10 free sample questions from a bank of 57, with the correct answers and explanations. No signup required — start practising right now.
Verdict: B - Oracle Spatial data cartridge.
Tracking where buses are and firing an alert when one reaches a given stop is a geospatial event pattern. The Oracle Spatial data cartridge is the one built for it: it carries spatial payloads (SDO_GEOMETRY points and routes) and lets a CQL query ask spatial questions of the event stream, such as which bus is within a distance of this stop.
The event carries a position, and the condition to detect is a spatial one. The Oracle Spatial cartridge understands the geometry model and exposes spatial operators to CQL, so the condition "this bus is at stop X" is evaluated in the stream itself and the alert fires without the application doing geometry maths.
A. JDBC Data - the JDBC cartridge reads relational tables and rows. It has no idea what an SDO_GEOMETRY point is, so it cannot evaluate a distance or containment condition against the event stream.
C. Hadoop Big Data - the Hadoop cartridge ingests events from HDFS and big-data sources. It solves where the data lives, not what the condition means; there is no geospatial reasoning in it.
D. NoSQLDB Big Data - the NoSQL cartridge is for document and key-value stores (for example Coherence or MongoDB-style data). Again it is a source cartridge with no spatial operators.
E. Java Data - the Java cartridge deserializes POJOs or JavaBean events. It would get the coordinates to the application as plain fields and leave any geometry work to your own code, which defeats the point of tracking positions in the stream.
Pick the cartridge by the nature of the payload and the question you must ask: geometry and distance means Oracle Spatial, rows means JDBC, documents means NoSQL, POJOs means Java, files on HDFS means Hadoop.
Verdict: B and C.
Two things decide how well Oracle Adapters interoperate with Service Bus and SOA Suite: they talk to the packaged application through native APIs rather than re-implementing its interface, and they are deployed on the same WebLogic Servers as the rest of Fusion Middleware, so they share the JCA runtime, MDS, security and transaction infrastructure.
B is the functional side of interoperability: the adapters use the back end's own APIs, so operations map onto real application semantics instead of approximations. C is the architectural side: because the adapters are deployed to the same WebLogic Servers as Service Bus and SOA, the JCA container, MDS repository and security framework are shared, so there is no cross-domain or remote-container penalty in invoking them.
A. Oracle Adapters are based on the JCA 1.5 specification - the specification level is an implementation detail of the connector runtime, not a factor that determines interoperability with Service Bus and SOA. What matters is native API access and co-deployment, and the item's framing turns a compliance statement into a false "most influential factor".
D. Use of Oracle Adapters Extended Architecture (XA) features is completely optional - XA support is how an adapter takes part in a global transaction alongside the rest of the composite. Treating it as purely optional misrepresents how transactional integrity is achieved when an adapter operation must commit or roll back with the SOA transaction.
For adapter questions, separate the two axes: functional coverage comes from native APIs, runtime behaviour comes from co-deployment on the Fusion Middleware WebLogic servers.
Verdict: BD - Proxy Service and Business Service.
Service virtualization means exposing an existing service behind a virtual endpoint that clients call instead of the real one. In Oracle SOA Suite that is done with a proxy service in front and a business service behind it: the proxy service publishes the contract the clients consume and its pipeline carries the message to the business service, which owns the actual route to the third-party HR provider.
Routing is simple and one-for-one, and the payloads need minimum transformation. That is exactly the profile of the proxy service plus business service architecture. The proxy service gives the third-party HR service a virtual home and absorbs protocol and routing detail; the business service holds the concrete endpoint and connection settings. No transformation logic and no process choreography is required, so nothing else is added to the pair.
A. BPEL Process - a BPEL process is an orchestration engine: it brings scopes, state management, dehydration and timers, all of which are dead weight for a synchronous one-for-one call with minimal transformation. The banked key includes it, but the exam's own definition of virtualization here is the proxy/business service architecture, not process orchestration.
C. Mediator - a Mediator is the SOA-composite routing and transformation component. It would work only if you deliberately built an SCA composite to hold the mediation logic, which adds a composite, wiring and a deployment for something Service Bus does natively with a proxy service. The community vote pairs it with the proxy service and drops the business service that actually carries the route to the third party.
