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 13,960 of 14,669   
   glen herrmannsfeldt to mark_cruzNOTFORSPAM@hotmail.com   
   Re: Session Initiation Protocol   
   02 Apr 13 00:44:08   
   
   From: gah@ugcs.caltech.edu   
      
   Mark  wrote:   
      
   (snip, I wrote)   
   >> I believe both SIP and RTP run over UDP, which tends to be much   
   >> more flexible with its ports.   
      
   (snip)   
      
   > Thanks for feedback!   
   > What are the mechanisms to validate the UDP payload in case of SIP? SIP   
   > message doesn't contain size of header or the body, so how to verify that   
   > the payload is valid? In contrast, RTP indicates the size, so given the   
   > length value from UDP header it is possible to check RTP for validit and   
   > integrity. Can something similar be done for SIP?   
      
   I don't know either SIP or RTP in that much detail. Normally UDP runs   
   with a checksum. Most often, it runs over a hardware layer that supports   
   a CRC on the packet.   
      
   There are stories that in the early days of NFS (Sun's Network File   
   System) that was usually run only across a local network, that Sun would   
   turn off the UDP checksum. (The normal checksum can't be zero. If it is,   
   it is considered to be turned off.) They then relied on ethernet to   
   check the data. As people started using NFS over longer distances, or   
   across networks within an organization, and as machines got faster,   
   that stopped.   
      
   For some years, I had my home computers connected to the internet   
   through SLIP, which doesn't have a packet CRC. I never tried to run   
   NFS over it, though. As far as I know, all the popular link level   
   protocols now have a CRC. Errors inside a router are still possible.   
      
   Well, it is usual for a router to update the checksum when changing   
   the TTL on the packer instead of recomputing it. Maybe not always,   
   though.   
      
   Otherwise, if the SIP data is wrong then that packet will likely not   
   function, but the next one will.   
      
   -- glen   
      
   --- 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