2 of 43

Prof. Dr. Christian Baun – 3rd Slide Set Cloud Computing – Frankfurt University of Applied Sciences – SS2019

2/79

Current Situation

Software as a Service (SaaS) Provider runs web applications Customers only need a browser

Platform as a Service (PaaS)

Provider run scalable runtime environment(s) Customers run their own web applications in the infrastructure of the service provider

Infrastructure as a Service (IaaS)

Provider runs physical servers Customers run VMs with (almost) any operating systems and unmodified applications

Customers have administrator privileges in their VMs and define the firewall rules themselves

18.2

Operating System Concepts – 10^th Edition

3 of 43

Software Service Examples

Prof. Dr. Christian Baun – 3rd Slide Set Cloud Computing – Frankfurt University of Applied Sciences – SS2019

3/79

Cloud Computing SaaS, PaaS and IaaS Opportunities and Risks Current Situation

(Free) solutions for building software services exist since > 10 years

Web server: Apache HTTP server, nginx,. . .

Application server for web applications: Apache Tomcat (Java), JBoss (Java), Zope (Python)

Scripting language for dynamic web pages: PHP

Software services exist longer than the term „Cloud Computing“

18.3

Operating System Concepts – 10^th Edition

4 of 43

Prof. Dr. Christian Baun – 3rd Slide Set Cloud Computing – Frankfurt University of Applied Sciences – SS2019

4/79

Opportunities and Risks

Current Situation

Cloud Computing SaaS, PaaS and IaaS

Humans as a Service (HuaaS)

Principle of crowdsourcing

Human creativity is offered for low cost or donated from volunteers Interesting for. . .

Low-skilled jobs

Activities, which a computer cannot to, or requires an unreasonably high development time

Possible applications are among others:

Image recognition

Personal Perspective (subjective) reviews for products Translations

(Product) assignments to (product) categories

Examples of public Cloud HuaaS

Investigation of the British expenses scandal by The Guardian in 2009 GutenPlag, VroniPlag

Marketplace for HuaaS: Amazon Mechanical Turk

In the private Cloud area: HuaaS does not take place

18.4

Operating System Concepts – 10^th Edition

5 of 43

Prof. Dr. Christian Baun – 3rd Slide Set Cloud Computing – Frankfurt University of Applied Sciences – SS2019

5/79

Current Situation

Cloud Computing SaaS, PaaS and IaaS Opportunities and Risks

Recommended literature to Crowdsourcing

Christian Papsdorf. Wie Surfen zu Arbeit wird. Campus (2009)

Consumer Write reviews, develop ideas, create logos,. . .

These value-adding activities are of high economic significance

Companies use the internet culture (participation, engagement,

self-realization,. . . ) to let the users mostly work for free

Why do the consumers accept this and work for free?

18.5

Operating System Concepts – 10^th Edition

6 of 43

Prof. Dr. Christian Baun – 3rd Slide Set Cloud Computing – Frankfurt University of Applied Sciences – SS2019

6/79

Opportunities and Risks

Current Situation

Cloud Computing SaaS, PaaS and IaaS

Pril Competition

On April 1st 2011, Henkel launched a crowdsourcing campaign

Despite the date, it was no joke!

Everyone was able to crate a new design proposal for the 600ml bottle at http://mein.pril.de

There were material prizes to win

The two best designs should go on sale for a short time Users of Facebook were able to vote their favorite Huge feedback: > 30,000 proposals were submitted

18.6

Operating System Concepts – 10^th Edition

7 of 43

Prof. Dr. Christian Baun – 3rd Slide Set Cloud Computing – Frankfurt University of Applied Sciences – SS2019

7/79

Opportunities and Risks

Current Situation

Cloud Computing SaaS, PaaS and IaaS

Pril Competition – Outcome

Not all proposals matched Henkel’s expectation

After a short time, 2 proposals of Peter Breuer (a professional advertising copywriter) became favorites

