interface output drops

[lukasz4]

HI All,

I have some concern on congestion that happens on interface. On C3K I can see some output drops which are present on 10 Gbps interface. Utilization of this interface is maybe around 2 Gbps so it's not havely utilized but the output drops are sometimes visible in the "show interface" output.

How can congestion happen here while the interface has more "place" to send more bits ? Why it's dropping the packets while it could send it on a link ? 

I understand that congestion can happen when we are trying to send bits slower that they arrive but if we have packets that travers from 10 Gig to 10 Gig how can this happen ? 

For example if we have 2x10Gig interfaces on the input, that are utilized each receiving 2 Gbps of traffic and they both want to send this traffic to the same output interface at the same time, then the congestion can occur or it should send 4Gbps to the output without any problem here ?

[balaji.bandi]

Can you post the model and IOS

also interface which has drops, if high utilise link some drops normal.

[Georg Pauwen]

Hello,

output drops are often related to microbursts, sudden spikes in traffic that can overload the output queue.

Check if your C3K supports the global command:

qos queue-softmax-multiplier 1200

That usually has an effect on output drops.

[lukasz4]

Yes, I used this command:

qos queue-softmax-multiplier 1200

But it only partially helps. Less drops but still visible.

But my question is more general. Is it possible to have congested interface when it's not highly utilized ? I think this is actually the case for microburst as Georg wrote. I have found following example in the internet:

"Assume that Port1 and Port2 respectively send 5 MB data to Port3 at a line rate of 10 Gbps. The total transmission rate is 20 Gbps. Port3 supports only a rate of 10 Gbps, which is a half of the total transmission rate. It sends only 5 MB data out and buffers the other 5 MB data for transmission later. However, the switch has only 1 MB buffer space. Therefore, 4 MB data is discarded due to insufficient buffer space. Without considering overhead data such as the inter-frame gap, preamble, frame checksum, and packet header, the microburst duration is 4 ms (5 MB/10 Gbps). "

Can you help explain why in such case the total transmision is 20 Gbps and 4 MB will be drop out while we send this on 10 Gbps interfaces ?