L10 - FaaS

Question 1

severless meaing

Answer 1

application owners do not have to worry about the underlying execution infrastructure, mostl importantly physical or virtual servers.

Question 2

What is often used as a synonym for FaaS?

Answer 2

Serverless computing

Question 3

A serverless computing platform may provide one or both of the following

Answer 3

1. Function-as-a-Service (FaaS)
2. Backend-as-a-Service (Baas)

Question 4

What is FaaS?

Answer 4

Function-as-a-Service

Typically provides event-driven computing. Developers run and manage application code with functions that are triggered by events or HTTP requests. Developers deploy small units of code the FaaS, which are executed as needed as discrete actions, scaling without the need to manage servers or any other underlying infrastructure.

Question 5

What is BaaS?

Answer 5

Backend-as-a-Service are third-party API-based services that replace core subsets of functionality in an application. Because those APIs are provided as a service that auto-scales and operates transparently, this appears to the developer to be serverless.

Question 6

CNCF

Answer 6

Cloud native computing forum

Question 7

Pros of FaaS

Answer 7

• no provisioning of servers
• automatic scaling
• reduction of costs. Do not pay for idle VMs and do not run replicas for resilience
• underlying servers shared among different function invocations.

Question 8

Cons of FaaS

Answer 8

• focused on stateless functions (if function should store something then an external storage is needed)
• performance variation due to restart latencies
• not suited for heavy computation workload, own VMs might be cheaper
• limited security: shared VMs, no control over the network

Question 9

Amazon Lambda

Answer 9

• first serverless offering by a public cloud provider
• designed for:
• image or object uploads to Amazon S3
• updates to DynomoDB tables
• responding to website clicks
• reacting to sensor readings from an IoT connected device

Question 10

Lambda functions

Answer 10

Anonymous functions in functional programming:
- not bound to an identifier (no function name)
- often used as arguments being passed to higher-order functions
- used in node.js applications

Question 11

In which programming languages for example can you use Lambda functions?

Answer 11

Node.js
Python
Java
C#

Question 12

4 advantages of using Amazon Lambda

Answer 12

• Bring your own code
• Simple resource model (select power rating up to 3Gb and CPU and network are allocated automatically)
• Flexible use (integrated with other AWS services)
• Flexible authorization

Question 13

3 Lambda execution models

Answer 13

• Synchronous (push)
• Asynchronous (event)
• Stream-based

Question 14

What is Amazon Cognito?

Answer 14

For user management
- secure user sign-up, sign-in, and access control
- supports integration with social and enterprise identity services
- you have a user pool and application id

Question 15

What is OpenWhisk?

Answer 15

What is Apache OpenWhisk?
Apache OpenWhisk is an open source, distributed Serverless platform that executes functions (fx) in response to events at any scale. OpenWhisk manages the infrastructure, servers and scaling using Docker containers so you can focus on building amazing and efficient applications.

Question 16

What are the concepts of OpenWhisk?

Answer 16

• trigger (events)
• actions (code):
• functions or binary code in a container
• Json object as input parameter and as return value
• rules connect triggers with actions
• actions (functions) can also be called from e.g. the CML, API

Question 17

What are the ways to invoce an action on OpenWhisk?

Answer 17

A blocking action is an action that holds up the flow of execution in the system until it completes its execution and returns the result. This means that if one action is executing a blocking operation, no other actions can be executed in the system until the first one is complete.

On the other hand, a non-blocking action does not block the flow of execution in the system. Instead, it allows the system to continue executing other actions even if the non-blocking action has not completed its execution and returned the result. Non-blocking actions are executed asynchronously, which means that the system does not wait for them to complete.

In general, using non-blocking actions is recommended in OpenWhisk to improve the overall performance and scalability of the system.

Question 18

What is blocking in OpenWhisk?

Answer 18

Call action, wait until action is delivered, then get return-value

Question 19

What is non-blocking invocation of actions in OpenWhisk?

Answer 19

Call immediately returns and does not wait for the action to be executed.

