[SAA + SAP] 12. EC2 Stroage (EBS + EFS)

SAA

EBS

Highly Available

Automatically replicated within a single Availability Zone to protect against hardware failures

Scalable

Dynamicly increase capacity and chagne the type volume with no downtime or performance impact to your live systems.

EBS Snapshot

• Snapshot is available in Region
• "Copy snapshot" can help to move the EBS to another region
• "Create Volume" can help to move EBS to another AZ

 

AMI

 

 

• For GP3, the IOPS can be set independently
• But for GP2, IOPS and volume are linked together

• You got already 8000GB >5334 GB, increase volume size won't help anymore

 

• Nitro EC2 for better IOPS

• Only for io1/2
• In same AZ

• EBS is always AZ bounded, so always Same AZ
• Multi attach is the same Volume attach to mulit EC2 instances 

 

 

• Performance
• But one Volume fail, whole data lost

• No change in Performance
• but tolerance

 

EFS

• Use cases: content management, web serving, data sharing, Wordpress
• Uses NFSv4.1 protocol
• Uses security group to control access to EFS

• Encryption at rest using KMS
• POSIX file system (Linux) that has a standard file API
• FIle system scales automatically, pay-per-use, no capacity planning

 

• Multi AZs

 

• General purpose / MAX I/O
• Bursting / Provisioned (set throughput regardless of storage size)

 

 

---

 

 

SAP

EBS - Volume Types

• gp2: General Purpose Volumes (cheap)
  • You get 3 IOPS / GiB
  • minimum 100 IOPS
  • as long as under 3000 IOPS, you are able to burst to 3000 IOPS
  • max 16000 IOPS
  • Volume size from 1 GiB - 16 TiB
  • increase 1 TiB = Increase 3000 IOPS (1 TiB = 1000 GiB)
  • After 5.5 TiB, you won't get any IOPS increase
• io 1: Provisioned IOPS
  • Min 100 IOPS, Max 64000 IOPS (for Nitro instances) or 32000 (other instances type)
  • 4 GiB - 16 TiB. Size of volume and IOPS are independent
• st 1: Throughput Optimized HDD
  • 500 GiB - 16 TiB, 500 MiB/s throughput
• sc 1: Cold HDD, Infrequently accessed data
  • 250 GiB - 16 TiB, 250 MiB/s throughput

EBS - Snapshot

• Snapshots will be stored in S3 (but you won't directly see them)
• Copy snapshots across regions (for DR)
• Can make Image (AMI) from snapshot
• EBS volumes restored by snapshots need to be pre-warmed
• Snapshots can be automated using Amazon Data Lifecycle Manager

 

EFS

• Can only attach to one VPC, create one ENI per AZ

• You can has EFS inside a VCP
• With ENI attach to each AZ
• Your EC2 instance can locate in another VPC
• Talk to EFS by using VPC peering

• On-Premise Server can also talk to EFS
• You can use Direct Connect or/and Site-to-Site VPN (due to whether you need redundancy in DX, failover)
• Access EFS only by private IPv4, not DNS name
• Which means you EFS can work across Multi accounts and even on-premise

 

• NFS by itself is not considered a secure protocol
• So not recommend running it over the raw internet without DC or VPN
• You can also use DataSync to keep the storage in sync with EFS or S3 over a DC and it does so securely
• DataSync supports EFS to EFS Sync

 

• On-prmise use DataSync to connect to EFS
• EFS mount points to multi AZ for EC2