Basics of SE

Question 1

Traditional Software Development vs. AI Projekt:

• Goal
• Quality

Answer 1

Traditional:
Goal: Satisfy functional and non-functional
requirements (meet a specification)
AI / ML
Goal: Optimize a certain metric and try to
improve it via experimentation

Traditional
- Quality of the software depends mostly on
code
AI / ML
- Quality depends on multiple factors (input
data, data cleaning, feature engineering,
tuning parameters, ML algorithm, etc.)

AI-enabled software system requires both!

Question 2

Nenne die Elemente

Was sind die 4 Elemente in einem ML-Lifecycle. Nenne auch Beispiele

(3) (3) (2) (5)

Answer 2

• Requirements Engineering: Project Requirements, Goals, Metrics
• Design + Architecture: MLOps, Infrastructure, integrated design
• Implementation + Testing: API-Design, non-functional properties (SLA testing)
• Deployment + Monitoring: Deployment infrastructure, heart beat, monitoring/logging capabilities, version management, feedback loop

Question 3

Define:

Was sind DevOps Goals?

(4 main) (6 more over)

Answer 3

Automate processes (testing, integration, feedback, error report, deployment, etc.)
thereby improving software quality through better (more frequent)
tests, similar environments in dev and prod, actually needed features (fast feedback),
more secure software (using standardized procedures)

• Minimize time to deployment
• Maintain your own product in production
• Enable agile development (iterative process without DevOps possibly infeasible)
• Own the environment (and infrastructure) your system is running on
• Bridge the gap between local and cloud (remote) computing
• Save (reliable) updates and rollouts of software versions

Question 4

Define

Was ist die Idee hinter Virtualization and Containerization?

Answer 4

Abstract from the (i) physical hardware and (ii) the environment your application is running on. Create your virtual
environment anywhere.

Question 5

Define:

Was sind SLA-Tests

Answer 5

Service-Level-Agreement Tests:
Es sind spezifische Tests, die auf meine Performance-Ziele hinarbeiten.

Question 6

Docker:
Was macht LABEL?

Answer 6

Provide key/value metadata (version, topic, etc.)
LABEL <key>=<value>

LABEL version="1.0"
LABEL description="This text illustrates \
that label-values can span multiple lines."

Question 7

Docker:
Was macht ARG?

Answer 7

Define Docker build-time variables
ARG <name>[=<default value>]

FROM busybox
USER ${username:-some_user}
ARG username
USER $username

Question 8

Was ist der Command für:
build image using tag <image_tag> und Version <version>

Answer 8

docker build -t <image_tag>:[<version>]

Question 9

Was ist der Command für:
Lists all the available images

Answer 9

docker image ls or docker images

Question 10

Was ist der Command für:
Create a new instance of the image (run a docker container) in detached mode with name <name> and remove image after exit

Answer 10

docker run [-d] [--name <name>] [--rm <image_tag>]

Question 11

Was ist der Command für:
Lists all currently running containers (use [-a] to show stopped as well)

Answer 11

docker container ls or docker ps

Question 12

Was ist der Command für:
Removes a docker image

Answer 12

docker image rm <image>

Question 13

Was steht für: IaC
Definiere Ziel, Idee und Vorteile

Z: 3, I: 1, V: 6

Answer 13

Infrastructure as Code:

Automate:
- configuration
- deployment
- management of infrastructure resources (virtual machines, containers, networks, storage)

Idea
Nutze wiederverwendbaren Code

Benefits
- In Code Repo
- Consistent Deployment
- Efficient
- Scaling
- Reproducibility of Environments
- Easy Sharing

Question 14

Was sind Acceptance-Tests?

Answer 14

Softwaretests, bei denen bewertet wird, ob ein System seine Geschäfts- und Benutzeranforderungen erfüllt. Letzte Testphase vor Deployment.