Question 20

How does the OpenWhisk Achitecture look like?

Answer 20

Kafka and OpenWhisk can work together by using Kafka as a source of events and OpenWhisk as a way to process and respond to those events in real-time.

Question 21

What is the cold start time when running a container in OpenWhisk?

Answer 21

Time to prepare the container to run the function.

Question 22

What are two optimization tasks in OpenWhisk?

Answer 22

• warm containers
• pre-warmed containers

Question 23

What are warm containers?

Answer 23

• do not delete the containers immediately, grace period
• re-use container for subsequent invocation of same function

Question 24

What is a pre-warmed container?

Answer 24

-keep a pool of initialized containers (stem cell containers) with heavily used language runtimes

Question 25

Does OpenWhisk support different container factories?

Answer 25

Container factories are a set of tools and services that are used to create and manage containers for running cloud applications and services. Yes. e.g.: -Docker - Kubernetes - Mesos

Question 26

How to use OpenWhisk on top of Kubernetes?

Answer 26

OpenWhisk can be run on top of Kubernetes by deploying the OpenWhisk components as containers within a Kubernetes cluster. - Kubernetes manages the control plane of OpenWhisk (Controller, Invoker etc.) - Docker Container Factory (Docker is used to create the containers) K8s container factory: - small number of invokers - user containers are wrapped into pods - invokers start pods on a worker node

Question 27

What are the pros and cons of using OpenWhisk on top of K8s?

Answer 27

Pros: - resuses infrastructure of K8s - can be used with other container runtimes e.g. containerd Cons: - longer cold-start latency (because of K8s overhead)

Question 28

What is the characteristic of OpenWhisk functions and how do they perform complex operations?

Answer 28

- small, simple, stateless - require sequences (chaining) for more complex operations - e.g first detect the language of a document and then translate it

Question 29

What are 4 approaches for function composition in OpenWhisk?

Answer 29

- Client side: client calls the functions - Server side: New function that calls the elementary function and passes the result (double billing problem) - Event driven composition: First function triggers the next --> drawback: you have to write special code - Primitive sequence in OpenWhisk

Question 30

What are the advantages of a primitive sequence in OpenWhisk?

Answer 30

- requires no change or knowledge of the composed functions - requires no change to the client code - does nto inflate costs - build into OpenWhisk runtime

Question 30

What are the advantages of a primitive sequence in OpenWhisk?

Answer 30

- requires no change or knowledge of the composed functions - requires no change to the client code - does nto inflate costs - build into OpenWhisk runtime

Question 31

Alternative to a primitive sequence?

Answer 31

Apache OpenWhisk Composer: - Java script library to program compositions - Workflow compositions: sequence, loops, parallel, if.. - AWS step functions

Question 32

What is the Serverless Framework?

Answer 32

Cross-platform deployment and management tool - supports major FaaS implementation of cloud providers - template for cloud provider and function language - CML tool that works for all cloud providers Automatic instrumentation: enabling monitoring and debugging internet: The Serverless Framework is an open-source framework for building and deploying serverless applications. It provides a platform for defining, deploying, and managing serverless functions and other cloud resources using a simple, declarative syntax. The Serverless Framework supports a variety of cloud providers, including AWS Lambda, Microsoft Azure Functions, Google Cloud Functions, and others, allowing developers to build and deploy their serverless applications across multiple cloud platforms.

Question 33

Name 4 paid FaaS providers and 2 open source providers

Answer 33

- Google Cloud Functions - IBM Cloud Functions - Microsoft Azure Functions - Huawei Cloud Functions Open Source: - OpenWhisk - OpenFaaS - Kubeless

Question 34

When to use serverless?

Answer 34

Use if: - easily decomposable functions - highly-variable demand (fast response time needed) - no high frequency in triggering - overhead of running instances is high - need for tight integration to cloud events

Question 35

When to be cautious about using serverless?

Answer 35

- you don't want to lock-in to a cloud - demand is not variable