Diff: AddressingProcess

--- Version 1 

+++ Version 2 

@@ -1,8 +1,8 @@ 

 !!! Before reading this 

-This is work in progress. You will need to know something about routing - see [http://home.pacific.net.au/~djames.hub/mesh/routing_faq.html] for a mesh oriented introduction, or look on the net. 

+This is work in progress. You will need to know something about routing - review the (http://home.pacific.net.au/~djames.hub/mesh/routing_faq.html Mesh Routing FAQ) for a mesh oriented introduction, or look on the net. 

 

 !!! Precedents 

-IPv4 addresses have been doled out by groups such as [www.apnic.net APNIC] for quite some time now. These people have ended up using a number of systems that reside in front of a master __whois__ database. This database keeps track of address delegations to technical contacts etc. The idea is that perhaps Melbourne wireless should use the same style of system. 

+IPv4 addresses have been doled out by groups such as (http://www.apnic.net APNIC) for quite some time now. These people have ended up using a number of systems that reside in front of a master __whois__ database. This database keeps track of address delegations to technical contacts etc. The idea is that perhaps Melbourne wireless should use the same style of system. 

 

 !!! Introduction 

 So why don't we just start sequentially giving out addresses that are not (http://www.freenetworks.org/moin/index.cgi/NetworkAddressAllocations in use) by other community wireless organisations (for Melbourne Wireless this is 10.10.0.0/16 and 172.16.80.0/20)? Because if we did that, then there would never be any real possibility to have efficient, __scalable__ routing. 

@@ -10,7 +10,7 @@ 

 Routing protocols require that addresses be dealt out in a fashion that matches the topology, in order to be able to aggregate address space, and say "all of 10.10.13.0/24 is over-here", rather than having to tell the world about all sixteen of the /28 that are in the same /24. By being able to aggregate we reduce the load on our routing nodes - so Melbourne Wireless can scale to our goals. 

 

 !!! OSPF areas and addressing 

-OSPF is probably the interior gateway routing protocol of choice for our network. In order to make configuration of areas in OSPF practical, we also need to have contiguous blocks of IP allocated to each area.For more detail on this see [http://www.cisco.com/warp/public/104/2.html]. One crucial point about OSPF areas and addressing: every area must be connected to area 0, ie. there must be an area border router that participates in the given area and area 0. Virtual-link or tunneling techniques can help get around this restriction though. 

+OSPF is probably the interior gateway routing protocol of choice for our network. In order to make configuration of areas in OSPF practical, we also need to have contiguous blocks of IP allocated to each area. For more detail on this see the (http://www.cisco.com/warp/public/104/2.html OSPF Design Guide). One crucial point about OSPF areas and addressing: every area must be connected to area 0, ie. there must be an area border router that participates in the given area and area 0. Virtual-link or tunneling techniques can help get around this restriction though. 

 

 !! Options 

 * (http://www.itee.uq.edu.au/~mesh/db2/address.html Carving). This has been oriented towards allocation of (abundant) IPv6 space, but could also work potentially for IPv4 allocation.