前言

官网: https://cloud.google.com/learn/certification/cloud-digital-leader

知识点

1. Regions and Zones

Regions and Zones

  • Google provide 20+ regions around the world
  • High Availability, Low Latency, Global Footprint, Government Regulation

  • Each Region has three or more Zones
  • Advantage of Zones: Increase availability and fault tolerance

2. Compute

Compute Engine

  • Compute Engine (GCE): Provision & Manage Virtual Machines

  • Create and manage lifecycle of VM instances
  • Load balancing and auto scaling for VM instances
  • Can attach storage, manage network connectivity and configuration

VM Setup

  • Startup Script: For boostrapping
    • Install OS patches or software when an VM instance is launched

  • Instance Templates: Specify VM instance details
    • Used to create VM instances and managed instance groups
    • Cannot be updated, need to be deleted and modified

  • Custom Image: Prefered way than Startup Script (推荐的做法)
    • Can have OS patches and software pre-installed

Use Discount

  • Sustained Use Discounts
    • Automatic discounts for running VM instances for significant portion of the billing month (意思就是在一个月内使用这个 instance 到一定程度就可以获得 discount)

  • Committed Use Discounts
    • 1 year or 3 year reservations for workloads with predictable resource needs

Preemptible VM

  • Preemptible VM: Short-lived cheaper (up to 80%) compute instances for non time-critical, fault-tolerant workloads (只能存在 24 小时)

Spot VM

  • Spot VM: Latest version of preemptible VM (这个就是 Spot Instance)
  • Does not have a maximum runtime, 没有 24 小时限制

Sole Tenant Nodes

  • Sole Tenant Nodes: 类似 Dedicated Host 或者 Dedicated Instance
  • Virtualized instances on hardware dedicated to one customer

Custom Machine Types

  • Custom Machine Types: 可以自定义 CPU, memory, GPU

VM costs

  • 2 primary costs in running VMs using GCE
  • Infrastructure cost (VM) & Licensing cost (OS)

Instance Groups

  • Instance Group: Group of VM instances managed as a single entity
  • Two Types of Instance Groups
    • Managed: Identical VMs created using a template
    • Unmanaged: Different configuration for VMs in same group
  • Location can be Zonal or Regional (Regional gives higher availability)

Managed Instance Groups

  • Managed Instance Groups: Maintain certain number of instances
    • Detect application failures using health check (self healing)
    • Increase and decrease instances based on load (auto scaling)
    • Add load balancer to distribute load

Cloud Load balancing

  • Cloud Load balancing: Distribute traffic across VM instances in one or more regions

Bare Metal Solution

  • Bare Metal Solution allows customers to run specialized workloads on dedicated hardware provided by Google Cloud. It enables leverage of existing software licenses and configurations without any modifications

3. Managed Services

IAAS (Infrastructure as a Service)

  • Use only infrastructure from cloud provider
    • Using VM to deploy your applications or databases

PAAS (Platform as a Service)

  • Use a platform provided by cloud
    • You are responsible for Configuration and Application code

SAAS (Software as a Service)

  • Centrally hosted software (mostly on the cloyd), minimize the level of management
    • Offered on a subscription basis (pay-as-you-go)

Containers

  • Container ensure we have one way of deploying any microservices

  • Create Docker images for each microservice, include
    • Application Runtime
    • Application code and Dependencies
  • Runs the same way on any infrastructure
  • Advantage: light weight (No Guest OS), isolation for containers, cloud neutral

  • Container Orchestration (Many build upon Kubernetes)
    • Feature: Auto Scaling, Load Balancing, Self Healing, Fast Deployment

Serverless

  • Focus on code and the cloud managed service take cares of other stuffs
  • Pay for use

Shared Responsibility Model

  • Securiet in cloud is a Shared Responsibility
  • SaaS: Content + Access Policies + Usage
  • PaaS: SaaS + Deployment + Web Application Security
  • IaaS: PaaS + Operations + Network Security + Guest OS
  • Google Cloud is always responsible for Hardware, Network, Audit Logging etc

