[MLB-WIRELESS] Nasa Tv over MW

Kim Hawtin kim.hawtin at adelaide.edu.au
Thu Sep 20 09:50:21 EST 2007


hi guys.

Rob Wise wrote:
> On 19/09/2007, Mike Dickens <michaeldickens at gmail.com> wrote:
>> "More importantly however what would happen if say 100 people all wanted
>> to watch NASA TV at the same time on the network eg due to a shuttle
>> launch? Would the streaming tv have to be carried 100 separate times
>> across the network thus overloading the system or is there some way to
>> distribute one signal across the network and people can just passively
>> receive the streamed packets?"
>> It'd have to be carried 100 times. That being said, it could be 'cachied' or
>> 'proxied' on more localised nodes (probably the closest unix routing box, if
>> it took off), so your end would only be transmitting it a few times then the
>> node closest to the user would deal with it.
> 
> Actually if it was broadcast over MW using Multicast then only 1
> stream would be carried over each network link and then broken out
> locally to x number of local viewers.  More info on multicast is
> available at:
> 
> http://en.wikipedia.org/wiki/Multicast
> 
> This is how Pay TV broadcasts over IP networks operate.  The same
> technique is often used in hotel TV systems too.  It's probably worth
> noting that many if not all of the current MW nodes probably wouldn't
> handle multicast distribution on their present hardware.

there are a number of problems with multicast.

mainly about half the AP kit out there doesn't broadcast it over its
wifi interface ... ever wondered why OSPF won't work?
also most APs don't rebroadcast it back out the wifi interface to the clients
now your linux box or .*bsd box might get it right ;) but in the big mix ...

its also an arse to configure. proxies and/or streaming servers are a
lot easier to configure. it can also be done from a server and not the
router device. that way you only have to consider the security of the
router as a router, and the server as a server and not necessarily mixing
everything up.

regards,

kim
-- 
Operating Systems, Services and Operations
Information Technology Services, The University of Adelaide
kim.hawtin at adelaide.edu.au

"I am not a finite state machine"



More information about the Melbwireless mailing list