[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: updating GSE for the new millennium


On Friday, May 2, 2003, at 10:28 Canada/Eastern, Brian E Carpenter wrote:

David Conrad wrote:
...
With regards to maintaining the mapping table, I see two generic ways
of doing this, either pushing data out like routing protocols or
pulling data in on demand like the DNS. Both are (obviously) tractable
and both have advantages and disadvantages. For obvious reasons I like
the DNS model (not necessarily the DNS itself), but I see this is a
side (albeit important) issue to the underlying architecture.
No, I don't think it's a side issue. DNS brings the risk of a big-time
circular dependency. In fact, Jon Postel thought about this one: he
contributed the following to RFC 1958:

   3.11 Circular dependencies must be avoided.

      For example, routing must not depend on look-ups in the Domain
      Name System (DNS), since the updating of DNS servers depends on
      successful routing.
Some circular dependencies can be managed, however.

The DNS itself contains a circular dependency; a resolver can't find a true current set of authoritative nameservers for the root without asking a question of a root server. In this case we make do with a hints file which doesn't need to contain a true current set of NS records; it just needs a set which is good enough to find at least one root nameserver.

So a solution which provided basic connectivity and sub-optimal routing for bootstrap purposes might be ok, if that connectivity and routing could be subsequently refined and optimised by some dependent infrastructure.

For example, a site might bootstrap itself as single-homed, ignoring all but one of the diverse paths available to connect itself to the rest of the world, and only become really multi-homed once it had used its single-homed access to look up some supporting data (using the DNS or something).


Joe