Cloud Computing Definition
- on demand delivery of
- compute
- database
- storage
- appications
- other IT resources
- through a cloud services platform
- via the internet
- with pay-as-you-go pricing
6 Advantages of Cloud Computing (VECSMG)
- Variable - Trade Capital Expense For Variable Expense
- Economies - Benefit from Economies of Scale
- Capacity - Stop guessing Capacity
- Speed - Increase speed and agility
- Maintenance - Stop spending money running and maintaining data centers
- Global - Go Global in Minutes
CLOUD COMPUTING TYPES
IaaS
- Servers (EC2)
PaaS
- Managed Platform (GoDaddy, LightSail)
SaaS
- (GMail)
Deployment Types
- Public - AWS, Azure, GCP
- Hybrid - Mix
- Private - On Premise - Openstack, VMWare
Services Types on AWS for Practitioner Exam
- Compute
- Storage
- Databases
- Security Identity & Compliance
- AWS Cost Management
Compute
- EC2
- Lambda
Databases
- Relational Database Service (DBS)
- DynamoDB (Non Relational Database)
Storage
- S3
- Glacier
Network
- VPC
- Route53
Regions (19 + 5) - Geographical Area
Availability Zones (57 + 15) - Data Center
Edge Location (150) - Endpoint used for caching content. Consists of CloudFront, Content Delivery Network (CDN)
How to choose Region? (SSL)
- Data Sovereignty Laws
- Latency to end users
- AWS Services
Free?
- EC2 - 750 Hours
- QuickSight - 1GB
- RDS - 750 Hours
- S3 - 5GB
- Lambda - 1 Million Requests
- Always Free -
Support Plans
*Pricing scales based on usage
- Basic - Free - Support Forums, can ask questions
- Developer - USD 29 / month* - technical support - 12-24 hours response time
- Business - USD 100 / month* - 24x7 support by phone - 1 hour response time
- Enterprise - USD 15000 / month* - Technical Account Manager (TAM) - 15 minute response time
Windows Users
Go to EC2 > Go to Network & Security Tab > Go to Key Pairs > Create Key > Download PEM file > Download PuTTY 64 bit > Install > Open > Click on Load > Upload the PEM file downloaded > Save Private Key
Create Billing Alarm from Cloud Watch under Management & Governance
IAM
Definition
- It is Global, not region specific
- Root Account has full administrative rights
GROUPS
- Group stores users and inherit all permissions of the group
POLICIES
- Policies are applied to group, it sets the permissions. They are in JSON format
- Under Security, Identity & Compliance
S3
- Add flat files (files that don't change in the place)
- Videos, pictures
- NOT database files
- Object based Storage
- NOT Block based storage that is used in EC2
- File Size: 5 terra bytes (TB)
- Files are stored in buckets. Buckets are folders
- Buckets must be unique globally
- https://s3-<region>.amazonaws.com/<bucket name>
- Objects in S3 has key and value
- On uploading object in S3, HTTP 200 status code is generated
- Key is name
- Value is the data made up of sequence of bytes
- Version ID for versioning
- Metadata Data about data that is stored
- Subresources
- Access Control Lists
- A Torrent
- Data Consistency in S3
- Read after write consistency for PUTS of new objects
- New files are immediately available after uploading
- Although updating or deleting takes time
- Eventual Consistency for overwrite PUTS and DELETES (can take some time to propagate)
- 99.99% - built for
- 99.9% - Guarantee availability
- 99.99999999999% (11 nines in total, 9 nines after decimal) durability
- Features
- Tiered Storage
- Lifecycle Management
- Versioning
- Encryption
- Secure data with Access Control Lists (file level) and Bucket Policies (bucket level)
- TIERED STORAGE
- S3 Standard
- Availability 99.99%
- Durability 99.999999999%
- S3 - IA (Infrequently Accessed)