GCP Service Category

  • Compute Engine: IAAS (provide customize OS)
  • Google Kubernetes Engine: CAAS
  • App Engine: PAAS
  • Cloud Functions: FAAS
  • Cloud Run: CAAS (run one container quickly)

4. Managed Compute Service in GCP

App Engine

  • App Engine: Simplest way to deploy and scale your applications in GCP
    • Automatic load balancing & auto scaling
    • Managed platform update & health monitoring
    • Application versioning
    • Traffic splitting
  • No usage charges -> pay for resources provisioned
  • App Engine does not provide the granular control over the virtual machine infrastructure

  • Compute Engine vs. App Engine
    • Compute Engine is IAAS, App Engine is PAAS
    • App Engine is serverless
    • Compute Engine you have more Responsibility than App Engine
    • App Engine is lower flexibility

App Engine Environments

  • Standard: Applications run in language specific sandboxes

    • Run in language specific sandbox
    • Supports scale down to Zero instances

  • Flexible: Application instances run within Docker containers

    • Support ANY runtime
    • CANNOT scale down to Zero instances

Google Kubernetes Engine (GKE)

  • Google Kubernetes Engine: Managed Kubernetes service
  • Minimize operations with auto-repair (repair failed nodes) and auto-upgrade (use latest version of K8S always) features
  • Provides Pod and Cluster Autoscaling
  • Two Modes: Standard (You manage) & Autopilot (GKE manage)

Cloud Functions (GCF)

  • Cloud Functions: Excute some code when an event happen
  • Don’t worry about servers or scaling or availability (only worry about your code)
  • Pay only for what you use
  • Time Bound - Default 1 min and MAX 60 minutes(3600 seconds)

Cloud Run & Anthos

  • Cloud Run: “Container to Production in Seconds”

    • Fully managed serverless platform for containerized applications

  • Cloud Run for Anthos: Deploy your workloads to Anthos clusters running on-premises or on Google Cloud

Compute Service in GCP

  • Create Virtual Machines: Compute Engine
  • Create a group of similar VMs: Managed Image Group
  • Distribute load among VMs: Cloud Load Balancing
  • Simplify setting up web application: App Engine
  • Easiest way to run one container: Google Cloud Run
  • Orchestrate containers: Google Kubernetes Engine
  • Build serverless event driven functions: Cloud Functions
  • Manage multi-cloud and on-premise Kubernetes clusters: Anthos

5. Storage

Block Storage & File Storage

Block Storage

  • Persistent Disk: Network Block Storage (Zonal or Regional)
  • Local SSDs: Local Block Storage

File Storage

  • Filestore: High performance file storage

Cloud Storage

  • Most popular, very flexible & inexpensive storage service
  • Store large objects using a key-value approach (Object Storage)
  • Provides REST API to access and modify objects
  • Store all file types - text, binary, backup & archives

Storage Classes

  • Different kinds of data can be stored in Cloud Storage
    • Can I pay a cheaper price for objects I access less frequently
  • Storage classes help to optimize your costs based on your access needs

  • Standard: Frequently used data/Short period of time
  • Nearline storage: Read or modify once a month on average (30 天内读写一次)
  • Coldline storage: Read or modify at most once a quarter (90 天内读写一次)
  • Archive storage: Less than once a year (365 天内读写一次)

Object Lifecycle Management

  • How do you save costs by moving files automatically between storage classes

    • Object Lifecycle Management

  • Identify objects using conditions based on

    • Age, CreatedBefore, IsLive, MatchesStorageClass, NumberOfNewerVersions
    • Set multiple conditions: all conditions must be satisfied for action to happen

  • Two kinds of actions

    • SetStorageClass actions (change from one storage class to another)
    • Deletion actions (delete objects)


Transferring data from On-Premises to Cloud

  • Most popular data destination is Google Cloud Storage
    • Online Transfer: Use gsutil or API to transfer data to Google Cloud Storage (data < 1 TB or from on-premise)
    • Storage Transfer Service: Recommended for large-scale (petabytes) online data transfers from your private data centers, AWS, Azure, and Google Cloud (data > 1 TB or from another cloud)
    • Transfer Appliance: Physical transfer using an appliance (data > 20 TB or time > 1 week)

