Elastic Block Store¶

TLDR¶

This is network storage attached to an EC2 instance.

Features¶

• Network Drive
• persist Data after Termination
• can only be mounted to one instance at once
• limited to AZ which it was created
• can be attached and detached quickly
• need snapshot to move AZ
• uses provisioned capacity (Size and IOPS)
• you get billed even if drive is not full
• delete on terminiation is an option
• can be modified while in use (iops and capacity)

Snapshot¶

• is a Backup, can be done durin attachment
• Backups can be copied to diffrent region
• snapshot can be done while the volume is being used

EBS Snapshot Archive¶

• cheaper Snapshot storage but longer restore time

EBS Recycle Bin¶

• can enable recycle for recovery (retention 1day - 1 year)

Fast Snapshot Restore (FSR)¶

• cost alot
• no latency with restore

Volume types¶

• only ssd volumes can be used as boot volumes

GP2/GP3¶

• general purpose SSD
• cost effective Storage, low latency
• System boot volumes, virtual desktops, development and test environments
• 1 GiB - 16 TiB

gp3¶

• baselane of 3k iops and 125mibs
• can be increased to 16k iops and 1k mibs

gp2¶

• small
• burst up to 3k iops
• size of volume and iops are linked up to 16k iops max
• 3 iops per gb

IO1/IO2¶

• high performance SSD
• mission critical, low latency
• 4Gib - 16TiB
• max piops 64k for nitro ec2
• max piops 32k for non nitro
• io2 ist just better
• iops gb ratio 50:1 (10GB max 500 iops)

io2 block express¶

• sub milisecond latency
• max piops 256k
• iops gb ratio of 1000:1

hdds¶

• 125mb - 16 tb
• cannot be boot volume

st1 - low cost hdd volume¶

• designed for frequently access and throughput intensive
• low cost
• bit data workloads , data warehouses, log processing
• max 500mbs and 500 iops

sc1¶

• cold ssd
• hdd less frequently accessed workloads (lowest cost)
• 250mbs and 250 iops

provisioned iops ssd (piops)¶

• critical buissness application with sustained iops performance
• application which need more than 16k iops
• good for database workloads, good storage pref and consistency
• support ebs multi attach

Instance Store¶

• highest performance but ephemeral
• is lost even on hibernation
• cannot be detached and attached to other EC2

EBS Multi attach¶

• Same AZ can attach same volume to multiple instances
• only available for io2 and io1
• each instance has full permissions
• cluster linux applications
• application must manage concurrency
• max of 16 EC2 instances
• file system must be cluster aware

EBS Encryption¶

• Data at rest is encrypted
• Data is encrypted in flight between EC2 and EBS
• Snapshot is encrypted
• Volumes created from encrypted snapshot are encrypted aswell
• handles by aws behind the scenes
• minimal impact of latency
• uses keys from KMS (AES-256)
• copying and unencrypted snapshot enables encryption

Encrypt currently not encrypted¶

• created snaphsot
• encrypt snapshot
• create volume from snapshot
• attach new volume

Raid¶

Raid 0¶

• use if IO is more important than fault tolerance

Raid 1¶

• use if fault tolerance is more important than IO