I have a question about "priority percent" command.
I want to deploy QoS at customer network to make sure voice traffic with DSCP46 and be treated with highest priority. All the packets makred with DSCP 46 will be put into LLQ on Cisco 3845 router, and then sent out without any latency.
customer connected 10Mbps lease line with a fast ethernet port of Cisco3845, so normal LLQ configuration should not be functional even 20Mbps mixed egress traffic traveling through the lease line, that is because the fa interface is not congested, LLQ configuration is useless in this enviroment.
I replaced the QoS configuration on Cisco3845 router with hiberarchy qos:
priority percent 100
shape average 10000000
Hiberarchy QoS configuration shaped the interface bandwidth to 10Mbps, and permit voice traffic can occupy up to 100 percent of the 10Mbit bandwidth.
I did a test to check if I can change the max bandwidth which voice can occupy, for example, degrease the percent to "40%", and then I used Smartbit, which is traffic generator and analyzer, to inject 2 flows into the WAN link, one marked with DSCP46, and another marked with DSCP0.
The result shows me that percent of bandwidth of voice traffic is not consistent with the setting in "priority percent" command, I mean if I change the setting to 40%, but indeed, the percent of banwidth which voice traffic can occupy is more than 40 percent, it's about 63%. If I change the setting to 1%, the voice traffic can still used around 40% bandwidth.
My question is if the setting of "priority percent" is the real percent(max value) which high priority traffic can used.
Actually, the Smartbit sent two seperated flows, each flow is 20Mbps which can make the 10Mbps WAN link congested.
I can see how many packets got passed and how many packets sent from port per-second on the output of Smartbit, and then I can calculate the percent of throughout.I dida test yestoday that decrease the priority percent at the interval of 10%, but the throughout of voice traffic was not consistent with the change.
So you are sending 2 streams of 20 mbps each enough to congest the 10 mbps. But what I am trying to figure out is that, how are the strems interleaved? are they sent serially and is each packet interleaved? It can so happen that stream A sends a burst of 100 packets and then stream B sends another burst of 100 packets. Together they send 20mbos but it is possible that in the switch queue, there are 100 packets of voice and just a few of data (for one time instant). So I was asking you to send the traffic through 2 different streams. This is just a guess but still worth a try???
I have tested similar configs on 6500 wan cards and have seen 99% + accuracy.
Also clear the counters, and start the traffic for 10 minutes continuously and later check the ratio of traffic recieved if it matches? Just another way to dodge the fluctuations....
The command "priority percent X" allocates a guaranteed bandwidth for your voice service during high load, however if you have low load and more than 60% voice traffic, then voice will be able to use 60% even though 40% is configured. If you like to control the amount of voice traffic sent, then you should use policing together with priority. For example:
priority percent 75%
police cir percent 75 conform-action transmit exceed-action drop violate-action drop
shape average 10000000
Moreover, I would not suggest to use 100%, because you need to take into account of voice control packets, which has by default COS 3, and layer 2 overhead like InterFrameGap,Delimeter etc. Therefore I suggest you to configure not more than 75% of the available link speed for real time traffic.
Furthermore, I also recommend you to checkout the policing functionality of this box by sending streams from your testgenerator. I found out that device 7204VXR creates a bit bucket based on Ethernet header + IP header + IPPDU = 14Bytes + (46-1500). CRC, IFG, Delimiter are not counted. However, device Cisco 7609 takes all into account.
Alco could you just compare the ratios of the 2 streams at the destination rather than calculating the throughput according to the frames sent and recieved? The priority % means the % of the bandwidth available of the total and doesnot refer to % of throughput. % of throughput will depend on the amount of traffic you are sending.
We are pleased to announce availability of Beta software for 16.6.3.
16.6.3 will be the second rebuild on the 16.6 release train targeted
towards Catalyst 9500/9400/9300/3850/3650 switching platforms. We are
looking for early feedback from customers befor...
Introduction Featured Speakers Luis Espejel is the Telecommunications
Manager of IENova, an Oil & Gas company. Currently he works with Cisco
IOS® and Cisco IOS XE platforms, and NX to some extent. He has also
worked as a Senior Engineer with the Routing P...
In this session you can learn more about Layer 3 multicast and the best
practices to identify possible threats and take security measures. It
provides an overview of basic multicast, the best security practices for
use of this technology, and recommendati...