The chicken proposal was ranked 1st place with several thousand votes ahead 2nd place

Reaction of Henkel: They changed the rules

Now, proposals needed to be previously evaluated and release by a jury

Only after the jury evaluation, the users were allowed to vote for the proposal

Result: Wave of anger

18.7

Operating System Concepts – 10^th Edition

8 of 43

Pril Competition – Manipulation of the Outcome

Prof. Dr. Christian Baun – 3rd Slide Set Cloud Computing – Frankfurt University of Applied Sciences – SS2019

8/79

Cloud Computing SaaS, PaaS and IaaS Opportunities and Risks Current Situation

Things got even worse

Henkel erased comments of angry users Henkel massively reduced the number of votes of several designs

Henkel stated they just „cleaned up“ the results

From this time, the affair went through the press

=⇒ Bad public relations work

Source: Jörg Breithut. Virale Werbefallen – Pril schmeckt nach Hähnchen. 12.4.2011

http://www.spiegel.de/netzwelt/web/0,1518,756532,00.html

Things do not necessarily need to end like this. . .

18.8

Operating System Concepts – 10^th Edition

9 of 43

Prof. Dr. Christian Baun – 3rd Slide Set Cloud Computing – Frankfurt University of Applied Sciences – SS2019

9/79

Opportunities and Risks

Current Situation

Cloud Computing SaaS, PaaS and IaaS

Otto Competition

Otto organized a „model montest“ in 2010

The winner with the most votes was planned to become the new face of the Facebook fan page

Winner against 48,488 other participants: „Der Brigitte“ (Sascha Mörs)

A 22 years old business administration student from FH Koblenz

Otto was not unhappy about the result

Approximately 1.2 million votes were submitted

=⇒ Great public relations work

18.9

Operating System Concepts – 10^th Edition

10 of 43

Amazon’s Mechanical Turk – Cloud Marketplace

Prof. Dr. Christian Baun – 3rd Slide Set Cloud Computing – Frankfurt University of Applied Sciences – SS2019

10/79

Cloud Computing SaaS, PaaS and IaaS Opportunities and Risks Current Situation

March 8th, 2006 – Sam Williams

Pennies for Web Jobs

Speaking to a room filled with Internet developers at the O’Reilly Emerging Technology Conference in San Diego this week, Luis Felipe Cabrera, Amazon’s vice president of software development, outlined a project to harness human intelligence for tasks that computers can’t handle well, such as recognizing objects in images.

The backbone of the plan is a Web-services platform called Mechanical Turk. It uses an auction-style system to farm out complex tasks – complex for a computer, that is – such as recognizing the difference between a human face and a nearby bush, or accurately transcribing an audio recording. Cabrera likes to call the platform „artificial artificial intelligence“ – it’s computers asking humans to do tasks, rather than the other way around.

. . .

Source: http://m.technologyreview.com/web/16519/

Image source: Google image search

18.10

Operating System Concepts – 10^th Edition

11 of 43

Another Crowdsourcing Marketplace – Samasource

Prof. Dr. Christian Baun – 3rd Slide Set Cloud Computing – Frankfurt University of Applied Sciences – SS2019

11/79

Cloud Computing SaaS, PaaS and IaaS Opportunities and Risks Current Situation

http://www.samasource.org

Founded in 2008

Nonprofit project, which gives digital work to people in developing countries

Workers are in Haiti, India, Kenya, Pakistan, South Africa and Uganda

In these countries, school education includes for historical reasons a good basic education in the English language

But these countries don’t have enough jobs

Infrastructure is financed from donations

Donors are among others the Rockefeller Foundation and Google

Wages of about $300 are low from a European perspective, but in developing countries this is a desirable monthly income

