cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
3238
Views
5
Helpful
3
Replies

input outputs rate

taufeeq taufeeq
Level 1
Level 1

Dear All,

I issued show interface <int no> On a switch 3750 and I noticed that the Output rate is lesser than Input Rate on switch interface ?? is it normal? Y?
OR
Does it means that the Switch is dropping frames?
I don't see any counters errors or drops . yet y is it that 5min OP rate< 5min IP rate always???


Thanks in Advance?

Regards,
Taufeeq

Sent from Cisco Technical Support iPhone App

3 Replies 3

Joseph W. Doherty
Hall of Fame
Hall of Fame

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

They very often don't agree, as traffic volume is rarely symmetrical.

However, normally the sum of all interface ingress rates will about equal the sum of all the interface egress rates.  (NB: the reason for "about", rates are moving averages.)

If you have a device with just two active interfaces, the ingress rate on one interface will (about) match the egress rate of the other interface.

The foregoing assumes packets aren't being dropped.

Bhishma Khanna
Cisco Employee
Cisco Employee

Hi Taufeeq,

The switch calculates in bits per miliseconds and the output we see is an abstract view of average flow. That is why it is not always equal.

If the input rate is more than the output rate then it is not always dropping the packets because most of the buffering in swithces takes place at the egress and not in the ingress.

Consider drops only if the counters for output drops increases.

Thanks.

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

If the input rate is more than the output rate then it is not always dropping the packets because most of the buffering in swithces takes place at the egress and not in the ingress.

That's an important comment, which is why I mentioned moving averages, but to further clarify this point . . .

Imagine a 10g interface ingress, with a 10 Mbps egress.  The 10g can accept at, of course, 10g, likewise the 10 Mbps can only transmit at 10 Mbps.  So, the 10g can accept and send to the 10 Mbps 1000x faster.  Or in 1 second, the 10 Mbps interface may need to queue 999 frames while it's still transmitting the first frame.

On Cisco devices network devices, generally interface receive and transmit rate stats are 5 minute moving averages, by default, and also generally, by default, cannot be set to less than 30 seconds.  So as the seconds tick by, assuming a 10g interface is accepting traffic at line rate, its receive rate will indeed "ramp up" faster than the 10 Mbps egress rate, which will need to queue or buffer packets.  So, besides the disparity of ingress/egress rates, on the slower egress interface, you should see frames queuing/buffering, and if the queue/buffer overflows, drops.

During shorter time measurements, a situation like the above can show different ingress/egress rates, without drops.  On longer time measurements, the average rates should agree, again assuming no drops.  The latter because the 10g rate will need to receive in bursts, lowering its average ingress rate, while the 10 Mbps egress interface is continuously transmitting to drain its buffers.  I.e. the 10g short term rate will jump up and down with such bursts, but the average of those bursts will agree with the 10 Mbps egress rate; again assuming no drops.

Also note, many switches do not have extremely deep buffering, so in a "real" situation of 10g to 10 Mbps, you're likely to see some drops.  Those devices with deeper buffering capacity, e.g. many routers, can queue transient bursts, but excessive queuing also many create excessive queuing delay.  So often allowing egress drops to happen will signal the sender to slow its transmission rate.