Method and apparatus providing flow control using on-off signals in high delay networks
A method and apparatus controlling, either individually or in the aggregate, traffic flows associated with a plurality of queues by transmitting traffic source FlowON messages at first times selected to avoid queue underflow conditions, and traffic source FlowOFF message at second times selected to avoid queue overflow conditions, where the first and second times are determined as a function of delay time D of a link supporting the traffic source.
1. A method for controlling traffic flows associated with a plurality of active queues, the method comprising:
for each active queue, transmitting toward a corresponding traffic source a FlowON message at first times selected to avoid a queue underflow condition;
for each active queue, transmitting toward said corresponding traffic source a FlowOFF message at second times selected to avoid a queue overflow condition;
said first and second times being determined as a function of delay time D of a link between said each active queue and said corresponding traffic source,
wherein time intervals between transmissions of any two messages out of a plurality of FlowON and FlowOFF messages is varied, said plurality of FlowON and FlowOFF messages being an integer of at least more than two and including said FlowON message and said FlowOFF message.
2. The method of claim 1 , wherein said first times are identified as said each active queue exhibiting a utilization level decreasing through a first threshold level.
3. The method of claim 2 , wherein the first threshold level is set to a queue utilization level at which a time to drain the remaining packets in the queue is approximately equal to a further time to begin receiving additional packets after transmission of the FlowON message.
4. The method of claim 1 , wherein said second times are identified as said each active queue exhibiting a utilization level increasing through a second threshold level.
5. The method of claim 4 , wherein the second threshold level is set to a queue utilization level at which a time to fill the queue with packets being received is approximately equal to a further time to stop receiving additional packets after transmission of the FlowOFF message.
6. The method of claim 1 , wherein said second times are identified as a time after transmitting the FlowON message increasing through a second threshold level.
7. The method of claim 1 , wherein:
the FlowON message is adapted to cause said corresponding traffic source to transmit packets towards said each active queue; and
the FlowOFF message is adapted to cause said corresponding traffic source to stop transmit packets towards said each active queue.
8. The method of claim 1 , wherein the first and second threshold levels associated with a queue i are determined using a queue size parameter Qi, a queue service rate parameter Ri, a link bandwidth parameter Bi and a link delay parameter Di.
9. The method of claim 1 , wherein each of the plurality of active queues is included within a set of active queues M receiving packets via a link having a delay time D and a bandwidth B, the method further comprising adapting said first and second times for said each active queue in accordance with a respective allocation of a link bandwidth B.
10. The method of claim 9 , wherein each of the plurality of active queues is allocated a proportionate share of the link bandwidth B.
11. The method of claim 9 , wherein one or more of the plurality of active queues is allocated a disproportionate share of the link bandwidth B.
12. The method of claim 9 , wherein the set of active queues M only includes active queues exhibiting at least a minimum utilization level.
13. The method of claim 1 , wherein a hysteresis parameter E is used to adapt said second times.
14. The method of claim 1 , further comprising:
logically representing at least a portion of said plurality of said active queues as an aggregate queue in which each member queue contemporaneously transmits said FlowON and FlowOFF messages.
15. A method for controlling traffic flows associated with a plurality of active queues, the method comprising:
for each active queue, transmitting toward a corresponding traffic source a FlowON message at first times selected to avoid a queue underflow condition;
for each active queue, transmitting toward said corresponding traffic source a FlowOFF message at second times selected to avoid a queue overflow condition;
said first and second times being determined as a function of delay time D of a link between said each active queue and said corresponding traffic source,
wherein said first times are identified as said each active queue exhibiting a utilization level decreasing through a first threshold level,
wherein the first threshold level is set to a queue utilization level at which a time to drain the remaining packets in the queue is approximately equal to a further time to begin receiving additional packets after transmission of the FlowON message,
wherein the first threshold level comprises a remaining amount of data within the queue calculated as a queue service rate parameter Ri multiplied by twice a link delay.
16. A method for controlling traffic flows associated with a plurality of active queues, the method comprising:
for each active queue, transmitting toward a corresponding traffic source a FlowON message at first times selected to avoid a queue underflow condition;
for each active queue, transmitting toward said corresponding traffic source a FlowOFF message at second times selected to avoid a queue overflow condition;
said first and second times being determined as a function of a delay time D of a link between said each active queue and said corresponding traffic source,
wherein each of the plurality of active queues has a drainage rate Ri and is associated with a respective timer for determining a time between said first times and said second times, the method further comprising:
for said each active queue, starting said respective timer in response to transmitting said FlowON message, said respective timer having a stop time determined as a function of the delay time D of the link and the drainage rate Ri of said each active queue.
17. A method for controlling traffic flows associated with a plurality of active queues, the method comprising:
for each active queue, transmitting toward a corresponding traffic source a FlowON message at first times selected to avoid a queue underflow condition;
for each active queue, transmitting toward said corresponding traffic source a FlowOFF message at second times selected to avoid a queue overflow condition; and
logically representing at least a portion of said plurality of said active queues as an aggregate queue in which each member queue contemporaneously transmits said FlowON and FlowOFF messages,
said first and second times being determined as a function of a delay time D of a link between said each active queue and said corresponding traffic source,
wherein contemporaneous first and second times for said each active queue are determined using an aggregated queue size parameter Qa, an aggregated queue service rate parameter Ra, a link bandwidth parameter B and a link delay parameter D.
18. A non-transitory computer readable medium including software instructions which, when executed by a processor, performs a method for controlling traffic flows associated with a plurality of active queues, the method comprising:
for each active queue, transmitting toward a corresponding traffic source a FlowON message at first times selected to avoid a queue underflow condition;
for each active queue, transmitting toward said corresponding traffic source a FlowOFF message at second times selected to avoid a queue overflow condition;
said first and second times being determined as a function of delay time D of a link between said eachg active queue and said corresponding traffic source,
wherein time intervals between transmissions of any two messages out of a plurality of FlowON and FlowOFF messages is varied, said plurality of FlowON and FlowOFF messages being an integer of at least more than two and including said FlowON message and said FlowOFF message.
19. A computer program product, wherein a computer is operative to process software instructions which adapt the operation of the computer such that computer performs a method for controlling traffic flows associated with a plurality of active queues, the method comprising:
for each active queue, transmitting toward a corresponding traffic source a FlowON message at first times selected to avoid a queue underflow condition;
for each active queue, transmitting toward said corresponding traffic source a FlowOFF message at second times selected to avoid a queue overflow condition;
said first and second times being determined as a function of delay time D of a link between said each active queue and said corresponding traffic source,
wherein time intervals between transmissions of any two messages out of a plurality of FlowON and FlowOFF messages is varied, said plurality of FlowON and FlowOFF messages being an integer of at least more than two and including said FlowON message and said FlowOFF message.