Storage in GCP

  • A shared space for collaborating on media projects that involve large files: Filestore (File Storage)
  • A cost-effective solution to store and serve a large amount of unstructured data (Videos, Music, Files) globally: Cloud Storage (Object Storage)
  • Data is automatically managed and transitioned between storage classes to reduce costs: Object Lifecycle Management in Cloud Storage
  • A massive, one-time migration of data to the cloud, where online transfer is not feasible: Using Transfer Appliance for large- scale, physical data migration

6. Database Fundamentals

Cloud SQL & Cloud Spanner (OLTP)

  • Applications where large number of users make large number of small transactions

    • Popular databases: MySQL, Oracle, SQL Server etc

  • Recommended Google Managed Services

    • Cloud SQL: Supports PostgreSQL, MySQL, and SQL Server for regional relational databases
    • Cloud Spanner: Unlimited scale (multiple PBs) and 99.999% availability for global applications with horizontal scaling
    • Cloud Spanner replicates data across regions in real time

BigQuery (OLAP)

  • Applications allowing users to analyze petabytes of data

    • Reporting applications, Data ware houses, Business intelligence applications, Analytics systems

  • Recommended Google Managed Services

    • BigQuery: Petabyte-scale distributed data ware house

OLAP vs OLTP

  • OLAP and OLTP use similar data structures, but different in how data is stored
  • OLTP databases use row storage
    • Efficient for processing small transactions
  • OLAP databases use columnar storage
    • High compression, Distribute data, Execute single query across multiple nodes

Cloud Firestore vs BigTable (NOSQL)

  • NoSQL databases trade-off “Strong consistency and SQL features” to achieve “scalability and high-performance”
  • NoSQL = not only SQL
  • Google Managed Services: Cloud Firestore (Datastore) & BigTable

  • Cloud Datastore - Managed serverless NoSQL document database

    • Designed for transactional mobile and web applications

  • BigTable - Managed, scalable NoSQL wide column database

    • Recommended for large analytical (> 10 TB) and operational workloads (not serverless)

Memory Store (In-memory Databases)

  • Retrieving data from memory is much faster than retrieving data from disk (Redis)
  • Recommended GCP Managed Service: Memory Store
  • Use cases: Caching, session management, gaming leader boards, geospatial applications

Database in GCP

Relational OLTP databases: Cloud SQL, Cloud Spanner

  • Have predefined schema and very strong transcational capabilities (Row storage)

Relational OLAP databases: BigQuery

  • Columnar storage with predefined schema. Datawarehouse & BigData workloads

NoSQL Databases: Cloud Firestore (Datastore), BigTable

  • Apps that need quickly evolving structure (schema-less)

In memory databases/caches: Memory Store

  • Applications needing microsecond responses

7. IAM

Cloud IAM

  • How do you identify users in GCP?
    • Identity and Access Management (Cloud IAM) provides this service

IAM Example

  • Provide access to manage a specific cloud storage bucket to a colleague
    • Choose a Role with right permissions (Ex: Storage Object Admin)
    • Create Policy binding member (your friend) with role (permissions)
  • Roles: A set of permissions (to perform specific actions on specific resources)

IAM Roles

  • Roles are Permissions: Perform some set of actions on some set of resources
    • Basic Roles - Owner / Editor / Viewer
    • Predefined Roles - Fine grained roles predefined and managed by Google
    • Custom Roles - When predefined roles are NOT sufficient, you can create your own custom roles

IAM Policy

  • Roles are assigned to users through IAM Policy documents
  • Represented by a policy object


8. Encryption

Data Lifecycle States

  • Data at rest: Stored on a device or a backup
  • Data in motion: Being transferred across a network
  • Data in use: Active data processed in a non-persistent state

Encryption

  • Symmetric Key Encryption
  • Use the same key for encryption and decryption

  • Asymmetric Key Encryption
  • Encrypt data with Public Key and decrypt with Private Key

Cloud KMS

  • Cloud KMS: Create and manage cryptographic keys (symmetric and asymmetric)

9. Organizing GCP Resources

Resource Hierarchy

  • Organization > Folder > Project > Resources
  • Resources are created in projects
  • A Folder can contain multiple projects
  • Organization can contain multiple Folders

Billing Accounts

  • Billing Account is mandatory for creating resources in a project
  • Setup a Cloud Billing Budget to avoid surprises - Alerts

IAM Best Practices

  • Principle of Least Privilege: Give least possible privilege needed for a role
  • Separation of Duties: Involve atleast 2 people in sensitive tasks
  • Constant Monitoring: Review Cloud Audit Logs to audit changes to IAM policies and access to Service Account keys

Public, Private, Hybrid Cloud

  • Public Cloud: You host everything in the cloud
    • DO NOT need a data center，NO Capital Expenditure needed
    • Hardware resources are owned by Google Cloud
  • Private Cloud: You host everything in your own data center
    • Needs Capital Expenditure

  • Hybrid Cloud: Combination of both (Public & Private)
    • Use Public Cloud for some workloads and Private cloud for others
  • Multi Cloud: Using Multiple Cloud Platforms with/without on￾premise infrastructure
    • Reduces the likelihood of system failure during high-demand events by distributing the load and services across multiple cloud providers

Cloud VPN

  • Cloud VPN: Connect on-premise network to the GCP network
    • Encrypted
    • For low bandwidth, Cloud VPN is recommended
    • 注意, Cloud VPN 还是会经过 Public Internet, 如果不想经过 Public Internet, 选择Partner Interconnect connection

Cloud Interconnect

  • Cloud Interconnect: High speed physical connection between on-premise and VPC networks
    • Highly available with high throughput
    • Use only for high bandwidth needs

Private Google Access

  • Private Google Access allows instances in your VPC without external IP addresses to reach Google APIs and services using private IP addresses

Network Service Tier

  • Standard network service tier can help reduce data traffic costs, designed to be a cost-effective solution for services where performance and latency are less critical

  • Premium network service tier provides better performance, reliability, and lower latency by using Google’s private global network, is more expensive

Organization Policy Service

  • How to enable centralized constraints on all resources created in an Organization?
    • Configure Organization Policy
  • Needs a Role - Organization Policy Administrator
  • IAM focuses on WHO, Organization Policy focuses on WHAT

Corporate Directory Federation

  • Federate Cloud Identity or Google Workspace with your external identity provider (IdP) such as Active Directory or Azure Active Directory

Identity Aware Proxy (IAP)

  • Identity-Aware Proxy (IAP) in your Google Cloud VPC network is primarily used for controlling access to your applications running on Google Cloud.

  • It provides secure, identity-based access to applications and VMs without needing to configure a VPN or manage firewall rules.

Identity Platform

  • Identity Platform: Customer identity and access management, handle user authentication, authorization, and user management
  • Difference between Cloud IAM and Identity Platform
    • Cloud IAM: Employees and Partners Authorization
    • Identity Platform: Customer identity and access management

10. DevOps

CI, CD Tools

  • Cloud Source Repositories: Fully-featured, private Git repository
  • Container Registry: Store your Docker images
  • Cloud Build: Build deployable artifacts from your source code and configuration

Cloud Build

  • Cloud Build is a fully managed continuous integration and continuous delivery (CI/CD) platform on Google Cloud

  • It automates the process of building, testing, and deploying applications

Cloud Deployment Manager

  • Cloud Deployment Manager is an infrastructure management service that automates the deployment and management of Google Cloud resources using templates

  • Cloud Deployment Manager is not designed to automate the build and test processes

Container Registry and Artifact Registry

  • Container Registry: Uses GCS bucket to store images, supports Container images only
  • Artifact Registry: Evolution of Container Registry, manage BOTH container images and non-container artifacts

Infrastructure as code

  • Treat infrastructure the same way as application code
  • Bring repeatability into your infrastructure
  • GCP service: Google Cloud Deployment Manager

Cloud Operations & Insights

  • Monitoring - Metrics and Alerts: Cloud Monitoring
  • Centralized Logging: Cloud Logging
  • Audit Logging: Cloud Audit Logs
  • Real-time exception monitoring: Error Reporting
  • Live Debugging: Cloud Debugger
  • Distributed tracing: Cloud Trace
  • Statistical, low-overhead profiler: Cloud Profiler

Site Reliability Engineering (SRE)

  • SRE teams focus on every aspect of an application
    • Manage by Service Level Objectives (SLOs)
    • Minimize Toil
    • Move Fast by Reducing Cost of Failure
    • Share Ownership with Developers

SRE - Key Metrics

  • Service Level Indicator(SLI): Quantitative measure of an aspect of a service
    • Categories: availability, latency, throughput, durability, correctness (error rate)
  • Service Level Objective (SLO) - SLI + target
    • 99.99% Availability, 99.999999999% Durability
  • Service Level Agreement (SLA): SLO + consequences (contract)
    • What is the consequence of NOT meeting an SLO? (Defined in a contract)
  • Error budgets: (100% – SLO)
    • How well is a team meeting their reliability objectives?

SRE - Best Practices

  • Handling Excess Loads: Load Shedding, Reduced Quality of Service
  • Avoiding Cascading Failures: Plan to avoid thrashing

  • Penetration Testing (Ethical Hacking)
  • Load Testing (JMeter, LoadRunner, Locust, Gatling etc)

  • Resilience Testing - “How does an application behaves under stress?”

Single Sign-On (SSO)

  • With SSO in place, when an employee’s Active Directory account is terminated, they will no longer be able to authenticate and, consequently, will lose access to Google Workspace tools automatically

Google Cloud Directory Sync

  • Google Cloud Directory Sync (GCDS) is specifically designed to synchronize your existing LDAP directory with Google Cloud

11. Pub/Sub

  • Synchronous Communication: Applications makes synchronous calls to the logging service (What if the logging service goes down?)

  • Asynchronous Communication: Create a topic and have applications put log messages on the topic. Logging service picks them up for processing when ready

Pub/Sub

  • Pub/Sub: Reliable, scalable, fully-managed asynchronous messaging service
  • Backbone for Highly Available and Highly Scalable Solutions
  • Event ingestion and delivery for streaming analytics pipelines

  • Publisher - Sender of a message
  • Subscriber - Receiver of the message
    • Pull - Subscriber pulls messages when ready
    • Push - Messages are sent to subscribers

Cloud Dataflow

  • Cloud Dataflow is a difficult service to describe
    • Pub/Sub > Dataflow > BigQuery (Streaming)
    • Pub/Sub > Dataflow > Cloud Storage (Streaming - files)
    • Cloud Storage > Dataflow > Bigtable/CloudSpanner/Datastore/BigQuery (Batch - Load data into databases)

12. Data Architectures in GCP

Loose Coupling with Pub/Sub

  • Whenever you want to decouple a publisher from a subscriber, consider Pub/Sub

Date Formats

  • Structured: Tables, Rows and Columns (Relational)
    • Cloud SQL, Cloud Spanner, BigQuery
  • Semi Structured: Flexible Schema
    • Cloud Firestore/Datastore
  • Unstructured: Video, Audio, Image, Text, Binary files
    • Cloud Storage

Cloud Dataproc

  • Cloud Dataproc: Managed Spark and Hadoop service

Big Data Flow - Batch Ingest

  • Use extract, transform, and load (ETL) to load data into BigQuery
  • Dataprep: Clean and prepare data
  • Dataflow: Create data pipelines (and ETL)
  • Dataproc: Complex processing using Spark and Hadoop

Steaming Data

  • Pub/Sub: Receive messages
  • Dataflow: Analyze, aggregate andfilter data
  • For pre-defined time series analytics, storing data in Bigtable
  • For ad hoc complex analysis, prefer BigQuery

IOT

  • IoT Core: Manage IoT (registration, authentication, and authorization) devices
  • Pub/Sub: Durable message ingestion service (allows buffering)
  • Dataflow: Processing data (ETL & more..)
  • Data Storage and Analytics:
    • Make IOT data available to mobile or web apps => Datastore
    • Execute pre-defined time series queries => Bigtable
    • More complex or ad hoc analytics/analysis => BigQuery

Data Lake

  • Single platform with combination of solutions for data storage, data management and data analytics

  • Storage
    • Cloud Storage (low cost + durability + performance + flexible processing)
  • Data Ingestion
    • Streaming data - Cloud Pub/Sub + Cloud Dataflow
    • Batch - Transfer Service + Transfer Appliance + gsutil
  • Processing and analytics
    • Run in-place querying using SQL queries using BigQuery or (Hive on Dataproc)
  • Data Mining and Exploration
    • Clean and transform raw data with Dataprep
    • Use Cloud Datalab (data science libraries such as TensorFlow and NumPy) for exploring

Data Governance

  • Bad data: Bad data leads to poor business decisions
  • Data leaks: Data leaks can lead to a reputation loss

Dataplex

  • Dataplex is a Data Mesh: Unified dashboard with visibility into all data assets (data lakes, data warehouses, ..)

13. API Management in GCP

API Management

  • Apigee API Management: Comprehensive API management platform
  • Cloud Endpoints: Basic API Management for Google Cloud backends
  • API gateway: Newer, Simpler API Management for Google Cloud backends

14. Trust and Security with GCP

Cloud Security

  • Control: Decide who gets access
  • Compliance: Follows legal rules
  • Confidentiality: Keeps information secret
  • Integrity: Ensures data stays accurate
  • Availability: Ensure apps & data are available always

Enhanced Security with 2SV

  • 2 Step Verification (2SV): Add a 2nd step to verify user
  • Make 2SV Mandatory: For Google Cloud accounts

Security Command Center

  • ·Security Command Center· (SCC) in Google Cloud Platform (GCP) is designed to provide a comprehensive view of the security state of your GCP resources, including virtual machines

  • It enables you to detect and respond to threats, ensures compliance by reporting on security findings, and offers insights into the security status of your resources across multiple projects

Google Cloud Armor

  • Google Cloud Armor is a service designed to protect web applications from DDoS attacks and other web-based threats. It offers features like IP blacklisting/whitelisting, rate limiting, and application-layer traffic monitoring

Cloud Data Loss Prevention

  • Cloud Data Loss Prevention (DLP) is specifically designed to identify, classify, and help protect sensitive information, including PII

GCP Security Offerings

  • KMS: Create and manage cryptographic keys (symmetric and asymmetric). Control their use in your applications and GCP Services
  • Secret Manager: Manage your database passwords, your API keys securely
  • Cloud Data Loss Prevention: Discover, classify, & mask sensitive data
  • Cloud Armor: Protect your production apps (at run time) from denial of service and common web attacks

  • Web Security Scanner: Identify vulnerabilities by running security tests
  • Binary Authorization: Ensure that only trusted container images are deployed to Google Cloud
  • Container Threat Detection: Detects container runtime attacks
  • Security Command Center: Get a consolidated picture of security in Google Cloud

Zero Trust Security Model

  • Zero Trust - “No person or device should be trusted by default, even if they are already inside an organization’s network”

15. ML in GCP

ML in GCP - Pre-Trained

  • Speech-to-Text API: convert speech into text
  • Text-to-Speech API: convert text into speech
  • Translation API: Translate texts into more than one hundred languages
  • Natural Language API: Derive insights from unstructured text
  • Cloud Vision API: Recommended for generic usecases

ML in GCP - Custom Models

  • AutoML: Build custom models with minimum ML expertise and effort
    • AutoML Vision: Build custom models based on Images
    • AutoML Video Intelligence: Add labels to Video
    • AutoML Tables: Automatically build models on structured data
  • BigQuery ML: Build ML models using Queries
  • Vertex AI: Build & deploy ML models faster

  • AutoML Vision is ideal for scenarios where there is a need to create specialized image recognition models tailored to specific use cases

16. Cloud Native

Cloud Native Pillars

  • Microservices: Fix issues and deliver new features quickly
  • Containers: Portable & Lightweight
  • Container Orchestration: Kubernetes (GKE) - Auto Scaling, Load Balancing, Self Healing, Zero Downtime Deployment etc
  • DevOps (Dev + Ops, CI/CD, IaC): Increased automation of processes

Container Compute Examples

  • Cloud Run: Develop and deploy highly scalable containerized applications
  • Google Kubernetes Engine: Orchestrate containerized microservices on Kubernetes
  • Anthos: Manage Kubernetes Clusters in Multi-cloud and On-premises

