Title : 
Information transmission over a finite buffer channel
         
        
            Author : 
Diggavi, Suhas N. ; Grossglauser, Matthias
         
        
            Author_Institution : 
AT&T Shannon Labs., Florham Park, NJ, USA
         
        
        
        
        
            Abstract : 
We study information transmission through a finite buffer channel modeled as a concatenation of a discrete memoryless channel (DMC) and a finite state erasure channel. The state of the erasure channel is determined by the buffer occupancy upon arrival of the transmission symbol; an erasure occurs when an input arrives to a full buffer. We show that the capacity of the channel depends on the long-term loss probability of the buffer and the capacity of the DMC. Thus, even though the channel itself has memory, the capacity apparently depends only on the stationary loss probability of the buffer. We also show that delayed feedback does not help in this channel. We also study the channel as a deletion channel where we do not know where the erasures have occurred
         
        
            Keywords : 
buffer storage; channel capacity; memoryless systems; packet switching; DMC; capacity; delayed feedback; deletion channel; discrete memoryless channel; finite buffer channel; finite state erasure channel; information transmission; long-term loss probability; stationary loss probability; Buffer overflow; Capacity planning; Channel capacity; Dairy products; Delay; Feedback; Laboratories; Memoryless systems; Packet switching; Propagation losses;
         
        
        
        
            Conference_Titel : 
Information Theory, 2000. Proceedings. IEEE International Symposium on
         
        
            Conference_Location : 
Sorrento
         
        
            Print_ISBN : 
0-7803-5857-0
         
        
        
            DOI : 
10.1109/ISIT.2000.866342