Some types can only be mounted on one instance; others can be attached to multiple instances
Must specify capacity
Has 6 Volume Types:
gp2
SDD
For general purposes. Balances price and performance.
Good for random reads/writes
Can be a boot drive
1 GB – 16 TiB
Max IOPS 16,000
Volume size and IOPS are linked (difference with gp3)
gp3
Similar to gp3 but IOPS is not linked to the volume size
io1
SDD
Good for random reads/writes
Use this if you want a Provisioned IOPS (sustained IOPS (like databases) or if you require more than 16,000 IOPS).
4 GB – 16 TiB
Can be a boot drive
Max IOPS is 32,000 (x2 of gp2/gp2) or 64,000 (if using EC2 Nitro)
Supports Multi-attach (for Linux and Windows) but requires:
instances must be on the same single region
not more than 16 instances built on the Nitro System
same Availability Zone
io2
Similar to io1 but newer
With Block Express:
Size can be from 4 GB – 64TiB
Max IOPS 256,000
Supports Multi-attach for Linux only
st1
HDD
Good for sequential reads/writes
Use cases: Data Warehousing, Log Processing, Big Data
125 MB – 16TiB
Max Throughput is 500MiB
Cannot be the boot volume
sc1
HDD
Has the lowest cost
Good for infrequently accessed data
125 MB – 16TiB
Max Throughput is 250MiB
Encryption:
Not enabled by default.
When a volume is encrypted:
Data at rest is encrypted
Data at flight is encrypted
Snapshots are encrypted
Volumes from snapshots are encrypted
If a volume is un-encrypted it will be un-encrypted throughout its lifetime and then its snapshot is un-encrypted as well.
How to encrypt and un-encrypted volume
Create a snapshot of the volume
Copy the snapshot to the new one but enable encryption
Create a volume for the encrypted snapshot
Attach the new volume to the EC2.
INSTANCE STORE
This is basically the storage of the Compute node
Better performance than EBS.
You will lose your storage when the EC2 is stopped (not restarted). The reason is that when a stopped EC2 is brought back, it may be on a different compute node.
Good for cache, temporary storage, buffer.
Only available on some EC2 instance types
EFS
A POSIX NFS filesystem.
Only works with Linux
Used for sharing file storage
No need to provision the size. Will grow automatically. Can grow in petabytes
Can have a max 1000 NFS clients connection.
Can control access through File System Policy.
File System Type:
Regional – redundant across all Availability Zones within an AWS Region.
One Zone – within a single Availability Zone.
To mount to an EC2 Linux:
mount -t nfs file-system-id.efs.aws-region.amazonaws.com:/ /<mount point>
The domain name will resolve to the IP address of the “mount target id” on the same AZ as the EC2.
Access Points:
are application-specific entry points into an EFS file system that make it easier to manage application access to shared datasets.
can enforce a user identity, including the user’s POSIX groups, for all file system requests that are made through the access point.
can enforce a different root directory for the file system so that clients can only access data in the specified directory or its subdirectories
Performance Mode:
Influence latency and IOPS
Cannot be changed once EFS is created
Has 2 modes:
General Purpose
default performance mode
recommended for the majority of workloads and faster performance
Max IO
previous generation performance type that is designed for highly parallelized workloads that can tolerate higher latencies than the General Purpose mode
recommended for large-scale workloads
scale to higher levels of aggregate throughput and operations per second
not supported in One-Zone File System Type
Throughput Mode
Has 2 modes:
Bursting
default throughout mode
scale based on storage size
baseline 1TiB = 50 MiB/sec
Provisioned
provision a fixed throughput regardless of the size of the file system
Elastic
spiky or unpredictable workloads and performance requirements that are difficult to forecast, or
your application drives throughput at an average-to-peak ratio of 5% or less.
Storage Types
Standard
default storage
Infrequently Accessed
lifecycle management will move data to this storage after N days
SNOW FAMILY
A physical device used to transfer a large amount of data to AWS or for Edge computing (e.g. at the ship, at mining location)
Can run EC2 instances or lambda functions.
Supports 80 TiB block or S3-compatible storage
3 Types of Devices:
Snowcone(Discontinued)
8 Tib of storage storage
Can connect to thenetwork
Data sync agent installed.
2 CPU 4GiB RAM
Snowball Edge
Can do clustering – i.e. three or more Snowball Edge devices used as a single logical unit for local storage and compute purposes.
It may take up to 4 weeks to provision and prepare the device for your job before it is shipped.
Has 2 flavours:
Storage Optimize
Has 210 TiB of NVME storage
104 vCPU 416GiB of RAM
Compute Optimize
Has 28 TiB storage
104 vCPU 416 GiB of NVME SSD RAM
(with AMD EPYC Gen2)
Optional GPU
Snowmobile
Can hold exabytes of data
migrate large datasets of 10PB or more in a single location
Use OpsHub (software installed on a machine to manage Snowcone or Snowball Edge.
FSX
A high-performance file system
Has 3 offerings:
FsX for Lustre
FsX for Windows File Server
FsX for NetApp ONTAP
FsX for Windows File Server
Fully managed Windows File System for sharing (like EFS for Windows)
Lustre s a type of parallel distributed file system, generally used for large-scale cluster computing (src Wikipedia)
For HPC
Millions of IOPS
Can integrate with S3 (expose S3 as a file system)
Can be accessed from on-prem
Support POSIX protocol
Deployment Options:
Scratch File System – temporary storage, single copy
Persistent File – long term storage, data replicated within the same AZ
STORAGE GATEWAY
A hybrid cloud storage service that gives you on-premises access to virtually unlimited cloud storage.
Exposes S3, FsX for Windows, EBS Snapshot & Glacier from on-prem.
Requires an agent running on a VM or an appliance ordered from AWS.
A client can connect either through a public network or AWS Direct.
4 Gateway Types:
S3 File Gateway
Presents a file interface that enables you to store files as objects in Amazon S3
Backed by S3
Can integrate with AD
The client talks to the agent using NFS or SMB protocol
Most recently used data will be cached by the agent.
FsX File Gateway
On-premises access to Windows file shares on Amazon FSx
Can cache frequently accessed data (unlike accessing FsX directly)
Reading and writing files, are all performed against the local cache, while Amazon FSx File Gateway synchronizes changed data to FSx for Windows File Server in the background
Volume Gateway
Presents your applications block storage volumes using the iSCSI protocol (like a disk)
Backed by EBS and S3 snapshot
2 Types of Volumes:
Cached – data is written to S3, while retaining your frequently accessed data locally in a cache for low-latency access. Maximum of 1 PB per gateway.
Stored – primary data is stored locally and your entire dataset is available for low-latency access while asynchronously backed up to AWS. Maximum of 512TB per gateway
Tape Gateway
Cloud-based Virtual Tape Library (VTL)
Works with leading backup software
Backed by S3, Glacier and Glacier Deep Archive
TRANSFER FAMILY
A fully managed FTP service
Support the following protocols:
SFTP
FTPS
FTP
HA, multi-AZ
User access the FTP endpoint directly or through Route 53