The Routing Over Large Clouds Mailing List Archive by date

["Tom Golway" <golway_thomas@jpmorgan.com>]

[This note has been sent to the following InterNet address(es):
rolc @ nexen.com
ip-atm @ nexen.com]


Excuse my ignorance on the context of this discussion, but I was wondering
about the scalability and scope of this discussion.  My question relates to
whether this discussion can be viewed as a general solution to environments
which have critical applications needing redudant servers, or is it a specific
solution for servers which support network services type applications.

My question stems from my interaction with various users who have custom
client/server applications which have primary and backup servers.  I see two
problems with the backup servers.  One is the synchronization of the
information (which can change rapidly), and two is how to dynamically redirect
traffic to the backup server when the primary goes down.

As I look at this question in the context of clients with VCs setup to their
primary server, which is more desireable,  reestablishing VCs to the backup
server in the event that the primary goes down, or performing some sort of VC
redirect.  In the Financial industry, stock/bond/etc. market information is
typically sent out via a multicast.  The though on this is in an ATM network to
send these feeds via a point to multipoint connection.  Of course the root
would require a backup, and again there needs to be someway for an auto cutover
to the new root upon a failure.

Anyway, I hope that I have been somewhat clear on my questions.  thanks in
advance for any constructive input.

Tom Golway
golway_thomas@jpmorgan.com