If I create multiple network interface devices, Parallels seems to route all traffic through the first interface regardless of which interface the OS thinks it is sending the traffic to. For example, I have configured five Parallels network interfaces, all using host-only network (although I see the same problem using bridged networking). In CentOS, you can see that each interface eth0 through eth4 are bound to a different MAC address, and therefore to a different Parallels virtual interface: eth0 Link encap:Ethernet HWaddr 00:16:23:22:C3F inet addr:192.168.1.50 Bcast:192.168.1.255 Mask:255.255.255.0 inet6 addr: fe80::216:23ff:fe22:c3df/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:1064 errors:0 dropped:0 overruns:0 frame:0 TX packets:1541 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:616587 (602.1 KiB) TX bytes:247626 (241.8 KiB) Interrupt:10 Base address:0x4c00 eth1 Link encap:Ethernet HWaddr 00:F1:9A:0D:03:2C inet addr:192.168.1.51 Bcast:192.168.1.255 Mask:255.255.255.0 inet6 addr: fe80::2f1:9aff:fe0d:32c/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:733 errors:0 dropped:0 overruns:0 frame:0 TX packets:88 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:84722 (82.7 KiB) TX bytes:16917 (16.5 KiB) Interrupt:11 Base address:0x5000 eth2 Link encap:Ethernet HWaddr 00:92:C2:1C:41:A7 inet addr:192.168.1.52 Bcast:192.168.1.255 Mask:255.255.255.0 inet6 addr: fe80::292:c2ff:fe1c:41a7/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:699 errors:0 dropped:0 overruns:0 frame:0 TX packets:89 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:77554 (75.7 KiB) TX bytes:16784 (16.3 KiB) Interrupt:9 Base address:0x5400 eth3 Link encap:Ethernet HWaddr 00:2E:95:77:9E4 inet addr:192.168.1.53 Bcast:192.168.1.255 Mask:255.255.255.0 inet6 addr: fe80::22e:95ff:fe77:9ed4/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:669 errors:0 dropped:0 overruns:0 frame:0 TX packets:90 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:70086 (68.4 KiB) TX bytes:16767 (16.3 KiB) Interrupt:9 Base address:0x5800 eth4 Link encap:Ethernet HWaddr 00:4C:93:AC:EB:A8 inet addr:192.168.1.54 Bcast:192.168.1.255 Mask:255.255.255.0 inet6 addr: fe80::24c:93ff:feac:eba8/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:643 errors:0 dropped:0 overruns:0 frame:0 TX packets:90 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:62657 (61.1 KiB) TX bytes:16619 (16.2 KiB) Interrupt:10 Base address:0x5c00 However, no matter which IP address I ping from outside of the machine (from a second Parallels installation, for example), the only interface which shows traffic is the first one (based on observation of the network icons on the bottom right of the Parallels window). I can confirm that this is what's happening by disconnecting the first network interface. If the first network interface--which only eth0 (192.168.1.50) should be bound to--is disconnected, all network traffic to and from this machine is halted. I should be able to ping IP addresses 192.168.1.51-54 from outside of the machine and see traffic on interfaces 2-4, but I do not. This is very frustrating, as it is preventing me from doing some software testing work for my day job. Has anyone else observed this behavior? Parallels support, is this a bug, or am I misunderstanding how the product is supposed to work?