Note_Tech

All technological notes.

Project maintained by simonangel-fong Hosted on GitHub Pages — Theme by mattgraham

AWS - EC2

Back

AWS - EC2

EC2

Elastic Compute Cloud (EC2)
- a highly configuration virtual server
- resizable compute capacity.
- takes minutes to launch new instances.
- Everything on AWS uses EC2 instance underneath.

ec2

EC2 Instance Families

Instance Families
- different combinations of CPU, Memory, Storage, and Networking capacity.
- allow to choose the appropriate combination of capacity to meet user’s application’s unique requirements.

family

Instance Types

Instance Types
- a combination of size and family.
EC2 Instance Sizes generally double in price and key attributes
can use different types of EC2 instances that are optimised for different use cases (https://aws.amazon.com/ec2/instance-types/)
AWS has the following naming convention: m5.2xlarge
m: instance class
5: generation (AWS improves them over time)
2xlarge: size within the instance class

General Purpose

Great for a diversity of workloads such as web servers or code repositories
- Balance between:
- Compute
- Memory
- Networking
t2.micro: a General Purpose EC2 instance

Compute Optimized

Great for compute-intensive tasks that require high performance processors:
- Batch processing workloads
- Media transcoding
- High performance web servers
- High performance computing (HPC)
- Scientific modeling & machine learning
- Dedicated gaming servers

Memory Optimized

Fast performance for workloads that process large data sets in memory
Use cases:
- High performance, relational/non-relational databases
- Distributed web scale cache stores
- In-memory databases optimized for BI (business intelligence)
- Applications performing real-time processing of big unstructured data

Storage Optimized

Great for storage-intensive tasks that require high, sequential read and write access to large data sets on local storage
Use cases:
- High frequency online transaction processing (OLTP) systems
- Relational & NoSQL databases
- Cache for in-memory databases (for example, Redis)
- Data warehousing applications
- Distributed file systems

EC2 Tenancy

3 levels of tenancy
- Dedicated Host
  - User has whole server and control of the physical attribute.
- Dedicated Instance
- Default

tenancy

Pricing Models / Purchase Options

5 ways to pay for EC2

pricing

On-Demand

On-Demand
- Pay-As-You-Go (PAYG), where user consumes compute and then pay.
- Pay for what you use:
  - Linux or Windows - billing per second, after the first minute
  - All other operating systems - billing per hour
- Has the highest cost but no upfront payment
- No long-term commitment
- no up-front payment
- Recommended:
  - for short-term and un-interrupted workloads, where you can’t predict how the application will behave
  - For users whos have a new app for deveplopment or want to run experiment.
It is by default using On-Demand pricing when an EC2 is launched.
EC2s are charged by per-second (minumum of 60 seconds) or per-hour. When looking up pricing it will always show EC2 pricing is the hourly rate.

Reserved Instances (RI)

Reserved Instances (RI)
- is designed for application that have a steady-state, predictable usage, or require reserved capacity.
- Up to 72% discount compared to On-demand
- reserve a specific instance attributes (Instance Type, Region, Tenancy, OS)
- Reservation Period (Reserved instances do not renew automatically.)
  - 1 year (+discount)
  - 3 years (+++discount)
- Payment Options
  - No Upfront (+)
  - Partial Upfront (++)
  - All Upfront (+++)
- Reserved Instance’s Scope
  - Regional
  - Zonal (reserve capacity in an AZ)
- Recommended:
  - for steady-state usage applications (like database)
can buy and sell in the Reserved Instance Marketplace
Convertible Reserved Instance
- Can change the EC2 instance type, instance family, OS, scope and tenancy
- Up to 66% discount
Reduced pricing is based on Term, Class Offering, RI Attributes, and Payment Option.
- The instance will use On-Demand with no interruption to service when expiring.
Class: The less flexible the greater the savings.
- Standard
  - Up to 75% reduced pricing compared to on-demand.
  - User can modify RI Attributes.
- Convertible
  - Up to 54% reduced pricing compared to on-demand.
  - User can exchange RI based on RI Attributes if greater or equal in value.
Payment Option: The greater upfront the greate the savings
- All Upfront
  - Full payment is made at the start of the term
- Partial Upfront
  - A portion of the cost must be upfront and the remaining hours in the term are billed at a discounted hourly rate.
- No Upfront
  - Users are billed a discounted hourly rate for every hour within the term, regardless of whether the Reserved Instance is being used.
RI can be shared between multiple accounts within an AWS Organization.
Unused RIs can be sold in the Reserved Instance Maketplace.

