Test Description

1. Network layout
2. BE data streams: single BE stream injected by INFN to congest the output interface:
   • BE payload size: 1000 by;
   • BE pack rate: 200 pack/sec (2 Mbps); CBR
   • BE protocol: UDP
3. Example of router configuration
4. Parameters:
   • number of EF streams: [1,100]
5. Stream profiles:
   • EF load: constant, equal to 640 Kbps, i.e. 120 kbps per site
   • 1 EF stream generated by the SmartBits
   • EF protocol: UDP, CBR
   • EF rate: same for each stream, constant EF IP packet size: 68 bytes
   • examples of scripts per test site:
     • 4, 8, 12, 16, 20, 24, 28, 32, 36, 40, 100
6. Test conditions:
   • PVC: bandwidth 2 Mbps
   • queueing algorithm: priority queuing applied to EF
   • EF agregation and congestion in each transit site
   • 10000 samples per test
7. Test methodology
   The number of EF sreams varies in the range [1, 100] and each site injects the same amount of traffic.
   For each test in each router on the data path we measure the amount of EF packets tail dropped by the priority queue (which serves EF packets) and we progressively increase the priority queue size Q until no packet loss caused by tail drop is observed during the whole test. Given the queue size Q at the time when such condition applies, we assume that the maximum burst size (in packets) is Q -1. The burst size in bytes can be derived since the EF packet size is constant and known.
   Measurement is applied to a single stream, which is called the reference stream.

Results in short:

• Burstiness is not a function of the number of EF streams: From 8 to 100 streams it only moderately increases.
• The maximum one-way delay experienced moderately increases with the number of EF streams.
• IPDV of a single streams is much smaller than in the case of multiple streams, however, with several streams the IPDV distribution is almost independent of the number of streams.

Comments:

• As Figure 1 shows, burstiness is only moderately a function of the number of EF streams which compose a Behaviour Aggregate: Burstiness creatly increases from 1 to 8 streams, while it slowly increases for a number of EF streams up to 100.
• Figure 2 and 3 compare the one-way delay distribution computed for increasing aggregation degrees. The performance difference is evident by comparing the ideal performance of a single EF stream (aggregation degree A = 0) with the corresponding performance achieved with multiple flows (A = 99/100 with 100 EF streams): when A increases one-way delay gets more spread and the maximum one-way delay experienced increases.
  The delay unit is equal to the minimum delay experienced, in this case it is equal to 107.01 msec.
  

  num EF streams		Avg one-way delay (msec)
  
  1			124.226
  8			128.027
  12			135.361
  16			133.948
  20			133.653
  24			133.513
  28			132.230
  32			134.979
  36			133.318
  40			132.629
  100			128.795
  

• As Figure 4 shows, also IPDV is influenced by the number of streams belonging to the BA. IPDV is expressed in transmission units, i.e. as a multiple of the tx time of one EF packet.
  The difference between the ideal case (1 stream only) and the other is considerable: with just only 1 stream IPDV is small and almost all the packets experience the same jitter. The larger is the number of streams, the greater is the maximum IPDV, however, the difference for different numbers of streams is almost negligible.

  num of EF streams	Avg one-way delay (microsec)
  
  1			1431
  8			8222
  12			6744
  16			4370
  20			7149
  24			7510
  28			7719
  32			5765
  36			7029
  40			8188
  100			6945