- Infrequently accessed data but rapid access when needed
- Lower storage fee than S3 but retrieval fee charged
- S3 One Zone - IA (Infrequently Accessed)
- Lower Cost Option
- Not frequently accessed data
- Do NOT require multiple availability zone data resilience
- S3 - Intelligent Tiering
- uses ML
- looks at usage pattern
- optimizes cost
- moves data to cost effective access tier
- S3 Glacier
- data archiving
- low cost
- secure, durable
- Retrieval time configurable from minutes to hours
- S3 Glacier Deep Archive
- Lowest storage cost on AWS
- Retrieval time greater than 12 hours
- S3 PRICING FACTORS (MR STAR)
- Storage
- Requests
- Storage Management Pricing
- Data Transfer Pricing
- Transfer Acceleration
- Cross Region Replication Pricing
- S3 Transfer Acceleration
- If this is enabled, users anywhere in the world can upload a file and it will not be sent to the data-center in any region via the internet. Instead, it is sent to a nearby edge location from where it is transfered to our data-center via Amazon's private network which is faster than the internet
- Cross Region Replication
- On activating it, buckets are replicated in another region
- S3 Policies
- Bucket Policies
- Applied to entire bucket
- Object Policies
- Applied to Object
- IAM Policies to Users & Groups
- Applies to Users & Groups
CloudFront
- Edge Location - content is cached
- Origin - S3 or EC2 or Elastic Load Balancer or Route53
- Distribution - Web and RTMP (for Adobe flash streaming)
- Comes under Networking and Content Delivery
- Content Delivery Network (CDN)
- Webpages and other web content are delivered to users from CDN where they are cached
- It stands between Servers and Users to quickly deliver cached version of the content
- CloudFront is at Edge Locations
- Web Distribution - for websites
- RTMP - for media such as Adobe Flash
- S3 Buckets can be set up for CloudFront Web Distribution
- On doing that, the links of the objects are ready to be cached on vatious edge locations if requested by users
EC2
- Definition
- Elastic Compute Cloud
- Virtual Servers
- Boots in minutes
- Quickly scale up and down
- Types
- On Demand
- Fixed rate
- Reserved
- Types
- Standard
- Convertible
- Scheduled - Between a given time everyday
- Longer contracts
- Discount on hourly charge
- Best Price - Pay for 3 years and get maximum discount
- Spot
- Bidding goes on all the time
- Flexible start and end times
- Good for applications where it is ok to lose the server
- If Amazon ends the session of an instance, you WON'T be charged for the hour in which it was stopped
- If you stop the session of an instance, in that case you WILL be charged for the hour in which it was stopped
- Dedicated Hosts
- Physical EC2 Servers dedicated for your use
- Reduces cost
- Fewer people use it
- License terms and conditions there
- Instance Types
- C5
- Compute Optimized
- CPU Intensive Apps/Databases
- D2
- Dense Storage
- Data Warehousing, Hadoop
- F1
- Financial Analytics, Genomics Research
- G3
- Graphic Intensive
- Video Encoding, 3D Application Streaming
- H1
- High Disk Throughput
- MapReduce workloads, distributed file systems such as HDFS & MarpR-FS
- I3
- High Speed
- NoSQL databases, Data Warehousing
- M5
- General Purpose
- application Servers
- P3
- Graphics/General Purpose GPUs
- Machine Learning, Bit coin mining
- R5
- Memory Optimized
- Memory Intensive Apps/Databases
- T3
- General Purpose
- Low Cost
- Webservers, small databases
- X1
- Memory Optimized
- SAP HANA, Apache Spark
- Z1D
- High Compute Capacity and high memory footprint
