XPost: comp.dcom.lans.ethernet, comp.dcom.sys.cisco   
   From: barmar@alum.mit.edu   
      
   In article ,   
    "Mark"  wrote:   
      
   > "Barry Margolin"  wrote in message   
   > news:barmar-2FD372.14174320092012@news.eternal-september.org...   
   > > Usually they each maintain separate tables.  Each routing protocol   
   > > typically needs to maintain its own state, because they operate   
   > > independently -- a RIP route shouldn't normally be advertised via BGP.   
   > > The hardware FIB is created by merging all the information from the   
   > > different protocols.   
   >   
   > Thanks for the reply. To be more specific -- a routing protocol maintains   
   > its state and RIB table should reflect the RIB information down to the   
   > kernel or/and hardware; so my question is -- should it be done on both   
   > kernel and hardware, i.e. FIB in the hardware, or only FIB in hardware?  (By   
   > the hardware let's consider L3 switch, or network processor). Obviously if   
   > we have a generic x86 machine with NICs and OS Linux, then the only FIB we   
   > update is the kernel's FIB, but what if we have Linux and user-level   
   > applications implementing routing protocols, and underlying network   
   > processor capable of maintaining its own FIB tables?   
      
   If you're implementing a router, you can do whatever you want. The main   
   benefit of keeping a copy of the FIB in the kernel is that it doesn't   
   have to interrupt the hardware if an application asks to view the FIB.   
      
   --   
   Barry Margolin, barmar@alum.mit.edu   
   Arlington, MA   
   *** PLEASE post questions in newsgroups, not directly to me ***   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)