ACE Certificate

GCP Compute

• App Engine: Run applications for you (no need to configure servers or other resourses) - PaaS
• Compute Engine: virtual machines - IaaS
• Kubernetes Engine: orchestration platform that manages clusters and organize how containers can run efficiently on a cluster
• Cloud Functions: serverless, event-driven computing service

GCP Storage

• NoSQL Databases
  • Bigtable: wide column database
  • Datastore: document database designed to store and query JSON-like documents
  • Firestore (next generation): document database
• Relational Databases
  • SQL: Managed database service (MySQL and PostgreSQL)
  • Spanner: Horizontally scalable global relational database
• File System
  • Filestore: Managed network file system (good for sharing)
• Object System
  • Storage: Data pockets for large amount of unestructured data.

Hierarchy

Organization -> Folder -> Project -> Resources

IAM

Identity

Represents a user or entity that has privileges to perform actions in GCP. There are 4 types of identities:

Google Account

• Represents developer, admin or other person.
• Email address associate with Google: myusername@gmail.com

Cloud Identity or G Suite Account

• G Suite user is a member of an organization's G Suite domain
• Cloud Identity like G Suite domain, but without access to G Suite services.

Google Group

• Collections of identities
• Useful for assigning roles to multiple users
• Identities in group acquire roles assigned to the group
• Identities lose roles when removed from group

Service Account

• Account associated with an application or instance rather than a user
• Can create as many as needed
• Represented as an email: 727080621142-compute@developer.gserviceaccount.com

Access Control Concepts

• Resources
• Permissions
• Roles
  • Predefined
  • Custom
  • Primitive

Policies

• Collections of statements that define which users have access to some resource
• Attached to a resource
• Roles attached to identitites, policies attached to resources
• Policy hierarchy

Compute Engine

Instance Groups

• Collection of instances that are managed as single entity
• 2 Types:
  • managed instance group
  • unmanaged instance group

Managed Instance Group

• Multiple identical VMs
• Configuration defined in instance template

Features:

• Autoscaling
• Autohealing
• Multi-zone deployments
• Auto-updating

Unmanaged Instance Group (for legacy clusters)

• Multiple possibly heterogeneous VMs
• Used to apply load balancing accross heterogenerous group of instances
• In generl, recommended for legacy clusters only
• No autoscaling, autohealing, or auto-updating

Autoscaling Configuration

• Automatically add or remove instances based on workload
• Set minimum and maximum number of instances
• Based on target utilization
  • CPU utilization
  • HTTP load balancing servicing capacity
  • Stackdriver metrics

Kubernetes Engine

• Managed service
• Provides Kubernetes Clusters
• Kubernetes
  • Runs containers on cluster of virtual or physical machines (VMs in GCP)
  • Known as container orchestration

Features

• Load balancing of workloads
• Node pools to segment nodes within a cluster
• Automatic scaling and updating
• Stackdriver monitoring

Kubernetes Process Objects

Services wraps Deployments which consists of Pods.

• Pods: Mecanism to encapsulate and run a container
• Deployment: Especification of how pods run (pod replicas, etc)
• Services: Mecanism for service discovery. For example, a way of assosiating a particular IP with a particular pod.

Kubernetes Storage Objects

• Persistent Volumes
• Persistent Volume Claims

App Engine

• Platform as a Service
• Serveless managed service
• Well suited to run microservices
• Originally limited to language specific runtimes
• Can now run containers
• Two types:
  • App Engine Standard
  • App Engine Flexible

App Engine Standard

• Run applications in preconfigured containers
• Applications run in secure sandbox
• Autoscales
• Standard Runtines: Python, Java, Node.js, PHP, Ruby, Go
• Components: Application, Service, Version and Instance

App Engine Flexible

• Runs customized Docker containers
• Not as restricted as sandbox used in App Engine Standard
• Support for Java, Python, Ruby, PHP and other languages but can be customized
• Run locally without App Engine SDK
• Custom health checks
• Higher CPU and memory limits
• Applications are run in regional managed instance groups, not zonal managed instance groups

App Engine Scaling

• Applications execute on App Engine managed instances
• Scale based on load when running dynamic instances
• Cona configure resident instances to run at all times but when autoscaling enabled, use dynamic instances

Configuring Autoscaling

• Configured in app.yaml
• Configuration options include:
  • Target_cpu_utilization
  • Target_throughput_utilization
  • Max_concurrent_requests
  • Max_pending-latency
  • Min_pending_latency

App Engine Traffic Splitting

• If more than one version of an app is running, you can scplit traffic between versions
• 3 ways to split traffic
  • IP address
  • HTTP cookie
  • Random selection

Traffic Splitting by IP address

• Provides stickiness so all traffic from IP address handled by same instances
• Can create problems if state is maintained and user changes IP address

Traffic Splitting by Cookie (preferred)

• HTTP request header for cookie named GOOGAPPUID contains hash value
• Hash value determines instance to route traffic too
• Tolerates user changing IP address

Traffic Splitting by Random Selection

• Useful for stateless applications
• No need to ensure traffic from a client always goes to the same instance
• Used when no GOOGAPPUID cookie