Example for a customer: Ask.com

Up to 50,000 requests from Ask.com are processed per month

18.11

Operating System Concepts – 10^th Edition

12 of 43

Google Cloud Print Image source: Google

Prof. Dr. Christian Baun – 3rd Slide Set Cloud Computing – Frankfurt University of Applied Sciences – SS2019

12/79

Cloud Computing SaaS, PaaS and IaaS Opportunities and Risks Current Situation

Provides printing via the Cloud

Internet enabled devices such as netbooks, touchpads and mobile phones get more and more popular

Connection of local printers is difficult

Printer drivers are missing

Some devices lack enough resources Several operating systems (iOS, Android, Windows Phone, Blackberry. . . ) exist

Solution: Google Cloud Print (https://developers.google.com/cloud-print/) HP and Samsung offer compatible printers

Via an e-mail address, the devices can be identified and added as a Cloud printer inside Chrome OS

The user sends his document to be print to the service, sets the printer settings and receives a feedback about the successful job execution

18.12

Operating System Concepts – 10^th Edition

13 of 43

Google Cloud Print (2 Types of Printers)

Prof. Dr. Christian Baun – 3rd Slide Set Cloud Computing – Frankfurt University of Applied Sciences – SS2019

13/79

Cloud Computing SaaS, PaaS and IaaS Opportunities and Risks Current Situation

Google Cloud Print compatible network printer

The printer is registered at the Service Print jobs are sent to a service

The service prepares the print job and forwards it to the printer

Legacy printer (not compatible with Google Cloud Print)

Locally attached printer (USB) or network printer A proxy is installed on a local PC

The proxy registers the printer and sends print jobs to the service

Prepared print jobs are sent via the proxy to the printer

Drawback: The proxy computer must be switched on for printing

18.13

Operating System Concepts – 10^th Edition

14 of 43

Cloud Gaming (1/4)

Prof. Dr. Christian Baun – 3rd Slide Set Cloud Computing – Frankfurt University of Applied Sciences – SS2019

14/79

Cloud Computing SaaS, PaaS and IaaS Opportunities and Risks Current Situation

Cloud gaming services make high-end video games available on low-end devices (older PCs, TVs, mobile phones)

The video games run at the servers of the provider The users’ devices are only used to display the games

The video output is transmitted as a compressed video stream User input is sent to the provider and processed there

18.14

Operating System Concepts – 10^th Edition

15 of 43

Cloud Gaming (2/4)

Prof. Dr. Christian Baun – 3rd Slide Set Cloud Computing – Frankfurt University of Applied Sciences – SS2019

15/79

Cloud Computing SaaS, PaaS and IaaS Opportunities and Risks Current Situation

Drawback: The required compression reduces the optical quality

Problem: The network latency must be low because the user input is transmitted to a remote server and processed there

Period between the user input and results on the local display must be small in order not to disrupt the game flow

Positive side effect for the providers: Pirate copies are impossible

18.15

Operating System Concepts – 10^th Edition

16 of 43

Cloud Gaming (3/4) – Providers

Prof. Dr. Christian Baun – 3rd Slide Set Cloud Computing – Frankfurt University of Applied Sciences – SS2019

16/79

Cloud Computing SaaS, PaaS and IaaS Opportunities and Risks Current Situation

http://www.onlive.com

Available in the U.S. between June 17th 2010 and April 30th 2015

Requirements:

Network link with low latency and < 1000 km distance to the OnLive data center used

The service itself is no longer available

http://www.gaikai.com

Available since February 27th 2011

July 2012: Sony buys Gaikai for $380 million

Is used to stream PS3 games to the PS4 and PC The service itself is no longer available

18.16

Operating System Concepts – 10^th Edition

17 of 43

Cloud Gaming (4/4) – Nvidia Shield Image source: NVIDIA

Prof. Dr. Christian Baun – 3rd Slide Set Cloud Computing – Frankfurt University of Applied Sciences – SS2019