- Electronic Design Automation, Certain relational database workloads
- General Purpose Instances:
- T3, M5, P3
- EBS
- Virtual Hardisk
- Every server has a disk
- Storage attached to EC2
- Run file systems, databases, etc. on them
- Stored in same availability zone as EC2
- Automatically replicated
- SSD
- GP2 - General Purpose SSD
- General Purpose
- IO1 - Provisioned IOPS SSD
- for High performing database
- Magnetic
- ST1 Throughput Optimized HDD
- Low Cost
- Frequently accessed
- throughput-intensive workload
- SC1 - Cold HDD
- Lowest Cost
- Less frequently accessed workloads
- Magnetic
- Previous Generation
- Common Ports
- Linux
- SSH
- Port 22
- Microsoft
- Remote Desktop Protocol
- Port 3389
- HTTP
- Port 80
- HTTPS
- Port 443
- Security Group is a Virtual Firewall in the cloud
- To allow any IP addresss using 0.0.0.0/0
- To just let one IP address we use the IP address followed by /32 (X.X.X.X/32)
- Production should have each instance in separate availability zone so that if one goes down other instance can take the load
AWS Command Line
- PuTTY is used on Windows to access AWS from my laptop's terminal
- User takes Access Key Id and Secret Access Key
- Command to configure Terminal with AWS
- aws configure
- Add access key id
- Add secret access key
- Add region name
- Command to make bucket from AWS terminal
- aws s3 mb s3://<bucket-name>
- Command to go in created bucket
- aws s3 ls
- Create a file in EC2
- echo "hello world" > hello.txt
- ls
- Command to copy file from EC2 to S3 Bucket
- aws s3 cp hello.txt s3://<bucket-name>
- AWS has a secret folder where it stores access key id and secret key. Go to root directory
- cd ~
- cd .aws
- ls
- nano credentials
- The above commands help you to view the file where access key id and secret key is stored
- Try using wrong access key id and secret key
- aws configure
- enter fake values
- aws s3 ls
One can interact with AWS in 3 ways
- Web Console
- Command Line Interface (CLI)
- Software Development Kits (SDKs)
ROLES
Create Role
- Go to AWS Web Console
- Go to IAM
- Go to Roles Tab
- Select EC2
- Click on Next for Permissions
- Select AmazonS3FullAccess
- Go to Next
- Add Tags
- Go Next
- Add Name and Description to the Role
- Create Role
Connect Role to EC2 Instance
- Go to EC2
- Go to Instances
- Select instance
- Click on Actions
- Select Instance Settings
- Click on Attach/Replace IAM Role
- From Drop down select the role just created
- Click Apply
The below commands help to remove .aws folder where access key id and secret access key is stored
- Go to CLI
- Type
- ssh ec2-user@<IP address> -i MyPrivateKey.pem
- ls
- aws s3 ls
- cd ~
- rm -rf .aws
- aws s3 ls
Roles are much better way than using access key id and secret access key
Roles are much more secure
Easier to manage
Can be applied to EC2 anytime and takes effect immediately
Roles are Global. Region need not be specified
Web Server
- Go to SSH CLI
- sudo su
- yum update -y
Install Apache
- yum install httpd -y
Turn it on
- service httpd start
Let's go to root directory of the website. Any file here will be visible on the IP address
- cd /var/www/html
- ls
Create a web page
- nano index.html
- <html><body><h1>Hello World!</h1></body></html>
- Ctrl + X
- Save? Yes
- ls
View Website
- Go to browser
- Add the public IP address to the address bar and run
LOAD BALANCERS (LB)
- Under EC2
- Click on Create Load Balancer
- Types (CAN)
- Application LB - good for applications
- Network LB - when high performances and static IP addresses
- Classic LB - old and cheap
- Choose Application LB
- Add a Name
- Select all availability zones - one AZ goes down, others are there
- Click Next
- Configure Routing - add name
- Next
- Click on the instance and click on Add to registered
- Next
- Create
Create a new instance
- Make sure it has subnet of different availability zone than the previous instance
- Click on advanced details
- In the text box add the following
- #!/bin/bash
- yum update -y
- yum install httpd -y
- service httpd start
- chkconfig on
- cd /var/www/html
- echo "<html><body><h1>Web Server 2</h1></body></html>"
- The above lines run at root level when EC2 instance boots up
- Next
- Next
- Add tag
- Next
- Select an existing Security Group
- Launch
Activate Load Balancer for two or more instances
- Go to Load Balancing > Target Groups
- Click on Targets tab below
- Add other instance
- Save
- Go to Load Balancers
- Get DNS name and run repeatedly in browser
- You will get both the websites alternatively
DATABASES
RDS - Relational Databases on AWS is called RDS
RDS allows (PA MOMS)
- Microsoft SQL Server
- Oracle
- MySQL
- PostgreSQL
- Amazon Aurora
- MariaDB
2 Key Features of RDS
- Multiple Availability Zones
- for disaster recovery
- lose one and other is used
- Read Replicas
- for performance
- production read copy
- can have upto 5 copies
Non Relational Databases
DynamoDB
- Each document have different number of fields
- Amazon's NoSQL Database is DynamoDB
OLTP vs OLAP
Online Transaction Processing vs
Online Analytics Processing
- OLTP
- A row of data from database - CRUD
- OLAP
Red Shift
- A group of data from database based on query - CRUD
- Data Warehousing was introduced for this
- Data Warehousing is used for Business Intelligence (BI)
- Amazon Data warehousing solution is called Redshift
ElastiCache
Definition
- web service
- makes easy to
- deploy
- operate
- scale
- in-memory caching solution in the cloud
- It improves performance of web application
- by allowing apps to retrieve info from
- fast
- managed
- in-memory caches
- ElastiCache caches most common queries
- Un common queries are made to database
Caching Engines (Open source and in-memory) that Elasticache supports
- Memcached
- Redis
AMI
- Amazon Machine Image
- Created from EC2 Instance
- To recreate the same instance again
Auto Scaling
Load Balancing
Route53
- Amazon's DNS - Domain Name System
- Named after Route 66 - interstate highway across United States
- DNS works on Port 53
- It is Global, not region specific, just like IAM and S3
- Under Networking in AWS services
- Provides
- DNS Management
- Traffic Management
- Availability Monitoring
- Domain Registration
- Used to
- direct traffic around the world
- register domain name
Elastic Beanstalk (Free)
Definition
- quickly deploy and manage applications
- in AWS cloud
- without worry of the infrastructure
- Upload application and
- Elastic Beanstalk automatically handles
- capacity provisioning
- load balancing
- scaling
- application health monitoring
How?
- On a click of button provision EC2, security groups, application load balancers
- Only code, rest taken care
- Choose stack, upload code and there you go
Name all Compute Services
- EC2
- LightSail
- Elastic Beanstalk
- Lambda
- etc.
Cloudformation (Free)
Definition
- Service
- that helps to model
- and set up AWS resources
- allowing one to spend less time managing resources
- Cloudformation takes care of configuration and provisioning
- It is free, but resources it deploys may not be free such as EC2, RDS instances
- Most powerful AWS tool
- Turns infrastructure into code
- With this code deploy within minutes
- Various templates available
- It creates all instances that are required
- EC2
- Security Group
- RDS
- etc.
