EC2

• Elastic Compute Cloud
• Root device data can be an instance or on EBS (elastic block storage). By using Amazon EBS, data on the root device will persist independent of the lifetime of the instance. This enables us to stop and restart the instance without impacting the data.
• Different instance type
• Good practice – assign role to instance rather than using access key
• Good practice – assign security groups to control access
• Cost perspective – Should know when to use which to save cost
  • On-demand
  • Reserved instance (RI) – for a 1 or 3 year term, significant discount
  • Spot – bid price, loose the instance with 2 minutes warning if price goes up
  • Convertible (RI) – Allow to convert from one instance type to another
• Termination protection check box
• Encrypt root volume check box
• Status of instance
  • start – start and run assign public IP
  • stop – stop instance, remove public IP, Data is persisted in EBS root and other volumes. RAM is cleared.
  • hibernate – Similar to stop. Data is persisted in EBS root and other volumes. RAM is NOT cleared
  • terminate – Terminate the instance, all data is lost except the one on saved EBS volume. Cannot restart a terminated instance
• Storage EBS
  • gp2 – General purpose SSD (solid state drive), used for general purpose transactions. IOPS-16000. Can use for root device
  • io1 – High performance SSD (solid state drive), used for high IOPS workload. IOPS-64000. Can use for root device
  • st1 – Throughput Optimized HDD (hard disk drive), used for database, data warehouse, and ETL workloads. Max IOPS 500MB/s. Cannot use for root device
  • sc1 – Throughput Optimized HDD (hard disk drive), used for file system workloads. Max IOPS 250MB/s. Cannot use for root device
  • EBS cannot be shared by instance
  • Set ‘Delete On Terminate’ flag to “N” if you want your Amazon EBS volume to persist outside the life of the instance
• Storage EFS
  • Elastic File System
  • Mount to instance using mount command
  • Can be shared by instances
  • It is like network file system
  • Can be access by on-premise file servers
• Dedicated host – Use for strict license requirement. It is dedicated physical server
• Dedicated instance – Instance running on physical hardware dedicated to a single customer
• Placement Group – Arrangement pattern of instance
  • Cluster – Instances close together inside an AZ. For low latency
  • Spread – Each instance on separate hardware. For individual critical application
  • Partition – Spreads instances across logical partitions. For distributed and replicated database
• Ephemeral Instance
  • Physically attach to instance, can terminate or reboot, Will lose all data. No stop or hibernate
  • Can be attached at launch time only
• Snapshot
  • Create snapshot to backup a volume. It can be used to create new volume in a different AZ or region
  • Snapshots are stored on s3 incrementally
  • For snapshot of an EBS volume that serves as a root device, you should stop the instance before taking the snapshot.
  • Can copy to different region to create new volume there
  • Can encrypt while copying to another region
• AMI
  • Can create image from snapshot
  • Unregister AMI to remove it
• Boot-strapping script will run at first boot to configure the instance (e.g. to install web server etc.)
• Get meta data of instance – curl http://169.254.169.254/latest/meta-data/
• Get user data of instance (bootstrap ) – curl http://169.254.169.254/latest/user-data/
• Get dynamic data of instance  – curl http://169.254.169.254/latest/dynamic/

Home | IAM | S3 | CloudFront | SQS | EC2 | RDS | Route53