how brodcast influence L2 Switch

[aviyoshi10]

Hi all,

i need to uderstand the influence of broadcast traffic in regard to the number of active recievers.

for example:

i have server that send broadcast of 5Mbps. there are 10 active users on the same L2 switch that recieves the broadcast. the cpu on that situation is 60%.

if i have the same broadcast stream but now there are 20 active users that need to get this broadcast, should i accept for increase in the switch CPU ?

thanks Avi.

[Sergey Fer]

Strictly saying, it depends on many factors (switch model, type of packets/frames, additional tools you may use or not, ...). In it's basic you won't expect CPU usage increase at all because simple L2 broadcasts must come through hardware-based, not CPU-based path. But life sometimes is a little tricky.

[aviyoshi10]

Hi

The switch model is 2960s.

What do you mean by kind of broadcast packets?

If the frame is all ff:ff, how can the switch knows that this frame is not intend for him?

Sent from Cisco Technical Support iPhone App

[Joseph W. Doherty]

Disclaimer

The   Author of this posting offers the information contained within this   posting without consideration and with the reader's understanding that   there's no implied or expressed suitability or fitness for any purpose.   Information provided is for informational purposes only and should not   be construed as rendering professional advice of any kind. Usage of  this  posting's information is solely at reader's own risk.

Liability Disclaimer

In   no event shall Author be liable for any damages whatsoever (including,   without limitation, damages for loss of use, data or profit) arising  out  of the use or inability to use the posting's information even if  Author  has been advised of the possibility of such damage.

Posting

The number of users receiving a broadcast (not multicast?) stream would only impact the switch's CPU if the CPU were involved with actual replication of the frame to additional active ports.

As how the switch would know a broadcast frame isn't directed to the switch itself, it doesn't which is why it's suggested a switch's management IP be on a non-user VLAN.

[Sergey Fer]

As Joseph said, "The number of users receiving a broadcast (not multicast?) stream would only impact the switch's CPU if the CPU were involved with actual replication of the frame to additional active ports." This means that in a pure environment your switch's CPU does not do anything towards broadcast frames. Even if these frames are addressed for switch itself CPU impact does not depend on number of users. But you need to consider a number of situations:

- previous generation switches (like 2960, 3560, 3750) use internal ring and no special hardware replication engine. Newer switches (3750X, 2960S as a higher models like 4500) use fabric-based architecture and a special replication engine within. Here (in fabric-based switches) you might see some sorts of CPU impact especially in old K2-based 4500s. I've seen that some times.

- there are may way of improper broadcast usage. For example you may want to encapsulate IP multicast into Ethernet broadcast. Why not? Though it is not proper usage, it is not forbidden at all. If your switch is also IP multicast router, then (it is IOS-dependant) multicast replication will be performed by CPU.

- There are not only IP and Ethernet in the world of nets. Ethernet frame contains EtherType field which shows a host what type of protocol is contained inside. I do not know what does a switch make towards a frame that is a broadcast but, for example, is of STP EtherType. Really it is also IOS-dependant. You might also think about Gratuitous ARPs or alike features.

- And, like Joseph said, it is not a good idea to have Management IP of a switch inside user VLAN.