A company wants to use ODM for asynchronous processing with large payloads and very strict performance guidelines. They want to keep data extraction and execution close to reduce any latency…

Question

Options

• AA stand-alone POJO session on a single VM.
• BA custom MDB with a stateful or stateless rule session.
• CA J2SE rule session to create a tightly coupled application with performance gains.
• DAn HTDS web service and increase the size of the connection pool based on

• A
  69% (35)
• B
  10% (5)
• C
  4% (2)
• D
  18% (9)

Explanation

A stand-alone POJO session embeds the ODM rule engine directly inside the calling application's JVM, which eliminates network hops and serialization overhead - exactly what you need when payloads are large and latency budgets are tight. Because rule execution happens in-process alongside data extraction, data never has to travel across a network boundary, satisfying the "keep data and execution close" requirement. Horizontal scaling is achieved by deploying the application (with the embedded rule engine) across multiple JVMs or nodes, supporting concurrent execution from different applications without a shared bottleneck.

Why the distractors are wrong:

• B (Custom MDB): MDBs use JMS, which introduces message serialization/deserialization overhead - a major problem for large payloads and strict performance SLAs.
• C (J2SE rule session): J2SE sessions are for standalone, non-enterprise contexts and lack the scalability mechanisms needed for multiple concurrent calling applications.
• D (HTDS web service): HTDS exposes rules over HTTP, adding network round-trips and payload marshaling that directly conflict with the low-latency requirement.

Memory tip: Think "POJO = co-located." When the exam mentions large payloads + low latency + scale-by-adding-nodes, the answer is always the option that avoids the network - and POJO is the only choice that lives inside the application.

No community discussion yet for this question.