Flexible multilevel output traffic control
A two stage rate shaping and scheduling system and method is implemented to control the flow of traffic to at least one output interface. The system and method involves initially queuing incoming packets into type-specific queues and applying individual rate shaping rules to each queue. A first stage arbitration is performed to determine how traffic is queued from the type-specific queues to interface-specific queues. Packets that win arbitration and pass the applied rate shaping rules are queued in interface-specific queues. Rate shaping rules are applied to the interface-specific queues. The interface-specific queues are further distinguished by priority and priority-specific and interface-specific rate shaping rules are applied to each queue. A second stage arbitration is performed to determine how different priority traffic that is targeting the same output interface is dequeued in response to interface-specific requests.
1. A method for controlling the flow of packets through at least one output interface comprising:
queuing packets into type-specific queues;
applying per-type rate shaping rules to said type-specific queues;
queuing packets, from the type-specific queues that pass said applied per-type rate shaping rules, into interface-specific queues, wherein each interface-specific queue is specific to one output interface;
applying per-interface rate shaping rules to said interface-specific queues; and
dequeuing packets, from the interface-specific queues that pass said applied per-interface rate shaping rules, from said interface-specific queues to at least one output interface in response to interface-specific requests.
2. The method of claim 1 further including:
establishing priority-specific subqueues, which have different priority levels, within said interface-specific queues;
queuing packets to particular priority-specific subqueues of said interface-specific queues; and
applying per-priority and per-interface rate shaping rules to said priority-specific subqueues.
3. The method of claim 2 wherein dequeuing packets from said interface-specific queues in response to interface-specific requests further includes arbitrating among said priority-specific subqueues.
4. The method of claim 2 wherein applying per-interface rate shaping rules includes scaling, by a common multiple, refresh rates of credit buckets that are associated with said priority-specific subqueues that have a pending packet but do not have a passing credit value when no priority-specific subqueues have both a pending packet and a passing credit value.
5. The method of claim 2 wherein queuing packets into said interface-specific queues includes arbitrating among multiple type-specific queues that target the same priority-specific subqueue.
6. The method of claim 1 wherein queuing packets into said interface-specific queues includes writing queue identifiers of said type-specific queues into said interface-specific queues to identify packets in said interface-specific queues.
7. The method of claim 6 wherein queuing packets into said type-specific queues includes writing at least the payload of said packets into a separate packet buffer and wherein dequeuing packets from said interface-specific queues includes reading at least the payload of said packets from said separate packet buffer.
8. The method of claim 1 wherein queuing packets into said type-specific queues includes writing at least the payload of said packets into a separate packet buffer and wherein dequeuing packets from said interface-specific queues includes reading at least the payload of said packets from said separate packet buffer.
9. The method of claim 8 wherein applying said per-type rate shaping rules to said type-specific queues further includes deducting an assumed length value from a credit bucket value to determine a new credit bucket value if a type-specific queue passes an applied rate shaping rule and a packet is queued into an interface-specific queues.
10. The method of claim 9 further including correcting for the difference between said assumed length value and an actual length value after a packet is fetched from said separate packet buffer.
11. The method of claim 9 further including establishing said assumed length value as an average of the actual packet length of packets that were queued in respective type-specific queues.
12. The method of claim 1 further including:
associating a scaling factor with a first one of said interface-specific queues; and
scaling, by said scaling factor, per-type rate shaping rules that are applied to the type-specific queues that target said first interface-specific queue.
13. The method of claim 12 further including adjusting the scaling factor that is associated with said first interface-specific queue in response to packet depth of said first interface-specific queue.
14. A method for controlling the flow of packets through at least one output interface comprising:
queuing packets into type-specific queues;
applying per-type rate shaping rules to said type-specific queues;
queuing packets, from the type-specific queues that pass said applied per-type rate shaping rules, into interface-specific queues;
applying per-interface rate shaping rules to said interface-specific queues; dequeuing packets, from the interface-specific queues that pass said applied per-interface rate shaping rules, from said interface-specific queues to at least one output interface in response to interface-specific requests;
establishing priority-specific subqueues, which have different priority levels, within said interface-specific queues;
queuing packets to particular priority-specific subqueues of said interface-specific queues; and
applying per-priority and per-interface rate shaping rules to said priority-specific subqueues;
wherein applying per-interface rate shaping rules includes scaling, by a common multiple, refresh rates of credit buckets that are associated with said priority-specific subqueues that have a pending packet but do not have a passing credit value when no priority-specific subqueues have both a pending packet and a passing credit value.
15. A method for controlling the flow of packets through at least one output interface comprising:
queuing packets into type-specific queues;
applying per-type rate shaping rules to said type-specific queues;
queuing packets, from the type-specific queues that pass said applied per-type rate shaping rules, into interface-specific queues;
applying per-interface rate shaping rules to said interface-specific queues;
dequeuing packets, from the interface-specific queues that pass said applied per-interface rate shaping rules, from said interface-specific queues to at least one output interface in response to interface-specific requests;
associating a scaling factor with a first one of said interface-specific queues; and
scaling, by said scaling factor, per-type rate shaping rules that are applied to the type-specific queues that target said first interface-specific queue.
16. The method of claim 15 further including adjusting the scaling factor that is associated with said first interface-specific queue in response to packet depth of said first interface-specific queue.
17. A method for controlling the flow of packets through at least one output interface comprising:
queuing packets into type-specific queues;
applying per-type rate shaping rules to said type-specific queues; performing a first stage arbitration to determine how packets destined to contending interfaces are queued from the type-specific queues to interface-specific queues;
queuing packets, from the type-specific queues that pass said applied per-type rate shaping rules and said first stage arbitration, into the interface-specific queues, wherein each interface-specific queue is specific to one output interface;
applying per-interface rate shaping rules to said interface-specific queues; and
dequeuing packets, from the interface-specific queues that pass said applied per-interface rate shaping rules, from said interface-specific queues to at least one output interface in response to interface-specific requests.