The Client-Server Database Questions and Answers

Answer: D

Correct answer: Include the use of a file server.

A file server is an earlier architecture, not a component of client/server: it ships whole files across the network and leaves the processing to the client, which is what client/server was devised to replace. Understanding requirements, checking network capacity and confirming that the tools work with the chosen middleware are all genuinely necessary.

Answer: A

Correct answer: True

A massively parallel system gives each processor its own dedicated memory and no shared memory at all, so each must run its own complete copy of the operating system. Nothing is shared between them, which removes the memory contention of a symmetric system and replaces it with the need to pass messages between nodes.

Answer: A

Correct answer: True

Because all the processors share one memory, a large data set can sit in memory once and be worked on by every processor without being copied. Where the data must stay resident to meet the performance target, that shared memory is exactly the property required - and it is also the source of the contention that limits how far the architecture scales.

Answer: B

Correct answer: False

Stored procedures generally take longer to write than the equivalent in a rapid application development tool, and that is the recognised cost of using them. What is bought with the extra effort is better performance, less network traffic and consistent enforcement of rules across every application.

Answer: A

Correct answer: True

Windows applications are structured around a loop that waits for events - a click, a keystroke, a timer - and dispatches each to a handler. This is why the code is organised as responses to things that happen rather than as a sequence of steps, and it is the model that graphical interfaces generally follow.

Answer: A

Correct answer: True

Presentation logic covers everything at the boundary with the user: formatting output for the screen or printer, and accepting and interpreting input. Keeping it separate from processing and storage is what allows the same application to acquire a different front end without altering its rules.

Answer: A

Correct answer: True

QBE lets a user build a query by filling in a grid rather than writing a statement, and it is the direct-manipulation query interface most widely available - present in some form in nearly every desktop database product. It has never been standardised, so the details differ between products.

Answer: A

Correct answer: True

ODBC works by defining a common interface and requiring each database vendor to supply a driver implementing it for their product. The application calls the interface, the driver translates, and so the same code reaches any database for which a driver exists - but nothing at all without the specific driver installed.

Answer: B

Correct answer: Useful for situations where data must remain in memory to process.

Symmetric multiprocessing gives every processor access to one shared memory, so a large working set can be held once and worked on by all of them. The other options are wrong in detail: one copy of the operating system is shared rather than resident in each processor, and memory contention is the characteristic difficulty of the architecture rather than an unlikely one.

Answer: B

Correct answer: It is similar in concept to ODBC.

JDBC does for Java what ODBC does generally: a common call interface with a driver for each database, so the application code is written once and the driver deals with the vendor’s particulars. The differences follow from that - it is specific to Java rather than language independent, and so is not available to object-oriented languages in general.