07-09-2015 03:06 AM - edited 03-08-2019 12:54 AM
Hello all,
my question is, is the internal buffer for any give interface the same for Upstream and Downstream traffic, in regards to that interface?
Or is the internal buffer of the receiving interface holding the packet until the sending interface can handle it?
The question originates from a Cisco 2950 series switch, but I imagine the implementation for this would be the same for all switches.
An example to make the question clearer:
A specific interface, let's call it Fa1, has very high traffic, to the extend that 50% of the packets are Ignored.
A packet comes to interface Fa2 that needs to leave the switch through Fa1, will this packet wait in Fa2's internal buffer until Fa1 is able to handle it, or will it be ignored since Fa1 has used up all of it's resources?
I understand that the original question and the example are slightly different and can have different explanations, so please advise how you can.
Thanks in advance,
Best regards,
Orestis
07-09-2015 05:13 AM
Hi
you could take a look at these links below basically if buffers fill up and there's no where to hold the packet and the input queue is full your going to drop packets , the packets will only wait if there's space in the buffer or hold queue
http://www.cisco.com/c/en/us/support/docs/routers/10000-series-routers/6343-queue-drops.html
http://www.cisco.com/c/en/us/support/docs/interfaces-modules/channel-interface-processors/14620-41.html
07-09-2015 05:32 AM
Hello Mark,
thank you for the response.
I went through the document again, and I now understand that the switch has a pool of buffers that are allocated freely between interfaces by the Routing Processor, and are not preallocated to specific interfaces.
If this is true, then my question is baseless.
Thank you for the response,
Best regards,
Orestis
07-09-2015 05:37 AM
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
Ingress/egress queuing buffering features, depends on the device. Some devices allow tuning of one or both. Most devices do not congest on ingress, even when egress is congested. Generally packets will be dropped for lack of egress buffering. However, you might see ingress queuing when packets need to be processed switched.
Even when buffers can be adjusted, just increasing available buffers might not be an ideal solution. Yes, it avoids immediate packet drops, but the additional queuing latency might be as bad or even worse.
BTW, don't misunderstand, insufficient buffering can be very adverse to your traffic too, but too much buffering can be almost as bad as too little. Often your goal with buffer tuning (when it's available) is to maximize "goodput".
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide