Alvarion SW VERSION 5.1 User Manual

Page of 273
Chapter 4 - Operation and Administration 
 
BreezeACCESS VL System Manual
 
196 
 
Table 4-15: CIR Ranges and Defaults 
 
CIR Uplink  
CIR Downlink 
Unit 
Type 
Range (Kbps) 
Default (Kbps) 
Range (Kbps) 
Default (Kbps) 
SU-3 0-2,048 
0-2,048 
SU-6 0-4,096 
0-4,096 
SU-8 0-11,264 
0-11,264 
SU-54 0-45,056 
0-45,056 
SU-I 0-4,096 
0-6,016 
4.2.6.6.2.5  Maximum Burst Duration (SU and AU) 
Sets the maximum time for accumulating burst transmission rights according to 
the Burst Duration algorithm. 
Available values range from 0 to 2000 (milliseconds). 
The default value is 5 (milliseconds), enabling a maximum burst of (0.005 X CIR) 
Kbps after a period of inactivity of 5 milliseconds or more. 
4.2.6.6.2.6 Maximum Delay (SU only) 
Sets the maximum permitted delay in the buffers system. As certain applications 
are very sensitive to delay, if relatively high delays are permitted, these 
applications may suffer from poor performance due to data accumulation in the 
buffers from other applications, such as FTP. The Maximum Delay parameter 
limits the number of available buffers. Data that is delayed more than the 
permitted maximum delay is discarded. If the SU supports applications that are 
very sensitive to delay, the value of the Maximum Delay should be decreased. 
Valid values range from 300 to 10000 (milliseconds).  
The default value is 5000 (milliseconds). 
4.2.6.6.2.7 Graceful 
Degradation Limit (AU only) 
Sets the limit on using the graceful degradation algorithm. In cases of over 
demand, the performance of all SUs is degraded proportionally to their CIR 
(IR=(100%-k%) x CIR). The graceful degradation algorithm is used as long as  
k ≤ K, where K is the Graceful Degradation Limit. Beyond this point the simple 
“brute force” algorithm is used. The Graceful Degradation Limit should be raised 
in proportion to the demand in the cell. The higher the expected demand in a cell, 
the higher the value of the Graceful Degradation Limit. Higher demand can be 
expected in cases of significant over subscription and/or in deployments where a