home bbs files messages ]

Forums before death by AOL, social media and spammers... "We can't have nice things"

   comp.protocols.tcp-ip      TCP and IP network protocols.      14,669 messages   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]

   Message 13,264 of 14,669   
   Rick Jones to Ashwin Rao   
   Re: Interpretation of the tcp_info struc   
   14 Dec 09 19:29:18   
   
   6cd7ab12   
   From: rick.jones2@hp.com   
      
   Ashwin Rao  wrote:   
   > The interpretation of some of the values according to the linux/tcp.h   
   > file is   
      
   > tcpi_rto;            21000   
   > tcpi_ato;           40000   
   > tcpi_snd_mss;   1448   
   > tcpi_rcv_mss;    536   
   > tcpi_pmtu;               1500   
   > tcpi_rcv_ssthresh;   5792   
   > tcpi_rtt;                    1000   
   > tcpi_rttvar;               750   
   > tcpi_snd_ssthresh;   2147483647   
   > tcpi_snd_cwnd;         4   
   > tcpi_advmss;            1448   
   > 119 is 0111 0111 in binary, hence   
   > tcpi_snd_wscale : 4,     0111 =>  7   
   > tcpi_rcv_wscale : 4;     0111  => 7   
      
      
   > I have the following doubts.   
      
   I can take a try at some, but not all of your questions:   
      
   > 1. Why is the tcpi_snd_ssthresh so large? Shouldn't it be 1 or 2 as   
   > the getsockopt was called just after a few bytes (less than 10KB) were   
   > sent.   
      
   No.  ssthresh starts large and remains large until there is an actual   
   retransmission.   
      
   > 2. Does tcpi_snd_cwnd imply that the window size is 4 * 1<<7 = 512   
   > bytes.   
      
   I suspect you will find it implies 4 * 1448 (MSS) bytes.   
      
   > 3. What does the tcpi_rcv_ssthresh stand for? When is it used?   
      
   It is used by linux's socket autotuning, to advertise no more window   
   than the receiving application has demonstrated its ability to consume   
   in a timely fashion.  As the application demonstrates it can make use   
   of greater bandwidth, that value, and thus the advertised window, is   
   increased.  You might think of it as belt, suspenders, and duct tape   
   on top of the send-side congestion control and avoidance.  It may also   
   be touted as a way to keep a non-receiving application from tie-ing up   
   significant quantities of buffer space.   
      
   rick jones   
   --   
   web2.0 n, the dot.com reunion tour...   
   these opinions are mine, all mine; HP might not want them anyway... :)   
   feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]


(c) 1994,  bbs@darkrealms.ca