DoD to Put Internet Router in Space

narramissic writes "ITworld is reporting that the Department of Defense plans to launch in the first quarter of 2009 a satellite-based router to deliver military communications. Satellite operator Intelsat will manage the three-year Internet Routing In Space (IRIS) project, while Cisco will provide IP networking software for the on-board router. After testing, the satellite will be available for commercial use. From the article: 'Potential nonmilitary benefits of the IRIS program include the ability to route IP (Internet Protocol) traffic between satellites in space in much the same way packets are moved on the ground, reducing delays, saving on capacity and offering greater networking flexibility, Lloyd Wood, space initiatives manager in the Global Defense, Space & Security division of Cisco, said Thursday.'"

Re:How long until the rest of the world wants acce

[Anonymous Coward]

Transit times for packets routed through space (to geosynchronous satellites) is much much larger than those routed over the surface.

Think seconds, not milliseconds.

Re:This can't be the first time

[Anonymous Coward]

Well... there is this from 2005: http://www.gcn.com/print/24_21/36507-1.html [gcn.com] [gcn.com]

NASA has been investigating using TCP/IP for communications with satellite since at least 2000 ... http://ipinspace.gsfc.nasa.gov/documents/OMNIconce [nasa.gov] pt.pdf [nasa.gov]

and it was going so well that http://www.military-information-technology.com/art [military-i...nology.com] icle.cfm?DocID=998 [military-i...nology.com]

Not that new based on a quick google.

Re:Oy, vey...

[torqer]

Lag is exactly the reason why they want to put the satellite there. It will route traffic between satellites in space, without need for the lag of travelling to terrestial router and then return to space.

Not realtime, but 2x speed of existing system

[davidwr]

TFA said the existing system involves

source - satellite#1 - ground-based router - satellite#2 - destination

The new system will be
source - satellite #1 - space router - satellite #2 - destination

or even better

source - space router - destination

Depending on where the satellite is, you may have just shaved a few tenths of a second off the one-way trip.

Re:Oy, vey...

[arthurpaliden]

Realtime means predictable delay not no delay.

Not sure if DirectPC is a router or not

[davidwr]

Satellite Internet can work two ways: The satellite is the router, or the satellite is a bridge or medium-converter box just like some cheap DSL or cable modems.

If I were DirectPC I'd want my satellite to be as simple and lightweight as possible.

Is this like packets being routed like usa-satellite-satellite-satellite-africa?

See the article for an explanation.

The long and short of it is this is designed to replace traffic that is currently going
somewhere-satellite-ground based router-satellite-somewhere
where the "somewheres" on either end are optional.

Re:What do I know about it?

[Skippyboy]

I work at Johnson Space Center and we are testing this right now. The idea is that in the future (ie when we have a moonbase, etc) that all communication will be in IP packets over RF links.
Currently the RF links use multiplexed frames with different PN sequences and frame sync headers to communicate, so the position of each bit within the frame means something.
With IP packets, we wouldn't really have to decode/demux the frames to get the information. Each entity could send data based on its IP address. As mentioned before - the lag time issue is gonna be pretty messy, unless we used UDP or something similar. We are just in the beginning stages right now for our purposes, so just configuring the routers and getting the data into an RF link and be errorless is what we are fighting.
Hope that sheds some light on why. Also - the frequencies we are talking about are going to be S, K, Ku, Ka, and higher, so it isn't likely that a script kiddy will have access to that kind of equipment. Also - the links will be encrypted and PN spread, making it less likely to be intercepted/hax0r'd...

Those numbers aren't even close to right.

[User 956]

You might try actually fact-checking those numbers. [globalsecurity.org] The actual number of US casualties in Iraq is almost 3,300, with another 23,000 wounded.

And yes, "wounded" includes losing limbs, eyes, and all sorts of other body parts that don't, on average, sustain major damage when you're out and about in Washington DC.

Re:Stats??

[beartrash]

"We are talking about Internet router here. this will just be another router .. only special thing is the location (i.e. space)> I guess it will use BGP as all the other routers in todays Internet."

How many paths out will this router have? If it's only one or two, statically route it or use a couple of differently weight static routes and IGP wouldn't make much sense in most likely scenarios, maybe with NAT if a few nodes are going to access it.

If an IGP is necessary (if there are eventually several "space routers" and a few paths back to earth, EIGRP would make most sense, since Cisco is providing the hardware (EIGRP is a Cisco proprietary protocol) and EIGRP provides the most knobs to tweak to overcome latency and modify routing announcements.

BGP would only make sense if more than one ISP ends up providing connectivity in space (so you can easily multi-home) or we end up with several space agencies up there. We can't share an IGP with the ESA, JAXA or RKA.