Benefits Of Using Concurrent Copy; Session Id; Concurrent Copy Domain; Intercepted Writes - IBM System storage DS6000 Series Redbook

Session
A session is a logical concept that represents a single invocation of Concurrent Copy (a
single DFSMSdss DUMP/COPY command). A session can include one or more data sets or
volumes on the same DS6000 or across different DS6000s. An individual DS6000 can
support up to 16 simultaneous Concurrent Copy sessions per volume, with a maximum of 64
simultaneous sessions per LSS.

Session ID

The system assigns a unique session ID to each Concurrent Copy session. The system uses
the session ID to identify and coordinate all host and DS6000 resources associated with a
particular Concurrent Copy session.

Concurrent Copy domain

The set of devices and tracks identified during the initialization of a Concurrent Copy session
is called the Concurrent Copy domain. It represents the set of data that Concurrent Copy
copies.

Intercepted writes

When an application tries to update information that is included in a Concurrent Copy domain,
the DS6000 intercepts those writes, thus maintaining a copy of the data as it was at the time
when the Concurrent Copy was requested.
Sidefile
A sidefile is a temporary repository for Concurrent Copy domain tracks that still have not been
copied by the SDM and are about to receive an update. During the processing of an
intercepted write, the DS6000 copies a before-image of the track being updated into a sidefile
for later processing. Together, the DS6000 and the SDM maintain two sidefiles, one in the
DS6000 cache and another in processor storage.
Terminate
The Concurrent Copy terminates when DFSMSdss has copied the Concurrent Copy domain
and both the sidefiles are empty. In error situations, either the DS6000 or the SDM can
terminate the Concurrent Copy session before the entire domain has been copied.

Fuzzy copy

Without a tool like Concurrent Copy, if you make a copy of the data while the data is being
updated, then the copy does not reflect a point-in-time version of the original data. In this
case, the copy is fuzzy, because it does not represent any particular point-in-time status. A
fuzzy copy is a set of data with no logical consistency from the application perspective.

Consistent copy

A consistent copy is a set of data with logical consistency from the application perspective.
The logical consistency of the source data is guaranteed by the application itself. The logical
consistency of the copied data must be ensured by a mechanism like Concurrent Copy. With
the application taking care at all times that the original data is consistent, you are then able
with Concurrent Copy to produce a point-in-time copy of the complete set of data. The copy
will hold the same logical consistency status as the original data at that certain point-in-time.

Benefits of using Concurrent Copy

Concurrent Copy can dramatically reduce the amount of time that is required to back up your
application data, hence increasing the application's availability time. When you use
511
Appendix A. Concurrent Copy