[mythtv-users] multiple tuner question

Tom Trelvik mythtvuser at triple-t.org
Sat Jun 21 11:24:30 EDT 2003


Bruce Markey wrote:
>>    It actually runs over ATM AAL5 (network layer 2 stuff (IP is layer 
>> 3, and TCP is layer 4, for reference)),
> 
> 
> Realize that "TCP/IP network" != "TCP" + "IP" over ethernet.

	Oh of course, I didn't mean to imply that the usage of ATM precludes an 
  IP network sitting on top of it.  I've actually had desktop computers 
doing plenty of TCP/IP on an OC-3 ATM network.  Quite speedy, but far 
more effort than it's worth for desktop connections.

> Layer 1 is coax, layer 2 is ATM with nodes identified by MAC

	Actually ATM doesn't natively use MAC addresses.  Actual ATM addresses 
are 20 byte addresses.  But many means of connecting to ATM networks 
simply involve translating from a device talking ethernet protocols with 
a 6 byte MAC address to native ATM protocols, typically LANE.

> 4. In fact, TCP would be a horrible choice for a massively
> broadcast networks  (imagine the ACK storms!).

	Quite true.  =)

>>     Gregorio also mentioned the A/V side of things using ISO/IEC 
>> 13818-1, which is the "MPEG-2 Transport Stream" protocol (equivalent 
>> to layer 3 IP, if I understand correctly).
> 
> 
> I think that's more like layer 4. 3 would be more like an
> addressing scheme.  

	Yeah, you're right that it's not layer 3, but looking back at that doc 
I'm more confused now.  In some places they refer to it like a layer 2 
thing, and in other places they specifically refer to it as a transport 
layer (4) protocol.  There was little to no reference to layer 3 (any 
protocol) at all, save in the glossary definition of IP.

> I assume that the FAT channels are
> different frequencies so there probably wouldn't be a need
> for an addressing scheme if there is only one unidirectional
> stream of data being broadcast. 

	Yes and no.  The frequency range of the FDC's is within the larger 
frequency range used by the FAT channels.

> Whether or not these packets
> are or aren't encapsulated in IP headers is moot. The data

	Yeah, I just misinterpretted your original statement to mean that the 
cable itself all went over TCP/IP, which is primarily what originally 
piqued my interest.

> could be sniffed. However, I assume that if it could be
> easily decoded, thousands of people would be doing this
> already. I know we're not the first people to think it would
> be cool to grab the mpeg streams ;-). And, I know it would
> be a waste of my time to think I might be the first person
> to figure out how to do it.

	Not to mention PVR developers/users/companies would have a much more 
difficult time defending their legitimacy if they had to figure out how 
to descramble it on the line.  We'd be in a similar boat to Jon Johansen 
for publishing DeCSS, and the DMCA would start biting in our direction a 
whole lot more.

> Yep. That leaves no doubt that there is an IP network.

	Yeah, but just for PC data, and other signalling protocols, not for the 
actual A/V content.

Tom




More information about the mythtv-users mailing list