Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations Chris Miller on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

Spirent Smartbits SmartVoIPQoS & NetIQ Vivinet Assessor?

Status
Not open for further replies.

ADB100

Technical User
Mar 25, 2003
2,399
GB
I came across an issue with testing today using Spirent's Smartbits SmartVoIPQoS and NetIQ's Vivinet Assessor.

The scenario is this - customer network is provisioned to handle 100 G.711 calls over a link (I'll spare the details....). If I test with Smartbits SmartVoIPQoS and configure 100 G.711 RTP flows (actually 100 from A to B and 100 from B to A) I get a high percentage of packet drops on the switches due to shaping the egress queue (I can see the drops in the switch output). The shaper is set higher than what should be the actual maximum traffic rate - 1% of 1Gbps (10mbps) - 100 G.711 calls should be less than 10mbps based on about 90kbps per call. If I repeat the test but use NetIQ Vivinet Assessor with two endpoints simulating 100 calls I don't get the same packet loss (I don't get any loss). In theory both these tests should be about the same, however it looks like the Smartbits is creating too high a rate of traffic. I have verified that I am only sending a single Voice payload per packet with Smartbits (default is four). It is SmartVoIPQoS version 1.12 (I think?).

Has anyone any experience with either of these and any has any clues as to what I am seeing?

Thanks

Andy
 
What kind of switches?? Why is the bandwidth of the network links. How do you have QoS setup? Are you marking DSCP/TOS on the smartbits but not on the netiq or vice versa?


Kinda hard to give you an answer, there could be many reasons.. I would definately get a packet capture going of both tests to see what is different..


BuckWeet
 
Hi, I think I have sorted it out. The documentation for the Smartbits talks about 'the number of Voice Frames per IP Packet' but it's slightly misleading. I think my configuration of the Smartbits test I was performing was generating too much traffic, hence the packet loss.

The equipment in test is access-layer Catalyst 4500 with SupV-10GE connected via layer-3 Gig EtherChannels to a pair of Core 6500's with Sup720's. The QoS is very strict, i.e. shaping the egress PQ to 1%.

I am much happier now I understand the Smartbits better :eek:)

Andy
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top