Test Description (as in the previous test)

1. Parameters:
   • with/without best-effort traffic
   • sheduling alogrithm applied to the EF queue (Weighted Fair Queuing and Priority Queuing)
2. Stream profiles:
   • EF: 300 Kbps, packet size = 1024 bytes
   • BE: 2.0 Mbps, 200 packs/sec, 1000 bytes per packet
3. Test conditions:
   • EF queue-limit = 10 pack (constant)
   • tx-ring-limit: 5 particles
   • PVC: bandwidth 2 Mbps
   • queueing algorithm for EF traffic: WFQ or PQ
   • service rate of EF queue: 400 Kbps (constant)
4. Scheduler configuration:
   • WFQ: bandwidth is set to 400 Kbps, i.e. the EF service rate is slightly over-estimated. 400 Kbps was chosen for the sake of homogeneity with the PQ configuration.
   • PQ: no bandwidth needs to be assigned to the PQ queue, paramter 400 kbps is the policing rate applied to the EF queue. The EF stram deployed in the test is such that no packet drop occurred during the tests.
5. Router configuration

Results in short:

• one-way delay gain with PQ;
• with both PQ and WFQ one-way delay increases with the addition of best-effort background traffic because of the tx queue which is FIFO: this implies the presence of best-effort packets at the head of the tx queue when a EF packet arrives in a diffserv node.
• peaks in ipdv can be less frequent with PQ, however this result depends on the highly regular nature of both the BE and EF streams. More tests with a mixture of EF and BE packet sizes need to be carried out to verify the gain in ipdv with PQ.

Comments:

• As figure 1 and figure 2 show, without background traffic WFQ and PQ are equivalent both in terms of one-way delay and ipdv. 10.71 msec is the minimum delay experienced by the EF packet of 1024 bytes.
• With PQ the minimum increase in one-way delay is approximately of 9.7 msec, which corresponds to the transmission time of 2 BE packets. The transmission time of BE packets of 1000 bytes is 4.66 msec. Given the PQ rate of 300 Kbps - a relatively small fraction of the line rate -, whenever a PQ packet arrives, the tx queue (5 particles) is full: up to 2 BE packets can be stored at a time in the tx queue, 1 BE packet requires the allocation of 3 particles.
  Given the presence of a FIFO tx queue and of BE packets 1000 bytes long, 9.7 msec is the minimum increase which can be achieved by a scheduling algorithm. The instantaneous value of one-way delay can be higher, depending on the time at which a EF packet arrives, i.e. on the completion degree of the transmission of the current BE packet.
• The gain in one-way delay with PQ is considerable, in this example is up to 14 msec.
• As figure 2 shows, with PQ ipdv can be smaller: In this example the peaks height is the same in both cases, but with PQ ipdv peaks are less frequent.