Computing and reporting latency in priority queues
View Patent ↗Channel traffic in Internet Protocol (IP) networks may be placed in priority queues upon receipt. Providing an efficient way for operators to use the channel may enable the operator to better make decisions for configuring the priority queues, maximizing effectiveness. For example, measuring a queue latency and making a queue latency visible to an operator may enable a better configuration of priority queues. Further, an improved latency metric may assist an operator to judge a relative effect that priority assignments have on traffic over a channel. Such techniques may be applied to an out-of-band (OOB) channel to include an assessment of parameters important for optimizing the configuration of priority queues in a downstream channel. For example, an improved latency metric for assisting an operator to judge the relative effect that priority assignments have on traffic is a parameter assessed for better use of an OOB channel.
1. A method comprising:
receiving a plurality of packets in a burst of packets in a buffer;
queuing the plurality of packets in the burst of packets in at least one queue in the buffer, the at least one queue having a designated priority;
measuring a total latency of the burst of packets representing a total number of intervals over which the plurality of packets are serviced and waiting in the at least one queue prior to packet servicing, wherein a latency in terms of intervals for each packet in the plurality of packets is a number of intervals the respective packet in the burst of packets waits until being serviced, including the interval in which the packet is serviced;
determining an average latency of the at least one queue based on the total latency for the burst of packets serviced over the plurality of intervals divided by a number of packets in the burst of packets serviced during the burst; and
outputting the average latency of the at least one queue.
2. The method of claim 1 , further comprising servicing each individual packet one at a time in an order in which they are received.
3. The method of claim 1 , wherein the burst is an isolated burst and no packets are queued either prior to or following the receipt of the plurality of packets in the isolated burst.
4. The method of claim 1 , wherein a delay in servicing the burst of packets exists at a start of the receiving the plurality of packets in the burst.
5. The method of claim 4 , wherein there are at least two queues each having a designated priority, the at least two queues including a lower priority queue having a lower priority than a higher priority queue, wherein the lower priority queue is not serviced due to the higher priority queue having a packet scheduled for servicing.
6. The method of claim 4 , wherein the average latency accounts for the delay at the start of the burst.
7. The method of claim 1 , wherein a last packet in the burst is delayed by several intervals.
8. The method of claim 7 , wherein there are at least two queues each having a designated priority, the at least two queues including a lower priority queue having a lower priority than a higher priority queue, wherein the delay of the last packet in the lower priority queue is caused by the higher priority queue having several packets which need to be serviced before the last packet in the lower priority queue.
9. The method of claim 8 , wherein the average latency accounts for the delay of the last packet in the burst.
10. The method of claim 1 , wherein the burst occurs at an edge of an interval, and the total latency is attributed to a side of an interval boundary on which the burst occurred.
11. The method of claim 1 , wherein the total latency of the burst of packets is a total latency determined across multiple intervals and is a weighted average based on an aggregate of interval latencies of a total number of packets serviced during the multiple intervals.
12. The method of claim 1 , wherein the at least one queue is a priority queue of a multiplexer device.
13. The method of claim 1 , wherein the at least one queue is a priority queue of a multiplexer-modulator device.
14. The method of claim 1 , wherein the latency of each packet is measured in each interval without regard to when individual packets entered the queue or a duration over which each individual packet is in the queue.