RI Attributes / Instance Attributes
- are limited based on Class Offering and can affect the final price of an RI Instance.
There are 4 RI Attributes
- Instance Type:
  - is composed of the instance family and the instance size.eg. m4.large
- Region
  - The region in which the reserved instance is purchased.
- Tenancy
  - Whether instance runs on shared(default) or single-tenant(dedicated) hardware.
- Platform
  - The Operating System. eg.windows or Linux/Unix.

Regional and Zonal RI
When purchasing a RI, user determines the scope of the Reserved Instance.
- The scope does not affect the price.

Regional RI:purchase for a Region	Zonal RI: purchase for an AZ
does not reserve capacity	reserves capacity in the specified AZ
RI discount applies to instance usage in any AZ in the Region.	RI discount applies to instance in the selected AZ(No AZ Flexibility)
RI discount applies to instance usage within the instance family, regardless of size. Only supported on Amazon Linux Reserved Instances with default tenacy	No instance size flexibility. RI discount applies to instance usage for the specified instance type and size only
User can queue purchases for regional RI	User cannot queue purchases for zonal RI

RI Limit
- a limit to the number of Reserved Instances that user can purchase per month.
Per month user can purchase
- 20 Regional Reserved Instances per Region
- 20 Zonal Reserved Instances per AZ
Regional Limits
- The default On-Demand Instance limit is 20.
- User cannot exceed user’s running On-Demand Instance limit by purchasing regional Reserved Instances.
- Before purchasing RI user must ensure On-Demand limit is equal to or greater than the RI that user intends to purchase.
- Regional Limits <= On-Demand = 20
Zonal Limits
- User can exeed the running On-Demand Instance limit by purchasing zonal Reserved Instances.
- If user already have 20 running On-Demand Instances, and purchase 20 zonal RIs, user can launch a furhter 20 On-Demand Instance that match the specifications of the zonal RI.

Standard RI vs Convertible RI

difference

RI Marketplace
- EC2 Reserved Instance Marketplace allows user to sell unused Standard RI to recoup RI spend for RI user do not intend or cannot use.

Marketplace

Capacity Reservation

Capacity Reservation
- a service of EC2 that allows user to request a reserve of EC2 type for a specific Region and AZ.
- Reserve On-Demand instances capacity in a specific AZ for any duration
- You always have access to EC2 capacity when you need it
- No time commitment (create/cancel anytime),
- no billing discounts 独立无优惠
- Combine with Regional Reserved Instances and Savings Plans to benefit from billing discounts 联合有优惠
- You’re charged at On-Demand rate whether you run instances or not
- Suitable
  - for short-term, uninterrupted workloads that needs to be in a specific AZ
- EC2 instances are backed by different kind of hardware, and so there is a finite amount of servers available within an AZ per instance type or family.
  - So AWS may have ran out of a specific type of server when user launch an EC2.
The reserved capacity is charged at the selected instance type’s On-Demand rate wether an instance is running in it or not.
User can also use regional RI with Capacity Reservations to benefit from billing discounts.

Spot Instances

Spot instance provide a discount of 90% compared to On-Demand Pricing.
- Instances that you can “lose” at any point of time if your max price is less than the current spot price
- The MOST cost-efficient instances in AWS
- Useful for workloads that are resilient to failure
  - Batch jobs
  - Data analysis
  - Image processing
  - Any distributed workloads
  - Workloads with a flexible start and end time
- Not suitable for critical jobs or databases
- can be terminated if the computing capacity is needed by other On-Demand customers.
- AWS has unused compute capacity that they want to maximize the utility of their idle servers.
- For application that have flexible start and end times or applications that are only feasible at very low compute costs.
AWS Batch
- a aws service using Spot Pricing.
Termiantion Conditions
- Instance can be terminated by AWS at anytime.
- If instance is terminated by AWS, user does not get charged for a partial hour of usage.
- If user termiantes an instance, user will still be charged for any hour that it ran.

