Dr. Vinton Cerf - Internet, Infinity and Beyond - Page 12

Pages:  Previous    1  2  3  4  5  6  7  8  9  10  11  12

Let me conclude my formal remarks with an update on the interplanetary extension of the Internet. The first thing I want to make clear is that this is not a Google project so I don't want you all running out of the room saying that Google's business model is to take over the solar system, that's not what this is about.

This is really about instrumenting and standardizing communications in space now we've done a pretty good job of instrumenting Earth and we have to work even harder now that global warming has become such a threat. Its really critical for us to understand what's happening on planet Earth and for similar reasons we’re quite interested in understanding what has happened in the past or what might happen in the future in other planets like Mars.

We've been landing on Mars since 1976 with the Viking project and in fact we've used something called the Deep Space Network which was built in 1964 by the Jet Propulsion Laboratory in order to communicate with spacecraft that are in orbit around planets or flying past the asteroids. Some of you will know that the space network has a major node in Canberra (Australia), it has another major node in Goldstone (California) and it's third major noted is in Madrid (Spain) and so as the Earth rotates there’s always one of these big 70 metre dishes in view that can see out into the solar system to communicate to these various spacecrafts. Some of the spacecrafts actually landed on the surface of the planet this is the Pathfinder that landed on Mars in 1997 and sent back a substantial amount of very interesting imagery. This is one of the rovers landed in 2004 it's a much larger thing than the Pathfinder was and 150 kg and stands about four and a half feet tall.

The two rovers landed on Mars in early 2004, the original mission planned called for 90 days of operation. They’ve actually have been in operation for three years now.

There's a reason for that if you look at this more detailed picture of the rover that you'll see that the black winglike things are solar panels those panels convert sunlight into electricity which charges batteries and batteries of course help all of the instruments on board the Rover to function and also to move around.

The original expectation was that Mars been rather a dusty place would eventually deposit so much dust on the solar panels that one day it would not convert enough sunlight into electrical energy and the battery would not be sufficiently charged enduring a long Martian night, that the instruments would die and they would not reignite so to speak, when the sun came back up because the battery had not been sufficiently charged. Well the fact is that the solar panels have remained relatively clean, personally I think that there's somebody up their dusting them off but we actually haven't caught them on video so I can’t prove my theory; there is another theory which is probably more realistic and that is that there are little wind storms that take place, there little dust devils that take place on Mars and those dust devils actually blow the dust off of the solar panels you can see it if you sitting in the operations centre in Pasadena looking at the signal telemetry that's coming back saying what the voltage levels is on the solar panels you can see a sudden spike when the wind blows the dust off of the solar panels.

So these things have been running for almost three years now; one of them (I forget which one) now has a broken wheel so it’s dragging furrows in the Martian sand. But there's still an enormous amount of data coming back which raises a very interesting other issue associated with the data communication. There are now four orbiters around Mars although the Mars global surveyor has spoken its last as near as we can tell.

The newest one is the Mars reconnaissance orbiter and Mars Odyssey is very critical now to the rover operation. It turns out that the radios on the rovers were originally intended to deliver data directly back to Earth through the Deep Space Network. So this was a direct to earth transmission path. The data rate which could be sustained over the long distances between Earth and Mars is about 28 kilobits per second so it's already pretty impaired in terms of data rate. When they turn that radio on they discovered after about 20 minutes of use it overheated and had to turn it off from keep it from damaging itself.

Which meant that not only did have a low data rate but you had a low duty cycle and it was hard to deliver all the data that the rovers were accumulating until someone got the bright idea that the orbiters were programmable, had processing power, had memory and had radios.

So they reprogram the rover to use a different radio operating at 128 kilobits per second to transmit data up to the orbiter and then the orbiter as it came around in its orbit would then squirt the data again at 128 kilobits a second to the deep space network. So 95% of all the data is now being stored on the Mars Odyssey and is then forwarded as the Odyssey gets into the field of view of the Earth Deep Space Network.

So we’ve been thinking since 1998 that we should do this sort of thing as a standard as opposed to a special hack which is what they did in the case of the rovers. What we started out with was the belief that we could use the standard Internet protocols to do this deep space communications and that lasted about a week and those of you have some background and protocol design will instantly understand what the problem was. The first problem is the distance between the planets is literally astronomical.

At the speed of light between Earth and Mars when they are closest together it takes three and a half minutes for a signal to go at the speed of light between Earth and Mars and a similar delay to come back; when they’re farthest apart in their orbit thats about 235,000,000 miles (approx 378,200,000 kms) apart that's 20 minutes one-way and 40 minutes round-trip time.

The TCP protocols do a real simple flow control when you're in outer space, you say stop. Now on Earth the that's not too bad because it's only a few hundred milliseconds at worst before the other guy hears you say stop and shuts up; of course you have to buffer some of the stuff that has been sent between the time you said stop and the time that the other guy hears its. But suppose it's a 40 minute round-trip time, if you say stop it takes 20 minutes before the other guy hears you and in the meantime of course stuff is coming at you that you don’t know what to do with and its falling on the floor.

There’s another problem thats called celestial motion. You have a rover sitting on the surface of the planet you’re talking to it and then the planet rotates and pretty soon you can't talk to the rover anymore because it's on the wrong side of the planet and you have to wait til it comes back around again just like you might have to do with an orbiting satellite. So celestial motion is disruptive, it breaks the connectivity.

So we realized after a while that our Interplanetary network protocols had to be revised and not just use TCP/IP and that they were a special case of a much more general case of protocols that we call delay and disruption tolerant networking. We went to the Defence Department and said we think you need this capability for tactical communications and they agreed so we’re now testing the interplanetary protocols with the Marine Core in a terrestrial environment to see whether or not they are more robust in the presence of disruption and delay.

We’re also testing this in sensor networks that are being implemented to gather data currently terrestrially but in the future on other planets, so we are very far along now in trying out these interplanetary protocols we’re going to put some equipment on board the international space station later this year assuming the schedule holds to test a space qualified version of the delay and disruption tolerant networking protocols and we've been told by NASA that if the rovers stop moving but they’re still functional that we may be able to re-programme them to run the full interplanetary protocol suite and have some tests done with those rovers.

In the meantime the Mars reconnaissance orbiter has already been programmed with the lower two layers of the interplanetary standard and after it has completed its mapping mission in a couple of years time will put the rest of the protocols suite onboard so that we will have effectively at the end of the decade a two planet Internet in operation.

Now over time what we hope will happen is that NASA and the other the space agencies around the world will standardize on DTN protocols in the same way that they've standardized on TCP/IP and every mission that gets launched will simply carrying the standard protocols onboard so that when a new mission is launched if there still assets leftover from previous missions they can be integrated into the support systems and used as part of an increasingly large store and forward backbone.

So we're not trying to build this gigantic interplanetary backbone network and then hoping that someone will come what we're saying is just standardize the protocols and use them to ecrete a backbone network to support both robotic exploration of the solar system and eventually continue manned exploration so that's the up-to-date status on the interplanetary network.

Pages:  Previous    1  2  3  4  5  6  7  8  9  10  11  12