Just some general questions someone might know:
On 3750 egress queues, when you specify priority queue out, assuming q1-4 were defaulted to 25%, q1 gets served until empty: Do q1-3 bump up to 33% of remaining available time or do they stay at 25%, leaving 25% unscheduled?
On 3750 the internal ring is 32Gbps, does anyone know what it is on a 3560, same?
On egress queues, if you only have 2 classifications of traffic, 1 to the priority queue and an 'all other' to one of the other queues, is the bandwidth allocated to the 2 unused queues just idle time? Should you reconfigure to minimize that?
In a recent example setup I saw for IP phones (refer to my thread from 2 days ago)they set the input priority queue input bandwidth to 10%. Given that even at G.711 on a 10Mbps interface the bandwidth requirements seem to be only about .8%, this seems like extreme overkill. Even given that the setting is global and has to take into account uplink ports, you'd have to be running about 1500 simultaneous calls to approach this value on a 1Gbps uplink. Are there additional considerations that would justify setting it this high?
Thx
T&F
On 3750 egress queues, when you specify priority queue out, assuming q1-4 were defaulted to 25%, q1 gets served until empty: Do q1-3 bump up to 33% of remaining available time or do they stay at 25%, leaving 25% unscheduled?
On 3750 the internal ring is 32Gbps, does anyone know what it is on a 3560, same?
On egress queues, if you only have 2 classifications of traffic, 1 to the priority queue and an 'all other' to one of the other queues, is the bandwidth allocated to the 2 unused queues just idle time? Should you reconfigure to minimize that?
In a recent example setup I saw for IP phones (refer to my thread from 2 days ago)they set the input priority queue input bandwidth to 10%. Given that even at G.711 on a 10Mbps interface the bandwidth requirements seem to be only about .8%, this seems like extreme overkill. Even given that the setting is global and has to take into account uplink ports, you'd have to be running about 1500 simultaneous calls to approach this value on a 1Gbps uplink. Are there additional considerations that would justify setting it this high?
Thx
T&F