I am planning to implement QOS on the network for Voice, Data. In my opinion, i would like to keep it simple just using MQC and only classify, mark and priotises the traffic as the traffic is routed between trust boundries.
Data is give higher priority
Second is Voice
shaping retains excess packets in a queue. is this in memory? will this slow down our switches?
Using Policing will means Traffic policing propagates bursts. When the traffic rate reaches the configured maximum rate, excess traffic is dropped or remark. right? Is this the right thing to use.
Is it just better to just classify, mark and priotize without reserving fix bandwidth or queuing in buffer? i know switches have built in queue to queue traffic based on markings.