Jim Henderson wrote:
> Not really, a party line is many-to-many, but multicast is one-to-many.  

If someone in a multicast domain sends a multicast packet, it'll go to 
everyone on the LAN and downstream from there.

But the real answer is that IP multicast doesn't describe how to route the 
packet other than point to point. IP multicast is between routers. The 
mechanism for getting from the last router to the individual host clients 
isn't specified and depends on the underlying technology. If the underlying 
network between client machines on the subnet doesn't do true broadcasting, 
you're going to wind up repeating the packet on the subnets too.

So it's not really a valid comparison.  IP broadcast only works when the 
underlying network supports it, so it's still point to point.

-- 
Darren New, San Diego CA, USA (PST)
    Ada - the programming language trying to avoid
    you literally shooting yourself in the foot.

Post a reply to this message

>> Last time I checked, ISDN is point-to-point only
> 
> So is TCP.  It's connection-oriented, except for the D-channel, which is 
> datagram oriented.

My point being, ISDN lets you put a cable between two nodes. If you want 
to talk to a different node, you have to plug in a different cable.

>>>> Sure, you don't usually talk to them directly; usually you use one 
>>>> of the IP addresses from the block assigned to your ISP.
>>>
>>> Ding ding! Guess what? They aren't assigned centrally. Your ISP 
>>> assined you yours, mine assigned me mine.
>>
>> And where does the ISP get their IP block from?
>>
>> Yeah, exactly: They're assigned centrally.
> 
> Nope. They're assigned heirarchically. That's completely different.

I don't see how...

-- 
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*

Post a reply to this message

Orchid XP v8 wrote:
>>> Last time I checked, ISDN is point-to-point only
>>
>> So is TCP.  It's connection-oriented, except for the D-channel, which 
>> is datagram oriented.
> 
> My point being, ISDN lets you put a cable between two nodes. If you want 
> to talk to a different node, you have to plug in a different cable.

So there's only one possible ISP you can use over ISDN? There's only one 
person you can call on ISDN?

Since one of the nodes on the cable is a switch connected to a world-wide 
network, I don't really see your point.  Every network that's just a wire 
works this way. IP over CAT5 ethernet only lets you talk between two nodes, 
and if you want to use a different node, you have to move a wire. That's why 
there's such a concept as a default route.

>>> Yeah, exactly: They're assigned centrally.
>> Nope. They're assigned heirarchically. That's completely different.
> I don't see how...

Can't really help you with the english, there.

-- 
Darren New, San Diego CA, USA (PST)
    Ada - the programming language trying to avoid
    you literally shooting yourself in the foot.

Post a reply to this message

On Fri, 14 May 2010 20:00:11 -0700, Darren New wrote:

> Jim Henderson wrote:
>> Assuming the switch isn't multicast aware.  I wouldn't be surprised if
>> some were these days (but I haven't looked at it recently).
> 
> Um, no. Even so.
> 
> If you have one cable coming into the university campus, and a network
> for each building, the router is going to have to send the packets to
> each building, duplicating the packets, regardless of how "aware" anyone
> is.

Yes, but that's not a 1:1 transmission (compared to the receiving 
clients).

>> The difference between, for example,
>> using unicast to push an image down to 15 workstations on a subnet and
>> using multicast to push an image down to 15 workstations on a subnet is
>> a significant reduction in overall network traffic.
> 
> Only for that one subnet.

The router is part of the subnet, though.  It retransmits upstream if the 
TTL hasn't expired (and it's configured to do so).

> It's better than a unicast stream to each destination, yes. That doesn't
> mean IP isn't point-to-point, if you factor in the broadcast address as
> meaning "one subnet is the point."

That's a bit like saying shouting is point-to-point, though, because the 
target "point" is a group.

Jim

Post a reply to this message

On Fri, 14 May 2010 20:01:35 -0700, Darren New wrote:

> Jim Henderson wrote:
>> Not really, a party line is many-to-many,
> 
> Er, not really, no. A party line is one switch to many phones.
> 
>> The routers act as repeaters, but still operate a multicast "domain"
>> (if you will) within their local network, as long as there's a client
>> that has registered with the router to listen in on the multicast.
> 
> Yes, I know how it works. Now we're just arguing over analogies.

We tend to do that. ;-)

> My point is that it's *not* like a radio broadcast. It's point-to-point
> between routers. The only place you're saving bandwidth is on subnets.

Which is kinda the important point.

Jim

Post a reply to this message

Jim Henderson wrote:
>> If you have one cable coming into the university campus, and a network
>> for each building, the router is going to have to send the packets to
>> each building, duplicating the packets, regardless of how "aware" anyone
>> is.
> 
> Yes, but that's not a 1:1 transmission (compared to the receiving 
> clients).

I don't know what you mean.

For one thing, modern ethernet (the kind that goes over CAT5 instead of 
coax) is indeed 1:1 on every part of the network. Nothing I send from my 
computer to yours does not go through the router.

You need to duplicate the packets at least once per physical wire. Same as 
any network. IP doesn't force you to duplicate anything less.

Granted, if you're broadcasting from one point, the broadcaster only has to 
send one packet per physical connection regardless of the number of 
recipients. But that's the same with ISDN and all those technologies too. 
You call into a conference bridge and the bridge duplicates the data as 
necessary. You're not sending out the same data once for each receiver there 
either.

