[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [ns] Satellite question



On Mon, 20 Mar 2000, Andrea Detti wrote:

> I am a novice user of ns satellite part.

me too.

> I have done a simulation with the following topology:
>     2 satellite terminal (sender and receiver)
>     1 polar satellite (LEO)
> I report that if a terminal doesn't see the satellite the generated packets
> are dropped.

If the terminal does not have a valid route to the destination the
packets are dropped. So, if your destination terminal isn't connected
to an in-view satellite either, packets will be dropped immediately at
the source instead of making their way across a mesh of ISLs to the
vicinity of the last-connected satellite, say.

> It seems that, there isn't a queue that holds the packets until the
> satellite is able to receive.
> Is it true ?

That's the case; continuous coverage and something to hand over to is
presumed for polar satellites. Changing this may be undesirable from
an endtoend viewpoint; you could end up holding packets for virtual
hours and simulating infinite ARQ, which rather gets in the way of
sensible TCP-over-satellite simulation. Unless you're deliberately
trying to simulate AX.25, of course...

Watch for packets vanishing on cross-seam ISLs during ISL
handover; that's imo slghtly more problematic.

L.

<[email protected]>PGP<http://www.ee.surrey.ac.uk/Personal/L.Wood/>