| Yesterday | Today | Tomorrow |
|---|---|---|
| Limited number of tools and vendors | Many tools and vendors to work with | Integrated tools and vendors |
| One platform - few devices | Multiple platforms - many devices | Connected platforms and devices |
| Data is scarce but manageable | Overabundance of data | Data is used for important business decisions |
| IT has major influence and control | IT has limited influence and control | IT is strategic to the business |
| People only work when they are at work | People work wherever they want | People have access to what they need, wherever they are |
Lecture 2
Introduction to the cloud - extended lab
Amit Arora, Abhijit Dasgupta, Anderson Monken, and Marck Vaisman
Georgetown University
Fall 2023
Look back and ahead
- Great use of Slack
- Big data definitions
- Used the Shell in Linux on a virtual machine through Codespaces
Deadlines
Lab 1: Linux Basics Due Aug 29 6pm- Quiz 1: Background Skills Due Sept 2 12pm
- Assignment 1: Python Skills Due Sept 5 11:59pm
- Lab 2: Cloud Tooling Due Sept 5 6pm
- Assignment 2: Shell & Linux Due Sept 11 11:59pm
Agenda and Goals for Today
Lecture
- Background of the cloud and related concepts
- Tour of the cloud services that are used in the course
Lab
- AWS SageMaker Setup
- AWSAcademy Learning Management System (LMS)
- AWS Console Overview
- SageMaker Setup
- Azure Machine Learning Setup
- Create resource group
- Deploy Azure Machine Learning
- Create Azure Machine Learning compute instance
- Work with VSCode on Azure Machine Learning
- Browser-based VSCode
- Locally-based VSCode (if time permits)
- Lab deliverable on GitHub Classroom
Are you ready for the Cloud?
Working on a single machine
You are most likely using traditional data analysis tools, which are single threaded and run on a single machine.
The BIG DATA problem
Is Moore’s Law Dead?
Is Moore’s Law Dead?
New Hardware
Need
- The demand for data processing will not be met by relying on the same technology.
- The key to modern data processing is new semiconductors
- Not just squeezing more transistors per area
- Need new compute architectures that are built and optimized for specialized functions
- Specialized edge hardware for Edge Computing
- While many declare Moore’s Law to be broken or no longer valid, in reality it’s not the law that is broken but rather a heat problem.
What
Graphic Processing Units (GPUs)
Field Programmable Gate Arrays (FPGAs)
Data Processing Units (DPUs)
Photonic computing
So, we can’t store or process data on a single machine, what do we do?
We distribute
More CPUs, more memory, more storage!
How do we do that?
Simple, we use the cloud
Simple, we use the cloud
Cloud computing is a big deal!
Benefits
Provides access to low-cost computing
Costs are decreasing every year
Elastic
PAAS works!
Many other benefits…
What is the claaaaaaawd (the cloud)
What is the cloud?
\kloud\ noun
the practice of storing regularly used computer data on multiple servers that can be accessed through the Internet
What is the cloud?
\kloud\ noun
the practice of storing regularly used computer data on multiple servers that can be accessed through the Internet
Using someone else’s computer(s)
NIST Definition
Cloud Services Models
Cloud Services Models Analogy
The evolution of the Cloud
What does the cloud look like?
Virtual Visit to a Microsoft Azure Data Center
Microsoft Azure Data Center in Boydton, VA
Loudon County, VA is called “CLoudon”
How data centers power VA’s Loudon County: https://gcn.com/articles/2018/10/12/loudoun-county-data-centers.aspx
The heart of “The Cloud” is in Virginia: https://www.cbsnews.com/news/cloud-computing-loudoun-county-virginia/
CBS Sunday Morning Visits the Home of the Internet in Loudoun County: https://biz.loudoun.gov/2017/10/30/cbs-sunday-morning-visits-loudoun/
NoVa Data Center Map
70% of the world’s internet traffic passes through Loudon County, VA
Overview of services used in this class
AWS
Azure
Services we will be using
| Category | AWS Service | Azure Service | Description |
|---|---|---|---|
| Compute (virtual machines) | Elastic Compute Cloud (EC2) Instances | Virtual Machines | Virtual servers allow users to deploy, manage, and maintain OS and server software. Instance types provide combinations of CPU/RAM. Users pay for what they use with the flexibility to change sizes. |
| Storage | Simple Storage Services (S3) | Blob storage | Object storage service, for use cases including cloud applications, content distribution, backup, archiving, disaster recovery, and big data analytics. |
| Networking | Cloud virtual networking | Virtual Private Cloud (VPC) | Virtual Network |
Services we will use
| Category | AWS Service | Azure Service | Description |
|---|---|---|---|
| Machine Learning Platforms | AWS Sagemaker | Azure Machine Learning | Machine learning platforms that abstract away details of networking, compute, networking, etc. so you can focus on the analytics, big data, and modeling. |
| IoT Streaming | Kinesis Firehose, Kinesis Streams | Event Hubs | Services that allow the mass ingestion of small data inputs, typically from devices and sensors, to process and route the data. |
Services we may use
| Category | AWS Service | Azure Service | Description |
|---|---|---|---|
| Relational database | RDS | SQL Database Database for MySQL Database for PostgreSQL |
Managed relational database service where resiliency, scale, and maintenance are primarily handled by the platform. |
| NoSQL / Document DB | DynamoDB SimpleDB Amazon DocumentDB |
Cosmos DB | A globally distributed, multi-model database that natively supports multiple data models: key-value, documents, graphs, and columnar. |
| Big Data Processing | EMR | Databricks | Apache Spark-based analytics platform. |
| Big Data Processing | EMR | HDInsight | Managed Hadoop service. Deploy and manage Hadoop clusters in Azure. |
Mapping of AWS <–> Azure services
Interfacing with cloud services
- portal / graphical interface
- command line
- REST APIs
Lab Time!
Set Up GitHub Personal Access Token
Glossary
| Term | Definition |
|---|---|
| Local | Your current workstation (laptop, desktop, etc.), wherever you start the terminal/console application. |
| Remote | Any machine you connect to via ssh or other means. |
| EC2 | Single virtual machine in the AWS cloud where you can run computation (ephemeral) |
| SageMaker | Integrated Developer Environment in AWS where you can conduct data science on single machines |
| Azure Machine Learning | Azure's managed machine learning platform that also provides interactive computing |
| VSCode | Visual Studio Code |
| Compute Instance | A virtual machine running inside Azure or AWS |
| Ephemeral | Lasting for a short time - any machine that will get turned off or place you will lose data |
| Persistent | Lasting for a long time - any environment where your work is NOT lost when the timer goes off |
DSAN 6000 | Fall 2023 | https://gu-dsan.github.io/6000-fall-2023/