Packet loss on a 27Mbit/sec rtp transport stream


Just to recap the configuration:

VMW Infrastructure 3 composed by three DL380G5 ESX3.5 nodes, each with double FC connection to SAN and 6 Gbit NIC (2 for Service Console and 4 for VM in trunking mode). They are connected redundatly on a pair of Cisco WS-C2960G (12 on switch-1 and 12 on switch-2, no etherchanel, native access for Service Console and Truncking 5 vlans on a vSwitch dedicated for VMs). This switches are not connected directly but are conncted redundantly on a WS-C4500 pair (distribution layer).

In a separate vlan and in a separate switches there is an MPEGoverIP generator that multicast about 27Mbit/sec traffic. If in the switches where VMW Nodes are connected I put a physical machine I can retrive all the traffic (multicast routing is implemented using PIM sparse-mode on C4500) without lost a packet, but using VM on nodes this traffic has some interruption, generally lost randomly few packet in burst, enough to corrupt reception of this multicast stream. I try to recieve the multicast traffic with an Ubuntu 8.04.2 VM and with an Windows XP Pro VM and both present the same behaviour. On both I've installed vmware-tools and used "flexible" NIC Adapter (ubuntu does not have "enanched" or "e1000" capable vmware-tools...).

Any help?

0 Kudos
1 Reply

Hello Econgiu,

Were you able to make any headway on this.

Does RTP work through the ESXi virtual switches?


0 Kudos