«

I guess this one sounds a little foolish, but still: I would like to
understand the way devices treat multicast and broadcast frames on a
segment.

Let's assume a segment with 2 routers and 2 PCs.
Now let's assume that RIPv1 is enabled. Since it broadcasts its routing
table, all the rest of the devices on the segment will receive these
broadcasts.
In the case of RIPv2, multicast is being used.
Since the hub (or even a switch that floods multicast) will transmit the
information to the rest of the segment, all the devices, PCs and
routers, with receive these fames too.
So in both cases, all the devices receive these frames.
What's the difference, what am I missing here?

Thanks, Yaron

"Yaron" <> wrote in message
news:...
> I guess this one sounds a little foolish, but still: I would like to
> understand the way devices treat multicast and broadcast frames on a
> segment.
>
> Let's assume a segment with 2 routers and 2 PCs.
> Now let's assume that RIPv1 is enabled. Since it broadcasts its routing
> table, all the rest of the devices on the segment will receive these
> broadcasts.
> In the case of RIPv2, multicast is being used.
> Since the hub (or even a switch that floods multicast) will transmit the
> information to the rest of the segment, all the devices, PCs and
> routers, with receive these fames too.

RIP2 can run "compatibility mode" where it uses broadcasts rather than
m/cast.
> So in both cases, all the devices receive these frames.
> What's the difference, what am I missing here?

adevice with a reasonable interface chip can program the chip with the set
of multicast MAC addresses it is interested in, and ignore the rest in
hardware.

All broadcasts traditionally need the CPU to examine each packet to decide
whether to ignore it.

some layer 2 switches can use multicast control to limit where the packets
go within a subnet, reducing the average bandwidth overhead per port.

doesnt matter for a few packet / sec for RIPv2, but once you start sending
multiple 5 Mbps MPEG2 video on multicast it gets more important......
>
> Thanks, Yaron
--
Regards

- replace xyz with ntl

Yaron -

On a single segment like you mention, there isn't really that much of a
difference - other than the address that a packet is multicast to or
broadcast to (multicast will always be in the range of 224.0.0.1 -
239.255.255.255). However, your PC's will typically have to have an
application that uses multicast for them to respond / receive those
packets. Something like the Nortel Symposium product, which allows
certain users to view real-time data on PBX call routing.

When you get to multiple segments or even multiple subnets, then you
get into the real guts of multicast and how it works. Because then,
communication between nodes and network devices allows a multicast
server application (such as an Anti-virus app) to send feeds to only
the hosts who need them. A multicast group can exist with a single
multicast host on every subnet in a large vlan environment and each of
those hosts (and only those multicast hosts) will receive the server's
traffic.

This is a quick and dirty high-level overview, but I hope it helps. The
Cisco Press books "Routing TCP/IP" volume 1 & 2 are good books with a
large section on multicast.

Thanks for your answer, much appriciated.

V. Evans:
> Yaron -
>
> On a single segment like you mention, there isn't really that much of
a
> difference - other than the address that a packet is multicast to or
> broadcast to (multicast will always be in the range of 224.0.0.1 -
> 239.255.255.255). However, your PC's will typically have to have an
> application that uses multicast for them to respond / receive those
> packets. Something like the Nortel Symposium product, which allows
> certain users to view real-time data on PBX call routing.
>
> When you get to multiple segments or even multiple subnets, then you
> get into the real guts of multicast and how it works. Because then,
> communication between nodes and network devices allows a multicast
> server application (such as an Anti-virus app) to send feeds to only
> the hosts who need them. A multicast group can exist with a single
> multicast host on every subnet in a large vlan environment and each
of
> those hosts (and only those multicast hosts) will receive the
server's
> traffic.
>
> This is a quick and dirty high-level overview, but I hope it helps.
The
> Cisco Press books "Routing TCP/IP" volume 1 & 2 are good books with a
> large section on multicast.

Thank you stephen, great response.

stephen:
> "Yaron" <> wrote in message
> news:...
>> I guess this one sounds a little foolish, but still: I would like to
>> understand the way devices treat multicast and broadcast frames on a
>> segment.
>>
>> Let's assume a segment with 2 routers and 2 PCs.
>> Now let's assume that RIPv1 is enabled. Since it broadcasts its
routing
>> table, all the rest of the devices on the segment will receive these
>> broadcasts.
>> In the case of RIPv2, multicast is being used.
>> Since the hub (or even a switch that floods multicast) will transmit
the
>> information to the rest of the segment, all the devices, PCs and
>> routers, with receive these fames too.
>
> RIP2 can run "compatibility mode" where it uses broadcasts rather
than
> m/cast.
>> So in both cases, all the devices receive these frames.
>> What's the difference, what am I missing here?
>
> adevice with a reasonable interface chip can program the chip with the
set
> of multicast MAC addresses it is interested in, and ignore the rest
in
> hardware.
>
> All broadcasts traditionally need the CPU to examine each packet to
decide
> whether to ignore it.
>
> some layer 2 switches can use multicast control to limit where the
packets
> go within a subnet, reducing the average bandwidth overhead per port.
>
> doesnt matter for a few packet / sec for RIPv2, but once you start
sending
> multiple 5 Mbps MPEG2 video on multicast it gets more important......
>>
>> Thanks, Yaron
> --
> Regards
>
> - replace xyz with ntl

CGMP or IGMP snooping on Catalyst Switches.