Taking Care – Error Handling - Applied SOA Patterns on the Oracle Platform - page 495

Basis for proactive Fault Management
For TOGAF architects, the preceding diagram will present some resemblance to generic
horizontal layering, which is similar to the OSI Reference Model where each layer provides
services to the surrounding layers. We exclude some generic enterprise layers in order to
focus on the SOA enterprise model, as it is realized in OFM. The difference is that OSI
mostly depicts seven layers between applications, from one API to another (that is, integra-
tion), which is not applicable for service compositions where an individual service spans
across several technical and logical layers. We have no intentions here to map the TOGAF/
OSI model to SOA (actually, this theoretical exercise is done already). The sole purpose of
this diagram is to illustrate the KPI-monitoring sources, their types, and input according to
rules 5, 9, 14, and 15. WLS, obviously, is the main OFM server and the Oracle database
(XE, standard, or enterprise), so technical monitoring for proactive error handling/preven-
tion will be focused around WLS JMX (green) and database server metrics (red). Function-
al monitoring is based on the information provided by SOA applications and their APIs
(SCA and OSB, yellow flows). Generally, that's what we get from BPEL's activities to
Audit , Catch , CatchAll , Throw , and OSB's Log() . We have to omit the monitoring
of the network infrastructure for brevity; as an enterprise architect, you should keep this in
mind.