I have a 7604 with SIP-400 and a SPA-1x2GE. This is connected to our provider's Metro Optical Ethernet (QMOE) at layer 2 - we apply L3 routing on our interfaces. This is at our data center (DC) and is connected at full gig speed. Our other sites are connected to the QMOE at 100MB ethernet connections.
On the L3 sub-interfaces we shape to just below 100mb and have QOS policies to manage Voice, video, etc. Here is the config of the gig port on the 7604 and a couple of sub-interfaces:
class-map match-any Call-Control
match ip dscp cs3
class-map match-any Video
match ip dscp af41
class-map match-any Voice
match ip dscp ef
!
policy-map QOS_classes_24calls_1video
class Voice
priority 2200
class Video
bandwidth 460
class Call-Control
bandwidth 100
class class-default
random-detect
policy-map WAN-100MB-speed
class class-default
shape average 99920000
service-policy QOS_classes_24calls_1video
interface GigabitEthernet2/1/1
description test QMOE link
no ip address
ip flow ingress
ip flow egress
load-interval 30
negotiation auto
no snmp trap link-status
cdp enable
!
interface GigabitEthernet2/1/1.50
description VLAN-50 to SBK,TSC (100MB)
bandwidth 100000
encapsulation dot1Q 50
ip address 168.75.189.1 255.255.255.248
ip flow ingress
ip flow egress
service-policy output WAN-100MB-speed
!
interface GigabitEthernet2/1/1.52
description VLAN-52 to CMB,TSC (100MB)
bandwidth 100000
encapsulation dot1Q 52
ip address 168.75.189.9 255.255.255.248
ip flow ingress
ip flow egress
service-policy output WAN-100MB-speed
With all the other remotes sites connected to QMOE at 100mb, we're oversubscribing the DC gig connection so I'm getting a second gig connection to be installed in another SIP/SPA in the 7604.
My question is, what is the best way to configure the 2nd gig? I want to have redundancy between the two so I'm thinking of creating an etherchannel of the two and moving the config to the port channel.
Is this a good idea in this scenario? How will that affect the QOS structure seeing as how the QOS policies will be on a virtual interface instead of a hardware interface - will the shaping change? How does the queue structure work in a port channel as compared to the hardware queues on the gig ports when working with LLQ bandwidth and priority limits?
Would it be better to have the 2nd gig connection a separate L3 connection and move half the sub-interfaces to it? This will not provide redundancy to all sites if a link goes down. Is the QOS cleaner with two separate gig connections?
