CONTENTS

 1     Paradigm shift – Motivation       
        1.1     Purpose 
        1.2     Scope and initial objectives         
        1.3     Linearity assumption        
 2     References
 3     Terminology and definitions      
        3.1     Session versus Call          
        3.2     General definitions           
        3.3     BHxA-related definitions 
 4     Abbreviations 
        4.1     Mathematical symbols      
 5     Basic model for 2-party communication services          
        5.1     Network model    
        5.2     Session variants   
 6    Processing performance 
        6.1     Idealized model    
        6.2     Session processing performance  
        6.3     Context processing performance  
        6.4     H.248 performance classes          
 7     Capacity   
        7.1     Theoretical capacity         
        7.2     Engineered capacity         
 8     Reference Control Load 
        8.1     Session Processor load parameters          
        8.2     Context Processor load parameters          
 9     Session-to-Context relation        
        9.1     Background          
        9.2     1:1 relationship    
        9.3     1:N relationship   
10     Extensions for the basic control load quantum  
       10.1     Extension factors 
       10.2     Throughput reduction factors       
       10.3     Reduced effective throughput in case of extended H.248 context processing      
Appendix I – Fundamental relations    
        I.1        Relation between Effective Multiplication Factor k and Extension Factor e     
Appendix II – Basic traffic models for H.248 systems    
       II.1     Lost context model
       II.2     Overload Control Model  
       II.3     Combined control/user plane model for H.248 Contexts of type "Circuit‑to-X"  
       II.4     Effective throughput versus Context Holding Time: f_CoCPS = f(C_OHT)     
       II.5     Overload Control Model for access gateways      
       II.6     Overload Control Model for ITU-T Rec. H.248.11          
Appendix III – Examples of control processing capacity computations