[comp.dcom.sys.cisco] arp cache/rif cache aging

ACEE@CERNVM.BITNET (Acee Lindem) (04/24/91)

 
I recently posted the query below regarding accessibility of
IBM 8209 bridged Ethernet V2 stations when the rif cache entry
had aged and the arp cache was valid. Does anybody know what
the algorithm is for aging arp cache entries? The network is
graphically depicted below:
                                            Ethernet V2
                                           +--------+
                                           | IBM    |
                                           | 8209   |
                                           +--------+
                                               |
   (****************)                 (**************)
  (                  )               (                )
  (  Class C         )   +-------+   (  Class B       ) +-------+
  (  Bridged Token   )---| AGS   |---(  Bridged Token )-| IBM   | Ether-
  (  Ring Network    )   | GW    |   (  Ring/Ethernet ) | 8209  | net V2
  (                  )   +-------+    (**************)  +-------+
   (****************)                          |
                                           +--------+
                                           | IBM    |
                                           | 8209   |
                                           +--------+
                                               |
                                            Ethernet V2
 
  The scenario where the problem occurs:
 
 1. A station on the class C network attempts to access a
    station on the Ethernet. The AGS Gateway has a valid
    arp cache entry but the rif cache entry has aged out.
 2. The AGS gateway send an XID explorer frame to dynamically
    discover the source routing route.
 3. The IBM 8209 passes the XID explorer onto the Ethernet segment.
 4. However, Ethernet V2 stations do not know what to do not
    respond to the XID. Hence, the station cannot be accessed.
 
Does anybody know a work around for this (other than hard configuring
all the Ethernet stations)? I believe synchronizing the aging of
the arp and rif cache would do the trick but can't find anyway to set
the arp cache aging rate.
 
Original Query:
 
>When routing TCP/IP to a bridged token ring network with Ethernets
>attached via IBM 8209s, accessibility problems result in the following
>situation:
>
>  1. The target station is resident on the Ethernet.
>  2. The AGS Gateway contains a valid IP ARP cache entry for
>     the target station.
>  3. The token ring source routing RIF cache entry for the station
>     has been aged out.
>
>To find the best route, an XID explorer frame is sent to the target     on
>and the 8209 bridges it to the Ethernet. However, most Ethernet V2
>stations don't respond to XID frames. Consequently, the station cannot
>be accessed. A solution is to hard configure the Ethernet station
>as having a permanent RIF but this is not very flexible. In the other
>direction, everything works fine.
>
>Is there a fix available? Has anybody else experienced this problem?