I am not exactly sure what you mean by a "DISP like environment", but
the approach we pioneered at the University of Michigan was to use LDAP
itself as the replication protocol. That is what the U-M LDAP 3.3
slurpd/slapd combination do -- the directory server (slapd) logs changes
to a file and a separate replication agent (slurpd) picks up the changes
and sends them to slave servers over the standard LDAP protocol. There
is a small bit of special code in the slave servers that executes when
receiving replicated changes (to ensure that create and modify
timestamps are recorded unchanged).
Here at Netscape, in our Directory Server 1.0 product, we use a similar
approach. There are some important differences (or example, our
replication engine is built into the directory server process itself and
it is quite a bit more advanced than the one implemented by the U-M
slurpd server) but we still ship the changes around using vanilla LDAP.
It works quite well.
-- Mark Smith Netscape Communications Corp. The opinions I express are my own. Directory Server Engineering