EFCI scheme with early congestion detection

Abstract

An adjustable bit rate (ABR) feedback control scheme is provided where the effects of multiloop delays and high priority traffic transmission are built into the control model. The data traffic is filtered by a low pass filter. Then, the low frequency bandwidth of the filtered traffic is measured and compared to a predetermined threshold. If the measured value exceeds the threshold, the ABR traffic flow is reduced. If the measured value is less than the threshold, the ABR traffic flow is increased. In addition, a General Prediction Control (GPC) method may be applied to the control model for optimal performance. An object of the invention is to minimize the unused link capacity subject to no congestion, where the ABR traffic is adapted to the low frequency variation of high priority traffic flow for high efficiency.

1 . A method for detecting congestion in high speed network switches of a high speed network in order to optimize throughput, the method comprising: 
 adapting low priority data traffic in response to a low frequency variation of high priority data traffic; and    eliminating low frequency, high magnitude low priority data traffic oscillations.    
   
   
       2 . The method of  claim 1 , in which the eliminating further comprises applying a generalized predictive control (GPC) method.  
   
   
       3 . The method of  claim 1 , in which the adapting further comprises detecting a bandwidth threshold requirement.  
   
   
       4 . The method of  claim 3 , in which the detecting further comprises filtering the high priority data traffic.  
   
   
       5 . The method of  claim 3 , in which the detecting further comprises filtering the high priority data traffic and the low priority data traffic.  
   
   
       6 . The method of  claim 4 , in which the filtering further comprises employing a low pass filter.  
   
   
       7 . The method of  claim 1 , in which the high speed network switches comprise ATM switches.  
   
   
       8 . The method of  claim 1 , in which the adapting further comprises adapting a low priority traffic transmission rate to an unused link capacity.  
   
   
       9 . The method of  claim 4 , in which the filtering occurs periodically.  
   
   
       10 . The method of  claim 4 , further comprising employing traffic prediction to reduce a filtering delay effect.  
   
   
       11 . A computer readable medium storing a program for detecting congestion in high speed network switches to optimize throughput, the program comprising: 
 an adaptation code segment that adapts low priority data traffic in response to a low frequency variation of high priority data traffic; and    an oscillating code segment that eliminates low frequency, high magnitude low priority data traffic oscillations.    
   
   
       12 . The medium of  claim 11 , in which the oscillating code segment further comprises applying a generalized predictive control (GPC) method.  
   
   
       13 . The medium of  claim 11 , in which the adapting code segment further comprises a detecting code segment that detects a bandwidth threshold requirement.  
   
   
       14 . The medium of  claim 13 , in which the detecting code segment further comprises a filtering code segment that filters the high priority data traffic.  
   
   
       15 . The medium of  claim 13 , in which the detecting code segment further comprises a filtering code segment that filters the high priority data traffic and the low priority data traffic.  
   
   
       16 . The medium of  claim 14 , in which the filtering code segment employs a low pass filter.  
   
   
       17 . The method of  claim 11 , in which the high speed network switches comprise ATM switches.  
   
   
       18 . The medium of  claim 11 , in which the adapting code segment further comprises adapting a low priority traffic transmission rate to an unused link capacity.  
   
   
       19 . The medium of  claim 14 , in which the filtering occurs periodically.  
   
   
       20 . The medium of  claim 14 , further comprising a traffic prediction code segment that predicts traffic to reduce a filtering delay effect.