- One can design one's own template too
- Configures them
Difference between Elastic Beanstalk and CloudFormation
- Elastic Beanstalk is limited to what it can provision, CloudFormation can provision almost any AWS service
- Elastic Beanstalk is not programmable while CloudFormation is completely programmable
Architecting for the cloud (White paper)
- IT Assets as Provisioned Resources
- Global, Available, and Scaleable Capacity (GAS Capacity)
- Higher Level Managed Services
- Built-in-Security
- Multi level authentication, etc
- Architecting for Cost
- Cost efficient
- Operations on AWS
- Re-architecting, re-factoring
Scalability
- Scale Up
- Example, t2 micro instance to bigger instance
- Scale Out
- Example, adding multiple virtual machines behind an elastic load balancer
- Types: 6 types of scaling out
- Stateless Applications
- Example, Alexa is Lambda, stateless
- Distributed Load to Multiple Nodes
- Example, Multiple Web servers or database servers
- Stateless Components
- Example, uses cookie
- Statefull Components
- Example, transactions stored in RDS
- Implement Session Affinity
- Sticky sessions
- Cookie in user's browser
- Everytime they and visit your website, Application Load Balancer detects the cookie and sends user back to the same EC2 instance
- Implement Distributed Processing
- Elastic Map Reduce
- Have bunch of EC2
- and process large number of jobs
- reduce time to complete it
Disposable Resources instead of fixed servers
- Instantiating Compute Resources
- Bootstrapping
- Not manually instantiating
- Golden Images
- Taking image of instance and use it again and again
- Containers
- Hybrid
- Containers + EC2 instance
- Infrastructure as a code
- CloudFormation
- Automation
- Serverless Management & Deployment
- AWS Elastic Beanstalk
- EC2 autorecovery
- AWS Systems Manager
- Auto Scaling
- Alarms & Events
- Amazon CloudWatch alarms
- Amazon CloudWatch events
- Example, Someone uploads image to S3, lambda adds watermark to the image or creates a thumbnail, etc.
- AWS Lambda Scheduled Events
- Example, Run a code at midnight
- AWS Web Application File (WAF) security automations
- automatically respond to people trying to do harm
- example SQL injections, Cross site scripting
- Loose Coupling
- Well Defined Interfaces
- Amazon API Gateway
- Service Discovery
- Example, EC2 instance connecting to RDS instance using DNS name, if multi-AZ is on, then if an RDS instance fails, AWS will switch over to other AZ
- Asynchronous Integration
- Distributed Systems Best Practices
- Graceful failure in prcactice
- Example, if a page does not exist, user sees error page
Services and NOT Servers
- Managed Services
- Serverless Architectures
Databases
- Relational Databases (Aurora)
- Scalability
- High Availability - Multi AZ
- Anti Patterns - if no need of joins or complex transactions, use NoSQL
- Non-Relational Databases (DynamoDB)
- Scalability
- High Availability - Multi-AZ
- Anti Patterns - if you have joins or complex transactions, use SQL for them OR large binary files like audio, video, images, etc., use S3 for them
- Data Warehouses (RedShift)
- Scalability
- High Availability - Multi-AZ
- Anti-Patterns - not meant for Online Transaction Processing (OLTP), more used for Online Analytics Processing (OLAP)
- Search (Cloud Search, Amazon Elastic Search)
- Scalability
- High Availability - Multi-AZ
- Graph Databases (Neptune)
- Scalability
- High Availability - Multi-AZ
Managing Increasing Volume of Data
- A data lake
- architectural approach
- allows to store massive amounts of data
- in a central location
- so that readily available
- to be categorized
- processed
- analysed
- consumed by diverse groups in organization
- Data stored as-is
- no need to convert it to pre-defined schema
- no need to know what questions to ask about your data beforehand
- Great place to create data lake is S3
- Athena can be used to run SQL queries on this database
Removing Single Point of Failure
- Introducing Redunduncy
- Detect Failure
- Durable Data Storage
- Automate Multi-Datacenter Resilience
- Fault Isolation and Traditional Horizontal Scaling (Scaling Out)
- Sharding (Elastic MapReduce)
Optimize for Cost
- Right Sizing
- Elasticity
- events when higher usage is expected, needs to be provisioned
- Take advantage of variety of purchasing options
- Spot instances, on demand instances, reserved instances
Caching
- Application Caching
- Elastic Cache
- Edge Caching
- CDN Networks like CloudFront
Security
- Use AWS features for Defense in Depth
- Share Security Responsibility with AWS
- Reduce Privileged Access
- Security as Code
- EC2 instances with security patches on them
- CloudFormation Template
- Real-Time Auditing
- AWS Inspector, AWS CloudTrail, etc.
