The Routing Over Large Clouds Mailing List Archive by date

Cell Relay Retreat>List Archive>month:1994-Nov> msg00001



[Date Prev][Date Next][Thread Prev][Thread Next]  
  [Date Index][Thread Index][Author Index][Subject Index]

Multiple NHSs

  • From: Kanan F Shah <kshah@cmf.nrl.navy.mil>
  • Date: Tue, 08 Nov 94 09:01:41 -0500
  • cc: rolc@maelstrom.timeplex.com

Lorraine,

] 
] I am wondering if it is possible with NHRP to have the situation where
] a NHS is trying to resolve the NBMA address of a destination it doesn't
] a NHS is trying to resolve the NBMA address of a destination it doesn't
] serve and there are several NHSs to choose between when deciding which NHS 
] to forward the NHRP request to.  If so, how does the NHS decide which NHS to 
] forward the NHRP request to?
] 
] In NARP it states that NASs exchange reachability information with each other
] In NARP it states that NASs exchange reachability information with each other
] which can then be used to decide which NAS to forward a given NARP request to
] which can then be used to decide which NAS to forward a given NARP request to
] .
] Is the situation similar for NHRP?
] 

There are two answers to this question since there are two different "modes"
in which NHRP can be deployed.  Server mode assumes a loose coupling with IP
routing (i.e. routers in network are not NHRP capable) whereas in Fabric Mode 
NHRP capable routers exist throughout the NBMA network, hence leveraging the 
NHRP capable routers exist throughout the NBMA network, hence leveraging the 
existing routing infrastructure that is "overlaid" on the NBMA network i.e. 
existing routing infrastructure that is "overlaid" on the NBMA network i.e. 
unlike Server mode the routed path of NHRP packets, in Fabric mode, is similar 
to that of IP packets

Server mode is assumed to be the most likely scenario for the initial deployment
of NHRP.  As such, the peer NHS's are statically configured with each others
of NHRP.  As such, the peer NHS's are statically configured with each others
addresses as well as the identities of the destinations they each serve.  With
this information each NHS can look at its forwarding table to determine which,
if any, NHS serves the address to be resolved and can correspondingly forward 
the request to the correct NHS. 

In Fabric Mode this manual configuration is eliminated and the NHS's acquire
knowledge about destinations other NHSs serve via participation in intradomain 
and interdomain routing protocol exchange, which is similar to NARP.
and interdomain routing protocol exchange, which is similar to NARP.

Hope this helps.

--Kanan Shah



  • References:
    • Multiple NHSs
      • From: Lorraine de Vere <lorraine@snrc.uow.edu.au>