Method and system for weighted fair flow control in an asynchronous metro packet transport ring network
View Patent ↗A method and system for implementing weighted fair flow control on a metropolitan area network. Weighted fair flow control is implemented using a plurality of metro packet switches (MPS), each including a respective plurality of virtual queues and a respective plurality of per flow queues. Each MPS accepts data from a respective plurality of local input flows. Each local input flow has a respective quality of service (QoS) associated therewith. The data of the local input flows are queued using the per flow queues, with each input flow having its respective per flow queue. Each virtual queue maintains a track of the flow rate of its respective local input flow. Data is transmitted from the local input flows of each MPS across a communications channel of the network and the bandwidth of the communications channel is allocated in accordance with the QoS of each local input flow. The QoS is used to determine the rate of transmission of the local input flow from the per flow queue to the communications channel. This implements an efficient weighted bandwidth utilization of the communications channel. Among the plurality of MPS, bandwidth of the communications channel is allocated by throttling the rate at which data is transmitted from an upstream MPS with respect to the rate at which data is transmitted from a downstream MPS, thereby implementing a weighted fair bandwidth utilization of the communications channel.
1. In a metropolitan area network, a method for implementing weighted fair flow control on the network, the method comprising the steps of:
a) accepting a plurality of local input flows at each of a plurality of MPS for transport across a communications channel;
b) transporting data among the MPS via the communications channel asynchronously, wherein the plurality of MPS include at least one upstream MPS and one downstream MPS;
c) for each MPS:
c1) assigning a QoS to each local input flow;
c2) allocating a portion of insertion traffic bandwidth of the MPS to each local input flow in accordance with the QoS to implement weighted bandwidth allocation;
c3) inserting the insertion traffic of the MPS onto the communications channel using an available opening in the communications channel;
c4) if the insertion traffic needs to be reduced, reducing the allocation to those local input flows having a lower QoS before reducing the allocation to those local input flows having a higher QoS; and
d) if the downstream MPS experiences congestion, throttling the insertion traffic of the upstream MPS to implement fair bandwidth allocation.
2. The method of claim 1 wherein the QoS includes at least a first level and a second level, the first level having a higher priority than the second level.
3. The method of claim 1 wherein the communications channel is an ethernet communications channel.
4. The method of claim 3 wherein the communications channel is a 10 gigabit ethernet communications channel.
5. The method of claim 1 wherein the metropolitan area network is a ring topology metropolitan area network.
6. The method of claim 1 wherein transit traffic on the communications channel is given strict priority with respect to insertion traffic from each MPS.
7. The method of claim 1 further including the step of:
minimizing jitter for higher QoS local input flows by reserving a portion of the insertion traffic of each MPS for the higher QoS local input flow.