GLOBAL AWS SERVICES (IRCSS)
- IAM
- Route53
- CloudFront (CDN)
- SNS
- SES
GLOBAL VIEW BUT REGIONAL
- S3
AWS SERVICES ON PREMISE
- Snowball
- Gigantic Disc to upload data on and ship back to AWS (80 TB min)
- Snowball Edge
- CPU + Storage
- Example, Lambda on Premise
- Boeing uses it
- Storage Gateway
- Caching file on Premise
- It replicates to S3
- CodeDeploy
- Used for deploy to on-premise web service too
- OpsWorks
- Uses chef for automated deployment to EC2 instance in AWS cloud as well as on premise web server
- IoT Greengrass
- it is IoT but connects devices up to AWS cloud. has on premise service as well
AWS SERVICES CAN BE USED TO DEPLOY APPLICATIONS ON PREMISE
- CodeDeploy
- Used for deploy to on-premise web service too
- OpsWorks
- Uses chef for automated deployment to EC2 instance in AWS cloud as well as on premise web server
CLOUDWATCH 101
Definition
- Monitoring Performance Service
- to monitor AWS resources
- as well as the applications running on AWS
- Monitors once every 5 mins
- intervals can be reduced to 1 min by turning on detailed monitoring
- CloudWatch alarms trigger notifications
CloudWatch monitors
- Compute
- EC2 instances
- Autoscaling Groups
- Elastic Load Balancers
- Route53 Health Checks
- Storage and Content Delivery
- EBS Volumes (Virtual Hard disks attached to EC2)
- Storage Gateways
- CloudFront
- Host Level Metrics of :
- CPU
- Network
- Disk
- Status Check
AWS SYSTEMS MANAGER
- Allows to manage EC2 instances/virtual machines at scale
- Run a command on all EC2 instances at once
- Under Management and Governance in AWS Console
- It also integrates with CloudWatch to give dashboard for entire estate
PRICING
Capex - Capital Expenditure - Pay upfront
Opex - Operational Expenditure - Pay as you use
Opex - Operational Expenditure - Pay as you use
5 Policies (GRAMC)
- Pay as you go
- Pay less when you reserve
- Pay even less per unit by using more
- Pay even less as AWS grows
- Custom Pricing
4 Principles & Best Practices (FOFR)
- Understand the fundamentals of Pricing
- 3 Fundamental Drivers of Cost
- Compute
- Storage
- Data Outbound
- Start early with cost optimization
- Earlier you start with cost optimization, the better it is
- Helps grow right than have a complex infrastructure that cannot be optimized easily
- Maximize the power of flexibility
- AWS services are priced
- independently
- transparently
- No minimum commitment
- Use when required, stop when not
- This helps save upto 70% than running 24/7
- Use the right pricing model for the job
- On demand
- Reserved
- Spot
- Dedicated
3 Fundamental Drivers of Cost
- Compute
- Storage
- Data Outbound
7 AWS FREE SERVICES (VECOCAI)
- VPC (Virtual Data Center)
- Elastic Beanstalk
- But NOT services it deploys
- CloudFormation
- But NOT services it deploys
- IAM (users, groups, roles)
- Auto Scaling
- But NOT services it deploys
- OpsWorks
- But NOT services it deploys
- Consolidated Billing
What determines Price?
