trusted. As long as there's an appropriate node authorization presented at connection time, then the node can safely queue such data on the authorized party's behalf.

    • by kabocox (199019)

      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?

      Um, if you only have one retransmitting router, then you'll loose your critical connection to that end point i

    • Re:From TFA (Score:5, Insightful)

      by Mattsson (105422) on Thursday July 24, 2008 @07:12PM (#24327321) Homepage Journal

      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. =)
       

      • by gr8dude (832945)

        And so the notion of "aitm" (alien in the middle) attack came into life.

        Remember, you saw it here first!

    • by profplump (309017)

      You're right that it's not possible to lose data in the middle of the network with pure TCP. But if the original source gets hit by a meteor in the middle of (re)transmitting, you'd still lose the connection.

      Plus the worst-case scenario you describe here is this:
      1. I request data from a remote resource
      2. The remote resource transmits my data to some intermediate node
      3. Data is cached at some intermediate node, and the source is notified of the competed transfer
      4. The intermediate node is destroyed
      5. I notic

      • by Mattsson (105422)

        Transmission logs on the source server cannot be considered a reliable indicator of delivery

        It can if there's two types of entries.

        Node reception confirmed.
        Destination reception confirmed.

        If only the first one is logged, reception counts as failed.
        If both is logged, reception is verified.

        Just requires a notification of reception to be sent back from the receiver to the sender, apart from the notification sent to every node in the path.
        A little more overhead, but might be worth it. =)

  • Excellent (Score:3, Funny)

    by silentcoder (1241496) on Thursday July 24, 2008 @04:19PM (#24324873) Homepage

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

  • Soon Davidson will be able to spam martians.

  • by PPH (736903) on Thursday July 24, 2008 @04:23PM (#24324937)
    ... Comcast moves to block P2P (planet to planet) traffic.
  • Ya, but I bet it'll be used mostly for pr0n. And I bet the bandwidth will suck too, so we'll just end up with a bunch of ASCII art pictures of naked aliens.
  • ~$ 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

    • Re: (Score:2, Funny)

      by 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

  • MITRE (all caps) is the name of the company, not Mitre (in case someone wants to update the original post.)

  • in space... (Score:4, Funny)

    by notgm (1069012) on Thursday July 24, 2008 @04:58PM (#24325541)

    in space, no one can hear you ping.

  • by Locke2005 (849178) on Thursday July 24, 2008 @04:58PM (#24325551)
    "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 (Score:5, Insightful)

    by tgd (2822) on Thursday July 24, 2008 @05:08PM (#24325675)

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

    • by Mattsson (105422)

      Hmm...

      Why not simply put a few 3G towers around the solar system?
      Then all probes can send data via SMS, pictures via MMS and even do video calls to send realtime videos!
      And the phone companies can get their part of the NASA/ESA/Whatever-budgets they so rightfully deserve. =)

  • by ckthorp (1255134) on Thursday July 24, 2008 @05:12PM (#24325729)
    Interstellar networking: putting the Ether in Ethernet.
  • Will this extension of IP still allow virus uploads to alien ships via MacBooks [nitpickers.com]?
  • When we have a perfectly good system [wikipedia.org] already? We're going to need some really big flags, a couple of long poles. And a good telescopes.

  • How many IP addresses should they allocate to the rest of the universe??

  • Now I can get SPAM from fucking Jupiter ... bloody wonderful.

  • To deal with "meteors and weighty antennae"? Duh.... The #1 problem is *time*. You are going to have a *real* slow up/download when you're on Mars, and it's on the other side of the sun: hours.

                  mark "didn't even begin to talk about t a l k i n g t o T i T a n"

  • In the future, Scott envisions routers afloat in deep space

    Network management may be the biggest challenge. IP is already designed to deal with link failure and congestion. But, when the router out by Ganymede is down and won't come up you can't just dispatch a network engineer to zip out there with a laptop and console cable. (much as I'd like some aspects of that job ;-)

"The value of marriage is not that adults produce children, but that children produce adults." -- Peter De Vries

Working...