Database Reference
In-Depth Information
Tip
If you are an architect on the client's side and during the RFP process, you discover such
kind of "sensitivity" combined with the inability of your vendor to explain how events are
registered, filtered, and later used for EBM construction, be extra cautious. The presence
of the Callback pattern for Events processing also has to be justified.
Implementation of noninvasive ways of event recognition and filtering (or change data
capture, CDC) is not more difficult than what was discussed earlier. Even more, some
solutions can be much faster with highest selectivity and precision, with zero CDC misses.
One possible way can be based on the already existing data replication solution for your
critical mission applications (and OEBS is definitely one of them). There are many tools
you can employ for this task and SharePlex could be a good option—it has no footprint on
the source database; it's extremely quick, and most importantly, it performs replication for
data and statements, allowing you to "sniff" certain operations on certain DB objects.
Even more, SharePlex does support data transformation before applying on the target DB
and it can work in the hub-and-spoke mode, aggregating data from many sources, not to
forget to mention that it can move a tremendous amount of data. It's relatively simple, so
establishing filters for capturing events and recording them to the event log for further
processing won't be a problem. The solution can be truly elegant, as you kill two (or prob-
ably more) birds with one stone—establish CDC with no footprint on the source DB and
reliable 24 x 7 zero loss DB replication. Comparing different data replication solutions is
not the subject of this topic, but looking to products such as Golden Gate, Streams (from
Oracle), and SharePlex (now from Dell) from our experience, we can say that SharePlex
would be the optimal choice (at the moment of writing), especially considering the price
and CDC capabilities. GG is quite similar to SharePlex and also has data transformation
capabilities, so if you have budget for it, you can try implementing CDC on it as well.
The last thing that you can see from the previous code—all object- and context-related
variables are compliant with the SBDH model we presented in the earlier chapters, so we
will use it quite extensively in all our examples.
Events filtering
Earlier we explained the difference between a basic and complex event. It's obvious and
generally related to the amount of the object's context information utilized during the
event filtering/recognition phase. Initially, the event is registered in a completely business
agnostic way, just as an evidence of the business objects changes. All that we know is the
Search WWH ::
Custom Search