17/79

Cloud Computing SaaS, PaaS and IaaS Opportunities and Risks Current Situation

http://shield.nvidia.com/ Announced at CES January 2013 Available since Q2 2013

Android based

Set top box for TV

Offers 4K TV shows and movies Also as tablet and portable version

18.17

Operating System Concepts – 10^th Edition

18 of 43

Prof. Dr. Christian Baun – 3rd Slide Set Cloud Computing – Frankfurt University of Applied Sciences – SS2019

18/79

Current Situation

Cloud Computing SaaS, PaaS and IaaS Opportunities and Risks

19 of 43

Cloud Operating Systems

Prof. Dr. Christian Baun – 3rd Slide Set Cloud Computing – Frankfurt University of Applied Sciences – SS2019

19/79

Cloud Computing SaaS, PaaS and IaaS Opportunities and Risks Current Situation

Web desktops are also called Cloud operating systems

20 of 43

Prof. Dr. Christian Baun – 3rd Slide Set Cloud Computing – Frankfurt University of Applied Sciences – SS2019

20/79

Cloud Computing SaaS, PaaS and IaaS Opportunities and Risks

Current Situation

Cloud Cooking – the Future ?!

Image source: Heise Zeitschriften Verlag

18.20

Operating System Concepts – 10^th Edition

21 of 43

Snowflake SaaS Data Cloud

Snowflake is a cloud-native Software-as-a-Service platform.
Fully managed data cloud eliminating infrastructure management.
Runs on AWS, Azure, Google Cloud.
Supports warehousing, engineering, analytics, applications.
Key concept: separation of compute, storage, and services.

18.21

Operating System Concepts – 10^th Edition

22 of 43

Snowflake Architecture Overview

Three-layer architecture:
Storage layer – centralized cloud storage.
Compute layer – virtual warehouses for queries.
Cloud services – metadata, optimization, security.
Enables elasticity, concurrency, and cost efficiency.

18.22

Operating System Concepts – 10^th Edition

23 of 43

Compute–Storage Separation

Storage independent from compute.
Multiple compute clusters access same data.
Benefits:
No resource contention.
Pay only for compute usage.
High concurrency analytics.
Flexible scaling.

18.23

Operating System Concepts – 10^th Edition

24 of 43

Data Flow in Snowflake

Data ingestion via batch or streaming.
Processing through SQL and Snowpark.
Consumption through BI tools, dashboards, ML.
Supports structured and semi-structured data.

18.24

Operating System Concepts – 10^th Edition

25 of 43

Virtual Warehouses

Compute clusters execute workloads.
Different sizes available.
Auto suspend/resume saves cost.
Concurrency scaling supported.
Used for ETL, analytics, ML.

18.25

Operating System Concepts – 10^th Edition

26 of 43

Semi‑Structured Data

Supports JSON, Avro, Parquet, ORC, XML.
Variant column type stores flexible schema.
Direct SQL querying without transformation.
Key for modern data lakehouse.

18.26

Operating System Concepts – 10^th Edition

27 of 43

Security & Governance

Encryption at rest and transit.
Role-based access control.
Secure data sharing.
Compliance certifications.
Governance policies.

18.27

Operating System Concepts – 10^th Edition

28 of 43

Performance Optimization

Automatic clustering.
Query optimization services.
Caching improves performance.
Minimal DBA overhead.

18.28

Operating System Concepts – 10^th Edition

29 of 43

Common Use Cases

Enterprise analytics warehouse.
Data lake modernization.
Secure collaboration.
Real-time analytics.
Machine learning pipelines.

18.29

Operating System Concepts – 10^th Edition

30 of 43

Learning Path

Start with official tutorials.
Practice SQL labs.
Explore Snowpark APIs.
Build native data apps.
Continue advanced architecture learning.

18.30

Operating System Concepts – 10^th Edition

31 of 43

Snowflake on AWS – Enterprise Overview

Snowflake runs natively on AWS cloud infrastructure.
Storage uses Amazon S3.
Compute warehouses provisioned dynamically.
Cloud services manage metadata, optimization, security.
Enterprise deployments focus on scalability, security, cost control.

