What are the AWS EBS volume types?
General Purpose SSD (gp2 and gp3) volumes offer cost-effective storage that is ideal for a broad range of workloads.
Provisioned IOPS SSD (io1 and io2) volumes provide low latency and are designed to meet the needs of I/O-intensive workloads. They are best for EBS-optimized instances.
Throughput Optimized HDD (st1) volumes provide low-cost magnetic storage that is a good fit for large, sequential workloads.
Cold HDD (sc1) volumes provide low-cost magnetic storage that offers lower throughput than st1. sc1 is a good fit for large, sequential cold-data workloads that require infrequent access to data.
Magnetic
Magnetic (standard) volumes are best suited for workloads where data is accessed infrequently.
Name 4 characteristics of the General Purpose SSD gp2 volume type.
- Suitable for boot disks and general applications.
- Up to 16,000 IOPS per volume.
- Up to 99.9% durability.
- Small as 1GB, large as 16TB.
Name 3 characteristics of the General Purpose SSD gp3 volume type.
- Suitable for high-performance applications.
- Predictable 3,000 IOPS baseline performance and 125 MiB/s regardless of volume size.
- Up to 99.9% durability.
Name 5 characteristics of the Provisioned IOPS SSD io1 volume type.
- Suitable for OLTP and latency-sensitive applications.
- 50 IOPS/GiB(MAX).
- Up to 64,000 IOPS per volume.
- High performance and most expensive.
- Up to 99.9% durability.
- 4GB - 16TB
Name 5 characteristics of the Provisioned IOPS SSD io2 volume type.
- Suitable for OLTP and latency-sensitive applications.
- 500 IOPS/GiB(MAX).
- Up to 64,000 IOPS per volume.
- 99.999% durability
- 4GB - 16TB
Name 5 characteristics of the Throughput Optimized HDD st1 volume type.
- Suitable for big data, data warehouses, and ETL.
- Max throughput is 500 MB/s per volume.
- 500 IOPS(1MB IO Size).
- It cannot be a boot volume.
- Up to 99.9% durability.
- 125GB - 16TB.
Name 5 characteristics of the Cold HDD sc1 volume type.
- Max throughput of 250 MB/s per volume.
- 250 IOPS(1MB IO Size).
- Less frequently accessed data.
- It cannot be a boot volume.
- Lowest cost.
- Up to 99,9% durability.
- 125GB - 16TB.
Compare SSD and HDD.
SSDs (Solid State Drives) are much faster than HDDs (Hard Disk Drives). They also have lower latency and are more durable. However, they are also more expensive and have less capacity.
HDDs are slower than SSDs, but they have more capacity and are less expensive. They are a good option for applications that do not require high speed or low latency.
Hints on Snapshots and volumes.
- Snapshots are point-in-time photographs of data on the volume and are incremental in nature and linked to each other.
- Volumes exist on EBS, whereas snapshots exist on S3.
- The first snapshot will take some time to create(will not affect EBS performance), For consistent snapshots, stop the instance and detach the volume.
- You can share snapshots between regions, but first, you need to copy that snapshot to the target region.
- You can resize EBS volumes on the fly as well as change the volume types.
EBS vs Instance Store
- Instance store volumes are sometimes called ephemeral storage.
- Instance store volumes cannot be stopped. If the instance moves, resizes or the underlying host fails, you will lose your data.
- **EBS-backed instances can be stopped without data loss.
- You can reboot both EBS and instance store volumes and you will not lose your data
- By default, both root volumes will be deleted on termination. However, with EBS volumes, you can tell AWS to keep the root device volume.
- Instance store volume is included in instance price (if present).
- Instance store volumes can only be attached at launch time
- D3 instance has 4.6 GB/s throughput. i3 = 16 Gb/s. can reach millions of IOPS.
- High performance, not persistence, not resilience.
Explain encryption in volumes.
- Data at rest is encrypted inside the volume(KMS is used). Data are unecrypted only inside the RAM of the instance.