When a question says service virtualization with minimal transformation and one-for-one mapping, answer with the proxy service plus business service. Reach for a Mediator when the logic lives inside an SCA composite, and for BPEL only when there is real orchestration, state or compensation to do.
A and C are correct. A BAM 12c alert rule is scheduled against time and calendar, so it can fire between two specified times or on a particular day of the week. Row deletion and a web service invocation are changes and event sources, not conditions you can schedule an alert rule on.
Verdict: A, B and E.
An Oracle Adapter is a reference binding component: it is invoked from inside an SCA composite by a service component. BPEL Process, Mediator and Business Rule are all service components that can be wired to an adapter reference and call it. A proxy service belongs to Service Bus, and a human task is a passive participant that does not invoke adapters.
BPEL Process calls a reference to invoke the back-end operation as a step in the orchestration. The Mediator calls a reference in a routing rule when the content of a message must drive an outbound call. Business Rule can call a reference from a rule action or condition function. All three are SCA service components, and that is precisely the set the item asks for.
C. Proxy Service - a proxy service is the virtual endpoint inside an Oracle Service Bus project. It consumes a business service, and a business service can be adapter-backed, but the proxy service itself is not one of the SOA composite components that consume an Oracle Adapter reference.
D. Human Workflow - a human task service component waits for a person. It is invoked by the BPEL process or by BPM, and it communicates through the human workflow notification channels (email, worklist, UMS). It does not call a JCA adapter reference, so it is not a consumer of Oracle Adapters.
Rule of thumb: adapters hang off reference binding components, and only SCA service components - BPEL, mediator, business rule, human task, spring - can be wired to them. Service Bus resources cannot.
Verdict: C.
The quickest way is declarative, not structural: put an XPath expression in an annotation on the business rule component. The business rule service exposes a skipped annotation; when its expression evaluates to true the rules are not executed at all, and the process moves straight on - no extra activity, no extra branch, no extra scope.
The business rule activity supports a skip condition expressed as an XPath expression in an annotation. Evaluating an expression is a single configuration step with no process restructuring, which is exactly what "quickest way" is asking for: you do not add a switch, you do not duplicate scopes, and the skip decision is recomputed each time the activity is reached.
A. Model a switch activity to conditionally execute the business rule component - it works, but it is the slow way. You add a switch, an extra branch and often an extra scope, which lengthens the process and makes the intent harder to read than a single annotation.
B. Use a skip condition on the business rule component - there is no separate "skip condition" setting to fill in. The mechanism is the XPath expression carried in the annotation on the business rule component, which is what C describes.
D. Use a correlation set on the business rule component - correlation sets exist to correlate messages with a running instance. They have nothing to do with deciding whether rules should run.
When the question asks for the quickest or simplest way to bypass a component in BPEL, look for the declarative annotation - the skipped XPath expression - not a new activity.
Verdict: B.
When the configured ora-retry action for remote faults has used up its retries and no further action is configured, the instance is left open.faulted. An open faulted instance still has recoverable state, which is why it is the state that fault recovery - including auto-recovery - picks up. Nothing rolls back, the instance is not closed, and no fault is rethrown to the caller.
ora-retry, ora-recover, ora-abort and ora-rethrow.ora-retry - retry the activity a configured number of times before moving on to the next action in the policy.Retries are an attempt to keep the work alive, so when they are exhausted the platform does not throw the work away. The instance stays in the open faulted state, holding its state and its pending work, and that is exactly the population the recovery tooling works with - replay, skip or abort can be applied, and auto-recovery can act on it without an operator.
A. The transaction of Composite X rolls back - no. A remote-fault retry sequence is not a rollback trigger; the instance's state is preserved so the work can be recovered, which is the whole reason the fault policy offers recovery actions at all.
C. The instance is marked closed.faulted and is recovered from Enterprise Manager - the closed state is wrong here. Closed means the scopes have terminated and there is nothing left to replay; the exhausted remote-fault instance remains open precisely so that recovery actions remain possible.
D. The default retryFailureAction is initiated and rethrows the error - rethrow is a behaviour you get when a rethrow or retry-failure action is configured. With ora-retry as the action, exhausting the retries leaves the instance faulted and recoverable rather than pushing the error back to the caller service.
