Machine-translated by Google. Questions are authored in English.
Top

Serverless Computing Questions and Answers

Answer: A

Servers obviously still run the code; what disappears is the customer's involvement in provisioning, patching and scaling them. The name describes the operational experience rather than the underlying machinery, which trips up a great many people meeting it for the first time.

Please log in or register to report this question.

2.

Assertion (A): Serverless does not mean that no server is involved.
Reason (R): The provider still runs the code on servers that it provisions and operates itself.

Answer: B

Both are true and the reason gives what the name conceals, and the customer's gain is that none of that provisioning is theirs to do. The abstraction is about responsibility rather than about the physical arrangement of the hardware.

Please log in or register to report this question.

Answer: C

The event may come from an HTTP request, a queue message, a storage upload or a timer. Between events nothing runs and nothing is billed, which is the central economic property of the model. An unused function costs the same as one that was never deployed at all.

Please log in or register to report this question.

4.

Match each event source with a typical function it triggers.
List I
A. HTTP gateway
B. Object storage
C. Message queue
D. Timer
List II
1. consuming work items asynchronously
2. handling an interactive API request
3. running a scheduled maintenance task
4. processing a newly uploaded file

Answer: D

The same function runtime serves all four, and only the trigger configuration differs. This uniformity is much of the appeal, since one deployment mechanism covers work that would otherwise need several different kinds of server.

Please log in or register to report this question.

Answer: A

The platform must allocate a sandbox, load the runtime and run the initialisation code before the handler executes. Once warm the environment is reused for subsequent invocations, so the penalty is paid only occasionally.

Please log in or register to report this question.

6.

Consider the following statements about cold starts.
1. A larger deployment package generally lengthens the cold start.
2. Work done outside the handler is paid for once per environment.
3. Every invocation of a function incurs a cold start.

Answer: B

Correct answer: 1 and 2 only

Statements 1 and 2 are correct: a larger package takes longer to download and unpack, and initialisation code outside the handler runs once when the environment is created rather than on every call - which is why database clients and SDK objects are built there. Statement 3 is wrong: a warm environment is reused for subsequent invocations, and in a busy function the great majority of calls incur no cold start at all.

Please log in or register to report this question.

7.

Assertion (A): Expensive setup work belongs outside the function handler.
Reason (R): Code outside the handler runs once per execution environment rather than once per invocation.

Answer: C

Both are true and the reason gives the mechanism, and opening a database connection or loading a model there amortises the cost across every warm invocation. Putting the same code inside the handler repeats it needlessly every time.

Please log in or register to report this question.

Answer: D

Correct answer: increasing the function's configured timeout

The timeout caps how long an invocation may run and has no bearing on how long the environment takes to start. The other three all attack the cold start directly: a smaller package downloads and unpacks faster, a lightweight runtime initialises faster, and provisioned concurrency keeps environments warm and ready so the start is not paid at all.

Please log in or register to report this question.

9.

State whether the following is true or false.
A function should be written so that it holds no state between invocations.

Answer: A

Any given invocation may land in a fresh environment, and the platform may discard environments at any time. State that must survive belongs in a database, a cache or object storage rather than in the function's own memory.

Please log in or register to report this question.

10.

Assertion (A): Local disk in a function environment is not a reliable place to keep data.
Reason (R): The environment may be discarded at any time and a later invocation may not see it.

Answer: B

Correct answer: Both A and R are true, and R is the correct explanation of A

Both are true and the reason gives the cause. The execution environment is created and destroyed by the platform on its own schedule, and successive invocations may land in different environments, so anything written to local disk may be gone or may simply not be visible. The disk is useful as a cache within one invocation or across invocations that happen to share an environment, and for nothing that must be there.

Please log in or register to report this question.

Loading…

Other Subjects

Tags: Serverless Computing Questions and Answers || Serverless Computing MCQ Questions and Answers || Serverless Computing GK Questions and Answers || Serverless Computing GK MCQ Questions || Serverless Computing Multiple Choice Questions and Answers || Serverless Computing GK || GK on Serverless Computing || Cloud Computing Questions and Answers || Cloud Computing MCQ Questions and Answers || Cloud Computing GK Questions and Answers || GK on Cloud Computing