18.31

Operating System Concepts – 10^th Edition

32 of 43

AWS Data Pipeline Integration

Typical enterprise architecture:
Data sources → AWS ingestion → Snowflake → Analytics.
Common AWS services:
• S3 staging
• Glue ETL
• Kinesis streaming
• Lambda automation
Enables modern lakehouse analytics.

18.32

Operating System Concepts – 10^th Edition

33 of 43

AWS Security Architecture

Security integration includes:
IAM authentication integration.
PrivateLink secure networking.
End‑to‑end encryption.
Snowflake RBAC governance.
Supports enterprise compliance.

18.33

Operating System Concepts – 10^th Edition

34 of 43

Enterprise Best Practices

Recommended architecture practices:
Separate compute warehouses per workload.
Use S3 staging zones.
Automate ingestion pipelines.
Monitor credit usage closely.
Implement governance policies early.

18.34

Operating System Concepts – 10^th Edition

35 of 43

Mastering the Snowflake CLI

Modern Developer Workflows for the Data Cloud

Content:

Introduction to the open-source Snowflake CLI (snow).
Moving beyond legacy SnowSQL.
Enabling DevOps, CI/CD, and application development.

18.35

Operating System Concepts – 10^th Edition

36 of 43

What is Snowflake CLI?

Definition: An open-source, developer-centric command-line interface designed for modern Snowflake workloads.
Primary Focus: Not just for SQL, but for building and managing:
Snowpark functions and procedures.
Streamlit in Snowflake.
Snowflake Native Apps.
Snowpark Container Services (SPCS).
The Future: Snowflake’s strategic tool for all new feature enhancements (replacing SnowSQL over time).

Snowflake CLI vs. SnowSQL

Snowflake CLI (Modern):
Extensible: Built for developers, supporting project templates and lifecycle management.
Application-Aware: Dedicated commands for app, streamlit, and spcs.
Python Integration: Built-in support for managing Python packages and requirements.
Open Source: Community-driven and transparent.
SnowSQL (Legacy): Primarily for executing SQL, DDL/DML, and data loading/unloading.

18.36

Operating System Concepts – 10^th Edition

37 of 43

Installation and Setup

Prerequisites: Python 3.10+ (Recommended).
Installation Methods:
pip: pip install snowflake-cli-labs (Note: verify latest package name in docs).
OS Specific: Homebrew (macOS) or MSI installers (Windows).
Verification:

Bash

snow --version snow –help

Config File: Managed via config.toml (typically in ~/.snowflake/).
Structure:

Ini, TOML

[connections.my_conn] account = "orgname-accountname" user = "username" password = "password" role = "ACCOUNTADMIN" warehouse = "COMPUTE_WH"

Environment Variables: You can override settings using SNOWFLAKE_CONNECTIONS_... variables.
Command: snow connection list and snow connection test.

18.37

Operating System Concepts – 10^th Edition

38 of 43

Project Definition (snowflake.yml)

Core Concept: Snowflake CLI uses a "Project First" approach.
snowflake.yml: The central configuration file for your project.
Defines the project type (e.g., Streamlit, Native App, Snowpark).
Specifies artifacts to deploy (files, stages, etc.).
Benefit: Enables consistent deployments across environments (Dev/Test/Prod).

Stage Operations: * snow stage copy: Upload/download files.
snow stage list: View files in a stage.
Object Management:
snow object list <type>: List warehouses, databases, or schemas.
snow object describe <type> <name>: Get details of a specific object.
SQL Execution:
snow sql -q "SELECT 1": Quick ad-hoc queries.
snow sql -f script.sql: Execute local SQL files.