Fault policy mnemonic: retry, recover, abort, rethrow - and remember that an instance left recoverable after exhausted retries is open.faulted, not closed.
Verdict: B - Enterprise Manager.
The loggers that matter for SOA diagnostics - oracle.soa., oracle.bpel., oracle.soa.mediator., oracle.soa.bam. - are turned up from Enterprise Manager Fusion Middleware Control, which exposes Log Configuration / Log Levels pages per SOA managed server and lets you raise a logger's level without touching the domain.
More detailed logging of SOA components means raising the level of the SOA component loggers. Enterprise Manager is the console that owns those loggers: from FMW Control you navigate to the SOA managed server, open its Logs section and set the level per logger, which is the supported way to get verbose SOA traces without editing server start-up properties by hand.
A. WebLogic Console - the WebLogic Administration Console configures the WebLogic server log and the logging subsystem, not the SOA component loggers. It cannot set oracle.soa or oracle.bpel logger levels, so it will not give you more detailed SOA logging.
C. BPM Workspace - BPM Workspace is the human workflow worklist for end users: tasks, reports and the process dashboard. It has nothing to do with server-side logging levels.
D. WebLogic Diagnostics Framework (WLDF) - WLDF is the diagnostic and monitoring framework: diagnostic modules watch subsystem metrics, emit diagnostic events and trigger alerts. It is a plausible-sounding distractor, but it is not the tool for turning up SOA component logging; that is done in Enterprise Manager.
Logging configuration in SOA Suite belongs to Enterprise Manager Fusion Middleware Control. Keep WLDF in mind for monitoring and alerting, the WebLogic console for the server log, and never reach for the BPM Workspace for anything administrative.
Verdict: E - Escalate after.
You want a task to move up the management chain if the assignee does not act within a time period. That is the Escalate after setting on the human task: you give it a duration and a number of levels (or route it to the manager), and if the task is still untouched when the duration elapses the workflow service escalates it automatically.
Escalation needs two things: a trigger time and a routing rule. Escalate after supplies both. The workflow service watches the clock, and when it expires with the task still pending, it routes the task the specified number of levels up the management chain - or to the manager - without an administrator doing anything.
A. Never expire - a lifetime setting, not a routing one. With it, an untouched task simply stays with the assignee.
B. Expire after - completes the task at the end of its lifetime using the configured expiry outcome (complete, reassign to a named participant, or fail). It does not walk the management chain.
C. Route to manager - names a single routing destination, not the time-based expiration/escalation setting the question asks for. Escalation is what combines the time condition with management-chain routing.
D. Renew after - sends a further reminder to the current assignee so the task stays alive; it does not move the task.
F. Management chain - that is the participant hierarchy used as the escalation path, not an expiration or escalation setting you configure on the task.
Task timers come in three flavours: expire (end the task), renew (remind), escalate (move it up the chain). If the answer mentions management-chain routing, look for Escalate after.
Verdict: C.
A human task activity returns its outcome as an output string when the task is completed. You branch on that value with a switch containing a case for each outcome of interest, plus an optional otherwise, placed after the human task activity. That is the documented pattern, and it is why the cases are for the outcomes you care about rather than for every outcome imaginable.
Because the outcome comes back as one value, the process must test it. A switch with a case per outcome of interest plus an otherwise is the clean way: each case carries the condition that selects that path, and the otherwise catches everything else so no outcome is left unhandled. It also keeps the branch logic in one visible place in the process.
A. A condition present in each path of the human task activity - the human task activity does not contain per-path outcome conditions. The outcome is a single output you test after the activity.
B. onMessage or pick activities that receive the outcome from the human workflow engine - the human workflow engine does not push the outcome to you as a message. The human task activity invokes it and receives the outcome back on completion; onMessage and pick are for callbacks and other asynchronous messages.
D. A switch with one case for each possible outcome - you do not enumerate every conceivable outcome, and outcomes are extensible. You list the outcomes of interest and add an otherwise, which is what C says.
Human task outcome handling is always: human task activity, then a switch with a case per outcome of interest and an otherwise. Nothing else in BPEL branches on an outcome.
Want the full bank of 57 questions for Oracle SOA Suite 12c Essentials (1Z0-434)? See all practice exams.