- All the data in flight moving between the instance and the volume is encrypted.
- All snapshots are encrypted.
- All volumes created from the snapshot are encrypted.
- Accounts can be set to encrypt by default or use their own KMS key.
- KMS creates 1 unique DEC(data encryption key) for each volume.
- Snapshots and future volumes use the same DEC.
- Cannot change a volume to not be encrypted.
- The OS is not aware of the encryption… no performance loss.
How to encrypt unencrypted volumes?
- Create a snapshot of the unencrypted root device volume.
- Create a copy of the snapshot and select the encrypt option.
- Create an AMI from the encrypted snapshot.
- Use that AMI to launch new encrypted instances.
Is EBS a global, regional, or Availability Zone (AZ) service within AWS?
Amazon EBS is an AZ service within AWS.
What is the relationship between IOPS, throughput, and IO size?
IOPS stands for Input/Output Operations Per Second and measures the number of read and write operations that a storage device can perform per second.
Throughput measures the amount of data that can be transferred to and from a storage device per second.
IO size refers to the size of each read and write operation.
The relationship between IOPS, throughput, and IO size can be summarized by the following equation: Throughput = IOPS * IO size
This equation means that the throughput of a storage device is equal to the number of read and write operations per second (IOPS) multiplied by the size of each operation (IO size).
For example, a storage device with an IOPS of 10,000 and an IO size of 1 KB will have a throughput of 10 MB/s.
What are GP2 EBS volume I/O credits and what is the maximum IOPS that GP2 EBS volumes can burst to?
GP2 EBS volume I/O credits are a resource that allows GP2 EBS volumes to burst above their baseline performance. GP2 volumes earn I/O credits at a rate of 3 IOPS per GiB per second, and they can use their I/O credits to burst to up to 3,000 IOPS for up to 30 minutes(When full - 54 million credits).
Note: GP2 EBS volumes with a baseline IOPS of 16,000 IOPS (volumes greater than 1 TB) can burst to 16,000 IOPS for up to 30 minutes.
GP2 EBS volume I/O credits are a valuable resource that can be used to improve the performance of GP2 volumes. However, it is important to use them wisely to avoid being throttled back to the baseline IOPS. Great for boots and initial workloads, dev, and test.
How is GP3 different from GP2?
GP3 is a newer generation of general-purpose EBS volumes that offers several advantages over GP2, including:
- No credit bucket architecture: GP3 volumes do not use the credit bucket architecture of GP2 volumes. This means that GP3 volumes can consistently achieve their baseline performance, regardless of how they are used.
- Higher baseline performance: All GP3 volumes have a baseline IOPS of 3,000 IOPS, which is 10 times higher than the baseline IOPS of GP2 volumes.
- Lower base price: GP3 volumes are 20% cheaper than GP2 volumes on a per-GiB basis.
- Higher throughput: GP3 volumes can achieve up to 1,000 MiB/s of throughput, which is four times higher than the maximum throughput of GP2 volumes.
Overall, GP3 is a better choice for most general-purpose workloads than GP2. It offers higher performance, lower cost, and greater predictability.
What are io2 block express features?
- Up to 4000 MB/s throughout
- Up to 256 IOPS per volume
- 4GB - 64 TB storage
- 1000 IOPS/GiB(MAX).
- Latest generation Provisioned IOPS volume.
What is the maximum IOPS and throughput that you can achieve when using multiple Provisioned IOPS SSD volumes attached to an EC2 instance?
- io1: 256,000 IOPS, 7500 MB/s
- io2: 156,000 IOPS, 4750 MB/s
- io2 Block Express: 256,000 IOPS, 7500 MB/s
What are the recommended options to achieve optimal performance when restoring a volume from an EBS snapshot?
- Force a read of all data immediately after restoring volume from a snapshot. For example the dd tool can be used: dd if=/dev/xvdf of=/dev/null bs=1M
- Use FSR(Fast Snapshot Restore) - Immediate Restore. Up to 50 snaps per region. It costs more.