Cloud Functions

• Serveless compute service
• Execute code in response to events based on triggers
• Available for events in:
  • Cloud Storate
  • Cloud Pub/Sub
  • HTTP
  • Firebase
  • Stackdriver Logging

Events

• Storage
  • Upload
  • Delete
  • Archive
• Cloud Pub/Sub
  • Publish a message
• HTTP
  • POST
  • GET
  • PUT
  • DELETE
  • OPTIONS

Triggers

• Declaration of interest in an event
• Bind function to a trigger to execute it
• May be associated with a resource
  • Cloud Storage bucket
  • Cloud Pub/Sub topic
  • Etc

Functions

• Node.js
• Python
• Go

GCP Storage

Types of storage

Cloud Storage

Storage

• Unstructure data
  • Images
  • Video
  • Text
• Archived data
• Temporary storage between services
• Global access, Web accessible

Managed SQL Database

Relational databases:

• Structured data
• ACID transactions
• Complex queries
• Joins

Cloud SQL

• Up to 10 TB
• Regional

Cloud Spanner

• Horizontally scalable
• Global

Managed NoSQL Database

NoSQL databases:

• Semi-structured, flexible schema
• No joins

Datastore/Firestore

• Document, JSON structure
• Hierarchical structure

Bigtable

• Petabyte scale
• Wide column
• Low latency writes
• Analytics

Analytic Database

BigQuery

• Data warehousing
• Petabyte scale
• SQL query language
• Some support for joins
• Not transactional

Know about Storage

• Object vs Persistent Disk vs Managed Database
• Cloud Storage
  • Regional, multi-regional, Nearline, Coldline
  • Lifecycle policies
  • gsutil
• Persistents Disks
  • Used with VMs
  • Persist data even when VM shuts down

SELECT name, gender, SUM(number) AS total FROM bigquery-public-data.usa_names.usa_1910_2013 GROUP BY name, gender ORDER BY total DESC LIMIT 10

Storage Transfer Service, Transfer Service for on-premises data Transfer Appliance in the new Data Transfer section of the Cloud Console.

Networking

• Q: What service allows VPCs to share resources if the VPCs are in different organizations?
• A: VPC Peering
• Q: When creating a VPC and creating default subnets, where are the subnets created?
• A: In all regions
• Q: What two resources must be created to implement a GCP VPN?
• A: VPN Gateway and a tunnel

Basics

IP Address

• 2 parts
  • Routing prefix
  • Device indentifier
• Routing prefix specified in CIDR notation
  • 192.168.2.1/24
  • 24 bits used for routing prefix/subnet
  • 8 bits used for device address

Understand about Networking

• Purpose of VPC
• VPC Peering vs Shared VPC (same organization)
• Hybrid Cloud implementation options:
  • VPN
  • Cloud Interconnect
  • Peering

Virtual Private Cloud (VPC)

• Global resource
• Subnets in regions
• Resources can communicate using private IP addressing
• Can share VPCs within organization
• Can peer VPCs accross organizations

Virtual Private Network (VPN)

• LInks VPC networks to on-premises network
• Implemented with IPSec
• Traffic routed over public internet
• Up to 3 Gbps

Cloud Interconnect

• Google networking service
• Connect to Google network
  • Directly, at least 10Gbps, up to 100 Gbps
  • Partner, from 50 Mbps to 10 Gbps

Peering

• Low level network connection
• Linking networks
• Traffic routed using BGP
• Does not use GCP objects

Load Balancers

• Global or regional
• External or Internal
• Traffic type
  • 3 Global: HTTP(S), SSL Proxy, TCP Proxy
  • 2 Regional
    • Internal TCP/UDP (only internal load balancer)
    • Network TCP/UDP

When configuring a Load Balancer:

• Need an instance group already configured
• Specify the type of load balancer
• Associate instance group with load balancer
• Specify health check
• Specify back end service
• Specify front end service

Cloud DNS

When creating a DNS zone, by default two records are greated: NS and SOA

Operations (Stackdriver)

• Q: What Stackdriver service is used to monitor application and host performance metrics?
• A: Monitor
• Q: What Stackdriver service is used to search for information in the audit log?
• A: Logging
• Q: What Stackdriver service is used to monitor variables in running programs?
• A: Debug
• Q: What Stackdriver service is used to analyze performance issues in distributed systems?
• A: Trace

What you need to know in depth

IAM

• Roles and permission
• Identity types and when to use them

Networking

• VPC, Subnets, CIDR Blocks
• Firewalls, DNS
• VPN, Cloud Interconnect, Peering

Compute

• Compute Engine
• Kubernetes Engine
• App Engine
• Cloud Functions

Storage

• Cloud Storage
• BigQuery

Cloud Pub/Sub

Used for decoupling services and allow one service to write messages to a topic, and another application to consume those messages

Stackdriver

Know how to choose between

• Cloud SQL
• Cloud Spanner
• Cloud Datastore/Firestore
• BigQuery
• Cloud Storage

Know when to use

• Cloud Datproc: managed Hadoop (Apache Pig, Hive and Spark)
• Cloud Dataflow: stream processing and batch processing platform (Apache Beam)
• Deployment Manager

What to use