Why cisco switches have many mac-addresses???

Gngogh

Hi, im trying to find the reason why cisco switches have a base mac-address and one for each port?? Can anyone help me to understand this..

Hondabuff

On is the base hardware address that is the main mac address used for functions such as STP election. Each port gets its own for L2 operation.

Gngogh

and why do they need a mac-address for L2 operation??? if other vendors have only 1 mac-address per switch (im i right??).

OfWolfAndMan

Do you know what the purpose of a switch is? I would look into that before asking a question like that, especially if you're certified already.

Hondabuff

Gngogh wrote: »

and why do they need a mac-address for L2 operation??? if other vendors have only 1 mac-address per switch (im i right??).

That's how equipment talks to each device on the LAN by Mac addresses. If they need to reach a device on another network then it goes to L3 operations at the router. Once on the router then the router looks for the IP address and if it doesn't have it, it is sent out to the next hop "default route" It is sent out to the next hop by Mac address and not the IP. I also remember is was something to do with the Cisco Express Forwarding function for a faster backplane with the Asci chip. That's what makes Cisco switches Bad A$$.

This was the explantion on a Cisco forum.

For L2 switches, there are mac addresses for different functions (to be able to communicate (when we want to telnet, ping or do snmp to the switch) and spanning tree (to form the bridge ID). On some manufactures switches, they use 1 mac address for all functions, Cisco does not. On the 4/5/6x00 devices, there are 1024 mac addresses assigned to the upervisor (1 or more for the switch, 1000 for spanning tree (PVST, each instance has its ' own mac address)). To see the addresses on those switches, do a show module. On the XL series, there is a range, but it seems much smaller (at least on my 2916XL), if you do a show mac self, you would see the range of addresses assigned to the switch. It will use the first one in the range for it self and as part of the bridge id for spanning tree. For Cisco, they are burned in an eprom...but it varies on platform as to were they are located (mother board, supervisor or in the case of the 5500, on the backplane).
Catalyst 6000 family switches have a pool of 1024 MAC addresses that can be used as bridge identifiers for VLANs running under PVST+ or for MISTP instances. You can use the show module command to view the MAC address range.

Gngogh

OfWolfAndMan wrote: »

Do you know what the purpose of a switch is? I would look into that before asking a question like that, especially if you're certified already.

at a hardware level i do think switch use mac-addresses to map its own ports.. port numbers are for humans.. and this is why i asked the question to be sure of that..

xnx

I recommend you re-read your CCENT book, properly

Gngogh

xnx wrote: »

I recommend you re-read your CCENT book, properly

Maybe im not the only one who needs to read the book.. to understand when a switch use its own mac-address, when they are not needed for switching. As far as i know the CCent book doesnt talk about the control plane.

Dieg0M

Switches have different mac-addresses for each of their physical interfaces. Within a subnet or broadcast domain, every mac-address must be unique. As for the virtual interfaces like the VLAN's on a switch, it will depend on the platform and vendor. A lot of the vendor's have the same mac-address for every virtual interfaces. This is possible because each VLAN is a seperate broadcast domain so there is no possibility of overlap or duplicate mac-address.

EdTheLad

Hi Gngogh, you seem to have gotten a hard time for a valid question. STP uses port-id as a tie breaker, the port-id consists of a combination of the port priority and interface-id. Due to needing a unique interface-id for STP, Cisco paired this interface-id with the base mac and hence each interface gets a unique mac-address. STP aside there would be no issue using the one mac across all ports just like what happens with SVIs on some platforms.

xnx

My previous reply may have been pretty rude as I can't remember Port MAC addresses specifically being mentioned in the OCG, maybe just as a St tiebreaker. It's something that you'll definitely know if you're studying CCNP switch.

davenull

EdTheLad wrote: »

Hi Gngogh, you seem to have gotten a hard time for a valid question. STP uses port-id as a tie breaker, the port-id consists of a combination of the port priority and interface-id. Due to needing a unique interface-id for STP, Cisco paired this interface-id with the base mac and hence each interface gets a unique mac-address. STP aside there would be no issue using the one mac across all ports just like what happens with SVIs on some platforms.

I was under impression that a Port ID had nothing to do with mac addresses. It is just a combination of port priority and port identifier (which may not even match the port number):

Spanning Tree Protocol priorities - PacketLife.net
Spanning Tree Port Priority and Port ID Values

@OP
You asked a valid question - it is not covered in any CCENT/CCNA material, and I'm curious about it as well.

OfWolfAndMan

I think I misunderstood your question initially. Check out this thread, post four.


https://learningnetwork.cisco.com/thread/3451

Hope this helps!

davenull

That's a good explanation in the link at post 4. It makes sense for a L3 switch to have mac addresses for all ports, but what about L2 switches? I just checked my 2950 and all ports indeed have their own mac addresses.

Gngogh

I've digging around on this, and i would like to say that my initial question was not the best one, what i really wanted to know is when does a switch uses the mac-address of the interfaces.. i have to apologise for my silly question.
Anyway it is clear that a switch needs a mac-address when is sending traffic to the control plane, and when in case of layer 3 switches and SVI we configure one. i have tested it with wireshark.