EC2 ( MIS PILOT H )
- Hours
- By second or hour
- Type of Instance
- T1 micro, etc
- Pricing Model
- On Demand
- Reserved
- Spot
- Dedicated
- Instances Count
- Load Balancing
- Application LB - less expensive
- Network LB - more expensive
- Monitoring
- Standard - 5 min interval
- Detailed - 1 min interval (more expensive)
- Scaling Automatically
- More EC2 more price to pay
- Elastic IP Addresses
- Is turned on by default everytime EC2 instance in instantiated, it is chargeable
- Operating Systems and Software Packages
- Windows more expensive than Linux

Lambda Pricing Determinants
- Free upto 1 million requests
- USD 0.20 / million requests thereafter
- 400,000 GB-seconds per month free, upto 3.2million seconds of compute time
- USD 0.00001667 for every GB-second thereafter
- Other AWS services used by Lambda
EBS Pricing Determinants
- Volumes (per GB)
- Snapshots (per GB)
- Data Transfer
S3 Pricing Determinants
- Class of Storage (SIIIGG)
- Standard
- IA
- IA - 1 AZ
- Intelligent Tiering
- Glacier
- Glacier Deep Archive
- Storage (GBs)
- Requests - Get, Put, Copy
- Data Transfer
Glacier Pricing Determinants
- Storage (GBs)
- Max Data Retrieval Time Period
- Expedited
- 1-5 mins
- 0.03 USD per GB
- 0.01 USD per request
- 100 USD per capacity unit provisioned
- Standard
- 3-5 hours
- 0.01 USD per GB
- 0.05 per 1000 requests
- Bulk
- 5-12 hours
- 0.0025 per GB
- 0.025 per 1000 requests
Snowball Pricing Determinants
- Service Fee per job
- 50 TB - 200 USD
- 80 TB - 250 USD
- Daily Charge
- Free - 10 days
- thereafter - USD 15/day
- Data Transfer
- Inbound - Free
- Outbound - NOT
RDS Pricing Determinants
- Clock Hours of Server Time
- DB Characteristics
- DB Purchase Type - size
- Number of Instances
- Provisioned Storage
- Additional Storage used
- Requests
- Deployment Type
- Data Transfer
DynamoDB Pricing Determinants
- Provisioned Throughput (write)
- Provisioned throughput (read)
- Indexed data storage
CloudFront Pricing Determinants
- Traffic Distribution
- Requests
- Data Transfer Out
AWS Budget vs AWS Cost Explorer
- Budget is used to budget/predict before they are incurred
- Cost Explorer is used to analyse expense after they have been incurred
AWS Support Plans
- Basic
- Free
- No Tech Support, only forums
- Nobody can open cases
- Developer
- 29 USD / month
- Business hour - email
- 1 person - unlimited cases
- General Guidance: < 24 biz hrs | System Impaired: < 12 biz hrs
- Business
- 100 USD / month
- 24x7 email, chat, phone
- Unlimited Contacts - Unlimited cases
- General Guidance: < 24 biz hrs | System Impaired: < 12 biz hrs | Prod System: < 4 hrs | Prod System Down: < 1 hr
- Enterprise
- 15,000 USD / month
- 24x7 email, chat, phone
- Unlimited Contacts - Unlimited cases
- Dedicated Technical Account Manager (TAM)
- General Guidance: < 24 biz hrs | System Impaired: < 12 biz hrs | Prod System: < 4 hrs | Prod System Down: < 1 hr | Buziness Critical System: < 15 mins
TAGS
- Tags are key value pairs
- CloudFormation Tags are inherited
- Groups resources using tags
- Discover resources using tags
- Information put in tags
- EC2 - Public or private IP Address
- ELB - Port Config
- RDS - Database type
- Resource Groups is where on can play with tags
- Tag Editor helps to find resources by tag
- New tags can also be added using tag editor
- Groups can be created using tags
- Automation can be applied on entire group
- Example, stop all EC2 instances in Mumbai
ORGANIZATIONS & CONSOLIDATED BILLING
- Account management service enables to consolidate multiple AWS accounts into an organization and centrally manage
- Features
- Consolidated Billing
- All Features
- More AWS accounts used with Organization, more pricing benefits
- Volume Discounts
- Unused reserved instances of EC2 are applied on across the group
- Root is Paying account
- Others are Linked accounts
- Max 20 linked accounts - soft limit
Best Practices
- Multi-factor authentication on root account
- Strong and complex password
- Paying account only for billing purpose, do not deploy resources
- Linked Accounts - max 20 (soft limit)
- Billing Alerts for consolidated as well as account based
- CloudTrail for auditing all accounts from root
CloudTrail
- CloudWatch monitors performance - RAM utilization, EC2 utilization etc.