18.38

Operating System Concepts – 10^th Edition

39 of 43

Developing Snowpark & Streamlit

Snowpark: * snow snowpark build: Packages Python code and dependencies.

snow snowpark deploy: Creates functions/procedures in Snowflake.

Streamlit:

snow streamlit deploy: Syncs local Python code to a Snowflake-hosted Streamlit app.

Dependency Management: Automatically handles .zip packaging for Python libraries.

Initialization: snow init --template app_basic
Lifecycle Commands:

snow app run: Creates/updates the application package and instance.
snow app teardown: Cleans up local/remote resources.
snow app validate: Checks the manifest and setup scripts for errors.

18.39

Operating System Concepts – 10^th Edition

40 of 43

Advanced Features & Automation

Cortex AI: Access LLM functions directly via snow cortex.
Notebooks: Manage Snowflake Notebooks from the CLI.
Git Integration: Manage Snowflake Git repositories.
CI/CD Integration: * Perfect for GitHub Actions, GitLab CI, and Jenkins.

Automate deployment of infrastructure and code via the CLI.

Use Virtual Environments: Always install Snowflake CLI in a Python venv.
Keep Config Secure: Use Key-pair authentication instead of plain-text passwords in config.toml.
Template Everything: Use snow init to ensure consistent project structures.
Version Control: Commit your snowflake.yml and project files to Git.

Snowflake CLI is the modern interface for developers building on Snowflake.
Key Command: snow --help is your best friend.
Links:
Official Docs: docs.snowflake.com
GitHub: github.com/snowflake-labs/snowflake-cli
Community: community.snowflake.com

18.40

Operating System Concepts – 10^th Edition

41 of 43

Use case

Capital One
Industry Challenge
Massive volumes of transaction data
Need for real-time fraud detection
Regulatory compliance requirements
Migration from legacy on-prem data warehouses
Snowflake Use Case
Centralized cloud data warehouse
Real-time analytics for fraud monitoring
Scalable compute for risk modeling
Secure data governance for compliance
Business Impact
Faster fraud detection cycles
Improved data accessibility for analysts
Reduced infrastructure maintenance
High-performance concurrent workloads

Netflix

Industry Challenge

Petabytes of streaming behavior data
Personalized recommendation engine
Global user base with high concurrency

Snowflake Use Case

Central analytics platform
User behavior aggregation
A/B testing and experimentation analytics
Marketing performance tracking

Business Impact

Improved personalization algorithms
Faster experimentation cycles
Global analytics scalability

18.41

Operating System Concepts – 10^th Edition

42 of 43

Instacart
Industry Challenge
Real-time demand forecasting
Logistics optimization
Vendor analytics and reporting
Snowflake Use Case
Unified analytics platform
Supply chain performance dashboards
Customer purchasing behavior analysis
Elastic scaling during peak shopping
Business Impact
Improved delivery efficiency
Better inventory forecasting
Scalable infrastructure during demand spikes

Pfizer

Industry Challenge

Global clinical trial data integration
Secure regulatory reporting
Large-scale R&D analytics

Snowflake Use Case

Consolidated research datasets
Secure collaboration across global teams
High-performance research queries

Business Impact

Faster clinical data analysis
Improved collaboration
Strong data governance compliance

18.42

Operating System Concepts – 10^th Edition

43 of 43

Toyota
Industry Challenge
Connected vehicle data
IoT sensor streams
Manufacturing optimization
Snowflake Use Case
Centralized IoT analytics
Predictive maintenance modeling
Global supply chain data integration

Pattern	Snowflake Advantage
Large Data Volumes	Elastic storage & compute separation
Global Access	Cloud-native architecture
Data Sharing	Secure cross-org sharing
Real-Time Analytics	Scalable virtual warehouses
Governance	Built-in RBAC & encryption

Common Patterns Across Industries

Business Impact

Improved manufacturing efficiency
Reduced downtime
Better vehicle performance analytics

18.43

Operating System Concepts – 10^th Edition