Serverless Examples

  • Cloud Functions: Serverless compute for event-driven apps
  • Cloud Run: Run isolated containers, without orchestration (Serverless)
  • Cloud Firestore: Apps needing quickly evolving structure (schema-less)
  • Cloud Dataflow: Serverless Stream and Batch processing using Apache Beam
  • Cloud Pub/Sub: Realtime Messaging in the cloud. Pay for number of messages
  • BigQuery: Relational OLAP, Data warehousing & BigData workloads

17. Cost Management in GCP

CapEx vs OpEx

  • Capital Expenditure (CapEx): Money spent to buy infrastructure
  • Operational Expenditure (OpEx): Money spent to use a service or a product

Pricing Calculator

  • Pricing Calculator: Estimating the cost of a Google Cloud solution

GCP Cost Management

  • Cost Management: Tools for monitoring, controlling, and optimizing your costs

Billing dashboards

  • Billing dashboards in Google Cloud provide a comprehensive overview of cloud spending

Resource labels

  • Resource labels enable users to tag cloud resources with key-value pairs, which can then be used to filter and organize cost data

18. GCP review

Basic Compute Services

  • Compute Engine: Use VMs when you need control over OS OR you want to run custom software
  • Preemptible VMs: Short lived VMs for non time-critical workloads
  • Sole-tenant Nodes : Dedicated physical servers
  • VMware Engine: Run VMware workloads in Google Cloud
  • Managed Instance Groups: Create multiple Compute Engine VMs
  • Cloud Load Balancing: Balance load to multiple instances of an application or a service

Firebase

  • Firebase is a comprehensive cloud-based platform ideal for building web and mobile applications
  • Firebase is designed to handle dynamic content and user interactions seamlessly, providing an integrated backend-as-a-service (BaaS) solution that reduces the need for managing servers and infrastructure

Managed Compute Services

  • App Engine: PaaS. Deploy web apps and RESTful APIs quickly
  • Cloud Run: Run isolated containers, without orchestration (Serverless)
  • Kubernetes Engine: Managed Kubernetes Service. Provides container orchestration
  • Cloud Functions: Serverless compute for event-driven apps
  • Anthos: Manage Kubernetes Clusters in Multi-cloud and On-premises
  • Firebase: Google’s mobile platform. Build Apps for iOS, Android, the web, C++, and Unity

Storage

  • Persistent Disk: Block Storage for your VMs
  • Local SSD: Local ephemeral block storage for your VMs
  • Cloud Filestore: File shares in the cloud
  • Cloud Storage: Object storage in the cloud

Databases

  • Cloud SQL: Regional Relational OLTP database (MySQL, PostgreSQL, SQL server)
  • Cloud Spanner: Global Relational OLTP database. Unlimited scale and 99.999% availability for global applications with horizontal scaling
  • Cloud Firestore: Apps needing quickly evolving structure (schema-less)
  • Cloud BigTable: Large databases(10 TB - PBs). Streaming (IOT), analytical & operational workloads. NOT serverless
  • Cloud Memorystore: In memory databases/cache. Applications needing microsecond responses

Streams, Analytics, Big Data

  • Cloud Pub/Sub: Realtime Messaging in the cloud
  • BigQuery: Relational OLAP databases. Datawarehousing & BigData workloads
  • BigQuery ML: Simplified Machine Learning using data in BigQuery
  • Cloud Dataflow: Serverless Stream and Batch processing using Apache Beam (open-source)
  • Cloud Dataproc: Managed Service for Spark and Hadoop. Not serverless
  • Cloud Data Fusion: Visually manage your data pipelines
  • Data Studio: Visualize data
  • Looker: Enterprise Business Intelligence

Migration

  • Database Migration Service: Migrate to Cloud SQL
  • Storage Transfer Service: Online Transfer to Cloud Storage
  • Transfer Appliance: Physical transfer using an appliance
  • Migrate for Compute Engine: Migrate VMs and VM storage to GCE
  • Migrate for Anthos: Migrate VMs to GKE containers
  • BigQuery Data Transfer Service: Migrate your analytics data