Last week we discussed the details of the replication process,
specifically we talked about update sequence numbers (USNs). This week,
we look at the concept of propagation dampening, which manages updates
coming from two different servers.
Since it is possible for a domain controller (DC) in your Active
Directory (AD) to receive the same replication data from different DCs,
AD has implemented a technique called propagation dampening.
Propagation dampening is the process by which a DC only requests that
changes that have not been replicated. Propagation dampening uses two
"vectors" called the "up-to-date vector" and the "high watermark
vector". The up-to-date vector is nothing more than a set of USNs that
represents the highest originating update (see my first article in this
series for more information about "originating updates"). The
up-to-date vector numbers are incremented when replication occurs with
the originating server. Each domain controller has its own different USN
(They may not start at the same number). The high watermark vector is
the highest USN from each domain controller that is stored on other
domain controllers. If the high water mark received from the server
that sent the update request is the same as the high watermark for the
originating server on the server receiving the request, the receiving
server will not send the replicated information.
Replicating these changes among the various partners in the replication
process takes some time. The amount of time required for a change to be
replicated to domain controllers throughout the domain is called
"propagation delay". Additionally, the AD replication process uses a
ring topology where the replication partners form a ring. This adds
reliability to the process and helps decrease propagation delay.
Next week we will discuss AD conflict resolution and how AD handles an
attribute changing on two different DC's between replication intervals.
Active directory revisited - USNs and replication, part 3
Active directory revisited