Cloud Computing

Fundamental attributes of cloud computing:

• On-demand service
• Network access
• Ressource pooling
• Rapid elasticity
• Measured service

Google Cloud

Storage and Compute ressources are connected with an 1 Petbit/sec connection.

Big Data Team

Resource Hierarchy

Organization > Folder > Project > Resource

Compute	Storage	Big Data	Cloud AI
Computer Engine (IAAS)	Cloud Bigtable	BigQuery	Vertex AI
Google Kubernetes GKE	Cloud Storage	Pub/Sub	Vision AI
App Engine (PAAS)	Cloud SQL	Dataflow	Speech-to-Text
Cloud Function (FAAS)	Cloud Spanner	Dataproc	AutoML
	Cloud Datastorage	Notebooks	Tranaslation

Services

Compute

• Dataproc = Apache Hadoop

Storage

Big Data

• Bigtable = Apache HBase

Machine Learning

Compute

Architectures

Storage

• Standard: Hot data
• Nearline: once per month
• Coldline: once a quarter

• Archive: Backup

  $ gsutil cp sales*.csv gs://acme-sales/data/

Network

Google network 10 GB/s

Regions and Zones

• Region < 5ms latency
• Multi-Regional Services: App Engine, Cloud Datastore, Cloud Storage, BigQuery

Security

IAM

IAM (Identity and Access Management) defines WHO, can do WHAT, on which RESOURCE.

WHO

• Google account of Cloud Identity user: test@gmail.com
• Service account: test@project_id.iam.gserviceaccount.com
• Google group: test@googlegroups.com
• Cloud identity of G Suite domain: example.com

WHAT

Instance Admin > Role

Service	Resource	Verb
compute	instances	list
compute	instances	delete
compute	instances	start

On which RESOURCE

• A policy is set on a resource

Roles

• Primitive
• Predefined
• Custom

Primitve Roles

Predefined Roles

Custom Roles

Service Accounts

• Provide an identity for carrying out server-to-server interactions in a project
• Used to authenticate from one service to another
• Control Privileges used by resources
• Indentified with an email address:
  • PROJECT_NUMBER-compute@developer.gserviceaccount.com
  • PROJECT_ID@appspot.gserviceaccount.com

Google Cloud SDK

• gcloud
• gsutil (Cloud Storage)

• bq (Big Query)

• Available as (Docker Image)[https://hub.docker.com/r/google/cloud-sdk/]
• Available via Cloud Shell

RESTful APIs

Programmatic access to products and services. Enabled through the Google Cloud Platform Console. The API Explorer is an interactive tool that lets you easily try Google APIs using a browser.

Virtual Machines

Compute engine runs virtual machines.

Virtual Private Cloud Network

• Each VPC network is contained in a GCP project.
• Cloud VPC networks are global, subnets are regional
• Subnet have a regional scope

Compute Engine

• Pick memory and CPU
• Pick GPU, if you need them
• SSD
• Windows, or Linux Server
• Define a startup script
• Make snapshots for backups
• Preemptible VMs get terminated after completing the job.

Load-Balancing

Routing tables forward traffic within a network.

Cloud load balancing. Managed service.

Global HTTP(S)	Global SSL Proxy	Global TCP Proxy	Regional	Reginoal internal
Layer 7	Layer 4	Layer 4	Any Traffic	Traffic inside VPC
Different route URLs to different back ends	Specific port	Specific port	Any port	Internal tiers

Cloud CDN (Edge Caches)

Cache Content close to users. Lower latency.

Interconnection Options

VPN	Direct Peering	Dedicated Interconnect	Carrier Peering	Partner Interconnect
Secure mulit-Gbps connection over VPN tunnels	Private connection between you and Google for your hybrid cloud workloads	Connect N x 10G transport circuits for private cloud traffic to Google Cloud at Google POPs	Connection through the largest partner network of service providers	Connectivity between you on-premises network and your VPC network through a supported service provider. 99.99% SLA