Cisco Labs' QoS Testing






Cisco Labs' QoS Testing

Cisco conducted the following tests to show not only VoIP quality, but also how Cisco QoS tools work under load.

For the testbed, Cisco used a simple network with two Cisco VoIP gateways and one 56-kbps WAN. It completed two tests:

  • Test A consisted of testing with and without QoS enabled while steadily increasing the saturation of the WAN link.

  • Test B consisted of testing with and without QoS enabled while sending traffic across the WAN link in a bursty nature.

When no QoS was used, FIFO queuing was implemented. When QoS was used, MCML PPP, WFQ, and IP Precedence were utilized.

Cisco sent the traffic through two in-house traffic generation tools. It measured latency and voice quality using a voice-quality test tool that utilizes the ITU-T Perceptual Speech Quality (PSQM) recommendation P.861.

When measuring PSQM, the higher the score the worse the voice quality. Figure shows the results from Test A.

Results for Test A

Bandwidth Saturation (Percent)

Delay (ms)

PSQM

Delay/QoS (ms)

PSQM/ QoS

0

76

1.43

76

1.43

40

233

1.94

106

1.5

60

242

2.1

104

1.43

75

280

7

102

1.51

90

300

9

104

1.51

100

350

10

105

1.51


To better understand the difference QoS can make, see Figure.

Graph of Test A Results


Results for Test B are just as dramatic and appear in Figure. Cisco IOS Software QoS tools, when implemented properly, can definitely affect the caliber and stability of voice quality.

Graph of Test B Results




 Python   SQL   Java   php   Perl 
 game development   web development   internet   *nix   graphics   hardware 
 telecommunications   C++ 
 Flash   Active Directory   Windows