Dedicated Instances

Dedicated Instances
- are designed to meet regulatory requirements.
- For users who have strict server-bound licening that won’t support multi-tenancy or cloud deployments.
Instances run on hardware that’s dedicated to you
May share hardware with other instances in same account
No control over instance placement (can move hardware after Stop / Start)
Enterprises and Large Organizations may have security concerns or obligations about against sharing the same hardware with other AWS Cutomers.
Multi-Tenant
- When multiple customers are running workloads on the same hardware.
- Virtuall Isolation is what separate customers.
Single Tenant
- When a single customer has dedicated hardware.
- Physical Isolation is what separates customers.
Dedicated can be offer for :
- On-demand
- Reserved (up to 60% savings)
- Spot (up to 90% savings)

Dedicated Instances

Dedicated Hosts

Dedicated Hosts
- A physical server with EC2 instance capacity fully dedicated to your use
- single-tenant EC2 instances designed to let user Bring-Your-Own-License (BYOL) based on machine characteristic.
Allows you address compliance requirements and use your existing server-bound software licenses (per-socket, per-core, pe—VM software licenses)
Purchasing Options:
- On-demand – pay per second for active Dedicated Host
- Reserved - 1 or 3 years (No Upfront, Partial Upfront, All Upfront)
The most expensive option
Recommanded:
- Useful for software that have complicated licensing model (BYOL – Bring Your Own License)
- Or for companies that have strong regulatory or compliance needs

Dedicted Host

Saving Plan

Saving Plans
- offer users the similar disounts as RI but simplify the purchasing process.
- Commit to a certain type of usage ($10/hour for 1 or 3 years)
- Usage beyond EC2 Savings Plans is billed at the On-Demand price
- Locked to a specific instance family & AWS region (e.g., M5 in us-east-1)
- Flexible across:
  - Instance Size (e.g., m5.xlarge, m5.2xlarge)
  - OS (e.g., Linux, Windows)
  - Tenancy (Host, Dedicated, Default)
3 types of saving plans
- EC2 Instance:
  - provide the lowest prices, offering savings up to 72% in exchange for commitment to usage of individual instance families in a region.
  - automatically reduces cost on the selected family in that region regardless of AZ, size, OS, or tenancy.
  - offer the flexibility to change usage between instances within a family in that region.
- SageMaker
  - Help reduce SageMaker costs by up to 64%.
  - automatically apply to SageMaker usage regardless of instance family, size, component, or AWS region.
- Compute:
  - provide the most flexibility and help reduce costs by up to 66%. - automatically apply to EC2 instance usage, Fargate, and Lambda service usage regardless of instance family, size, AZ, region, OS, or tenancy.
Terms:
- 1 Year
- 3 Year
Payment Options
- All Upfront
- Partial Upfront
- No Upfront

Payment

Summary

Plan	Suiable
On demand	coming and staying in resort whenever we like, we pay the full price
Reserved	like planning ahead and if we plan to stay for a long time, we may get a good discount.
Savings Plans	pay a certain amount per hour for certain period and stay in any room type (e.g., King, Suite, Sea View, …)
Spot instances	the hotel allows people to bid for the empty rooms and the highest bidder keeps the rooms. You can get kicked out at any time
Dedicated Hosts	We book an entire building of the resort
Capacity Reservations	you book a room for a period with full price even you don’t stay in it

Shared Responsibility Model for IAM

AWS
- Infrastructure (global network security)
- Isolation on physical hosts
- Replacing faulty hardware
- Compliance validation
User
- Security Groups rules
- Operating-system patches and updates
- Software and utilities installed on the EC2 instance
- IAM Roles assigned to EC2 & IAM user access management
- Data security on your instance

Summary

EC2 Instance:
- AMI (OS) + Instance Size (CPU + RAM) + Storage + security groups + EC2 User Data
Security Groups: Firewall attached to the EC2 instance
EC2 User Data: Script launched at the first start of an instance
SSH: start a terminal into our EC2 Instances (port 22)
EC2 Instance Role: link to IAM roles
Purchasing Options:
- On-Demand,
- Spot,
- Reserved (Standard + Convertible + Scheduled),
- Dedicated Host,
- Dedicated Instance
- Saving Plan
3 Ways to access to Instance
- SSH, port 22
- EC2 Instance Connect, require SSH port
- Systems Manager (SSM) Session Manager, not require SSH port

TOP