My thoughts are currently like this then: we wait for an OSPF area 0
mesh to develop. When we start getting other nodes joining inside an
existing, reasonable sized cycle in the mesh, we partition it off and
call it a new area. If these nodes can later link _across_ the cycle
boundary to another area, they are then connected in three area (which
is legit). All nodes within the cycle continue to be allocated the
same address block....?
This means we can just start using 10.10.x.x blocks sequentially for
networks and 172.12.0.0 addresses for links, sequentially. When we get
to the point where we start having a large number of nodes, the growth
will happen to fall more and more in the existing cycles, and we will
be able to start doling out slabs of address space to the new OSPF
areas as they are created.
[EditText] [Spelling] [Current] [Raw] [Code] [Diff] [Subscribe] [VersionHistory] [Revert] [Delete] [RecentChanges]
Node Statistics | |
---|---|
building | 132 |
gathering | 193 |
interested | 515 |
operational | 233 |
testing | 214 |