>>> The difference between, for example,
>>> using unicast to push an image down to 15 workstations on a subnet and
>>> using multicast to push an image down to 15 workstations on a subnet is
>>> a significant reduction in overall network traffic.
>> Only for that one subnet.
> 
> The router is part of the subnet, though.  It retransmits upstream if the 
> TTL hasn't expired (and it's configured to do so).

Firstly, if your subnet isn't broadcast, it doesn't save you any bandwidth. 
The router still needs to duplicate all the packets. If your subnet *does* 
support broadcast (like coax ethernet, or alohanet, or something like that) 
then sure, you can broadcast to everyone, but that's because *any* 
transmission goes to everyone, taking up your bandwidth even if it's not 
addressed to you.

>> It's better than a unicast stream to each destination, yes. That doesn't
>> mean IP isn't point-to-point, if you factor in the broadcast address as
>> meaning "one subnet is the point."
> 
> That's a bit like saying shouting is point-to-point, though, because the 
> target "point" is a group.

Broadcast networks tend to be very small, and they tend to not allow two 
conversations at once. You're not saving bandwidth by broadcasting. You're 
just using it more efficiently. A targeted communication on a broadcast 
network still takes up everyone's bandwidth. They just ignore it. Multicast 
on a broadcast network isn't saving bandwidth as much as it's not wasting 
bandwidth.

Wireless is about the only common broadcast IP-based network still around, 
tho, exactly because it's so bandwidth-inefficient and impossible to squelch 
when something goes wrong.

-- 
Darren New, San Diego CA, USA (PST)
    Ada - the programming language trying to avoid
    you literally shooting yourself in the foot.

Post a reply to this message

Jim Henderson wrote:
>> My point is that it's *not* like a radio broadcast. It's point-to-point
>> between routers. The only place you're saving bandwidth is on subnets.
> 
> Which is kinda the important point.

Except even there, you're only saving bandwidth on broadcast networks, and 
on broadcast networks, everyone shares the same bandwidth, so even TCP 
packets go to everyone on the subnet.

The whole point of using multicast instead of broadcast is what it does 
between routers, not what it does on a subnet.

-- 
Darren New, San Diego CA, USA (PST)
    Ada - the programming language trying to avoid
    you literally shooting yourself in the foot.

Post a reply to this message

Le 15/05/2010 22:02, Darren New nous fit lire :
> Jim Henderson wrote:
>>> If you have one cable coming into the university campus, and a network
>>> for each building, the router is going to have to send the packets to
>>> each building, duplicating the packets, regardless of how "aware" anyone
>>> is.
>>
>> Yes, but that's not a 1:1 transmission (compared to the receiving
>> clients).

>>>> The difference between, for example,
>>>> using unicast to push an image down to 15 workstations on a subnet and
>>>> using multicast to push an image down to 15 workstations on a subnet is
>>>> a significant reduction in overall network traffic.
>>> Only for that one subnet.
>>
>> The router is part of the subnet, though.  It retransmits upstream if
>> the TTL hasn't expired (and it's configured to do so).
> 
> Firstly, if your subnet isn't broadcast, it doesn't save you any
> bandwidth. The router still needs to duplicate all the packets. If your
> subnet *does* support broadcast (like coax ethernet, or alohanet, or
> something like that) then sure, you can broadcast to everyone, but
> that's because *any* transmission goes to everyone, taking up your
> bandwidth even if it's not addressed to you.
> 
[snip]
> Broadcast networks tend to be very small, and they tend to not allow two
> conversations at once. You're not saving bandwidth by broadcasting.
> You're just using it more efficiently. A targeted communication on a
> broadcast network still takes up everyone's bandwidth. They just ignore
> it. Multicast on a broadcast network isn't saving bandwidth as much as
> it's not wasting bandwidth.
> 
> Wireless is about the only common broadcast IP-based network still
> around, tho, exactly because it's so bandwidth-inefficient and
> impossible to squelch when something goes wrong.
> 

And then came in play the (G)MPLS system(s), allowing to reserve
bandwidth, having "routers" to perform split/duplication of data, and
putting in place minimalist-covering trees for optimal diffusion (and
even worse: you can have 1:1 protection, or even 1+1 !) adapting in
nearly realtime to the destinations interested in the "channel" and
reacting to the network events too.

Post a reply to this message

On 10-5-2010 10:12, Invisible wrote:
>> Or Carl Sagan, for that matter.
> 
> Who?

See also Pioneer WIP in p.b.i

Post a reply to this message

>>>> Last time I checked, ISDN is point-to-point only
>>>
>>> So is TCP.  It's connection-oriented, except for the D-channel, which 
>>> is datagram oriented.
>>
>> My point being, ISDN lets you put a cable between two nodes. If you 
>> want to talk to a different node, you have to plug in a different cable.
> 
> So there's only one possible ISP you can use over ISDN? There's only one 
> person you can call on ISDN?
> 
> Since one of the nodes on the cable is a switch connected to a 
> world-wide network, I don't really see your point.

My point is that ISDN is a point-to-point system, so it doesn't have a 
concept of network addresses. When there are only 2 nodes in the system, 
you don't need addresses; if machine A talks, it can only be talking to 
machine B.

Having a global network without addresses doesn't really work.

And that's kind of my point: All the other network systems were designed 
for networks that are "small", where it's feasible to keep track of all 
the addresses by hand. None of them really supported having several 
networks connected together, managed independently.

Post a reply to this message