- While CloudTrail monitors API calls within AWS Platform
- It audits
- Any changes, CloudTrail tracks it
- Per account
- Per region
- Can consolidate logs in S3 bucket of paying account where it has only write rights, no read rights
AWS QUICK STARTS
- List of Templates of CloudFormation
- Deploys solutions on one click
- Templates designed by experts
- FREE, but resources it provisions could be paid
AWS LANDING ZONE
- Allows to setup multi-account setups
- Organization level solution templates
- Based on Best Practices
AWS COST CALCULATORS
AWS SIMPLE MONTHLY CALCULATOR
- Estimations
- https://calculator.s3.amazonaws.com/index.html
AWS TOTAL COST OF OWNERSHIP (TCO) CALCULATOR
- Estimates and compares to on premise cost of the same operation
SECURITY
AWS Artifact
AWS is compliant to
Global:
- SOC 1
- SOC 2
- SOC 3
- PCI DSS Level 1
- Payment Card standard
- CSA - CLoud Security Alliance
- ISO 9001
- ISO 27001
- ISO 27017
- ISO 27018
US:
- FISMA - Federal Information Security Management
- HIPAA - Protected Health Information
AWS Shared Responsibility Model
- AWS is responsible for their Regions, AZs, Edge Locations | Compute, Storage, Database, Networking
- Security in the cloud is responsibility of the customer - Configurations, IAM, Customer Data
- Amazon responsible upto guest operating system | customers responsible there on
- Amazon responsibility is security OF the cloud, customer's responsibility is security IN the cloud
- Encryption setting on is customer responsibility, encryption is Amazon's
- If key is procided by customer then even encryption is customer's responsibility
- RDS patching is Amazon's responsibility
WAF & Shield
- WAF stands for Web Application Firewall
- Level 7 - Level 1 is hardware, Level 7 is application
- WAF in front of Load Balancer,s tops attackers from SQL injections and cross site scripting
- Shield is managed Distributed Denial of Service (DDoS) protection service
- safeguards web applications
- shields from DDoS attacks - such as too much traffic at once directed to the application with an intent to crash it or incur you high billing
- Types:
- Standard
- Advanced
- DDoS attack charges are reimbursed by AWS
AWS Inspector vs AWS Trusted Advisor vs CloudTrail vs CloudWatch vs Config (I-T-C)
Inspector (ASSESS)
- Inspector is automated security assessment service for vulnerabilities
- reports generated
- It is installed on EC2 instances
Trusted Advisor (ADVICE)
- Online resources
- helps to reduce cost, increase performance, and improve security
- real-time guidance on best practices
- Cost optimization
- Performance
- Security
- Fault Tollerance
- On entire AWS account
- Core checks on recommendations - FREE
- Full Trusted Advisor - Busine and Enterprise only
- Global Service
CloudTrail (LOG)
- Records everything in AWS environment
- Actions and API callls
- Logs everything
- Saved in S3
CloudWatch (PERFORMANCE)
- Monitor AWS Resource Performance
- CPU
- Network
- Disk
- Status Check
- Custom Scripts
AWS Config (CONFIG)
- Configs and relations and settings
- Changes made to config
Athena vs Macie
Athena
- Interactive Query Service
- Query data located in S3 using SQL
- Serverless
- Pay per query/TB scanned
- NO need to set up ETL processes (Extract/Transform/Load)
- Can be used to Mining Data Lake, Querying Logs, Generate Business Reports, query click-stream data
PII - Personally Identifiable Information
- Personal Data - Credit Card, Bank Account Details, etc.
Macie
- Uses AI to identify objects containing sensitive data such as PII
- Great for PCI-DSS compliance and preventing ID theft
Comments
Post a Comment