Amazon EBS volume types - Amazon Elastic Compute Cloud
Solid state drives (SSD)Hard disk drives (HDD)Previous generation volume typesGeneral Purpose SSD volumes (gp3)General Purpose SSD volumes (gp2)Provisioned IOPS SSD volumesThroughput Optimized HDD volumesCold HDD volumesMagnetic volumesPerformance considerations when using HDD volumesMonitor the burst bucket balance for volumes

Amazon EBS volume types

Amazon EBS provides the following volume types, which differ in performance characteristics and price, so that you can tailor your storage performance and cost to the needs of your applications. The volumes types fall into these categories:

  • Solid state drives (SSD) — Optimized for transactional workloads involving frequent read/write operations with small I/O size, where the dominant performance attribute is IOPS.

  • Hard disk drives (HDD) — Optimized for large streaming workloads where the dominant performance attribute is throughput.

  • Previous generation — Hard disk drives that can be used for workloads with small datasets where data is accessed infrequently and performance is not of primary importance. We recommend that you consider a current generation volume type instead.

There are several factors that can affect the performance of EBS volumes, such as instance configuration, I/O characteristics, and workload demand. To fully use the IOPS provisioned on an EBS volume, use EBS-optimized instances. For more information about getting the most out of your EBS volumes, see Amazon EBS volume performance on Windows instances.

For more information about pricing, see Amazon EBS Pricing.

Solid state drives (SSD)

The SSD-backed volumes provided by Amazon EBS fall into these categories:

  • General Purpose SSD — Provides a balance of price and performance. We recommend these volumes for most workloads.

  • Provisioned IOPS SSD — Provides high performance for mission-critical, low-latency, or high-throughput workloads.

The following is a summary of the use cases and characteristics of SSD-backed volumes. For information about the maximum IOPS and throughput per instance, see Amazon EBS–optimized instances.

General Purpose SSD Provisioned IOPS SSD
Volume type gp3 gp2 io2 Block Express ‡ io2      io1
Durability 99.8% - 99.9% durability (0.1% - 0.2% annual failure rate) 99.8% - 99.9% durability (0.1% - 0.2% annual failure rate) 99.999% durability (0.001% annual failure rate) 99.999% durability (0.001% annual failure rate) 99.8% - 99.9% durability (0.1% - 0.2% annual failure rate)
Use cases

  • Low-latency interactive apps

  • Development and test environments

Workloads that require:

  • Sub-millisecond latency

  • Sustained IOPS performance

  • More than 64,000 IOPS or 1,000 MiB/s of throughput

  • Workloads that require sustained IOPS performance or more than 16,000 IOPS

  • I/O-intensive database workloads
Volume size 1 GiB - 16 TiB 4 GiB - 64 TiB 4 GiB - 16 TiB
Max IOPS per volume (16 KiB I/O) 16,000 256,000 64,000 †
Max throughput per volume 1,000 MiB/s 250 MiB/s * 4,000 MiB/s 1,000 MiB/s †
Amazon EBS Multi-attach Not supported Supported
Boot volume Supported

* The throughput limit is between 128 MiB/s and 250 MiB/s, depending on the volume size. Volumes smaller than or equal to 170 GiB deliver a maximum throughput of 128 MiB/s. Volumes larger than 170 GiB but smaller than 334 GiB deliver a maximum throughput of 250 MiB/s if burst credits are available. Volumes larger than or equal to 334 GiB deliver 250 MiB/s regardless of burst credits. gp2 volumes that were created before December 3, 2018 and that have not been modified since creation might not reach full performance unless you modify the volume.

† Maximum IOPS and throughput are guaranteed only on Instances built on the Nitro System provisioned with more than 32,000 IOPS. Other instances guarantee up to 32,000 IOPS and 500 MiB/s. io1 volumes that were created before December 6, 2017 and that have not been modified since creation might not reach full performance unless you modify the volume.

‡ io2 Block Express volumes are supported with R5b instances only. io2 volumes attached to an R5b instance during or after launch automatically run on Block Express. For more information, see io2 Block Express volumes.

Hard disk drives (HDD)

The HDD-backed volumes provided by Amazon EBS fall into these categories:

  • Throughput Optimized HDD — A low-cost HDD designed for frequently accessed, throughput-intensive workloads.

  • Cold HDD — The lowest-cost HDD design for less frequently accessed workloads.

The following is a summary of the use cases and characteristics of HDD-backed volumes. For information about the maximum IOPS and throughput per instance, see Amazon EBS–optimized instances.

Throughput Optimized HDD Cold HDD
Volume type st1 sc1
Durability 99.8% - 99.9% durability (0.1% - 0.2% annual failure rate) 99.8% - 99.9% durability (0.1% - 0.2% annual failure rate)
Use cases

  • Big data

  • Data warehouses

  • Log processing

  • Throughput-oriented storage for data that is infrequently accessed

  • Scenarios where the lowest storage cost is important
Volume size 125 GiB - 16 TiB 125 GiB - 16 TiB
Max IOPS per volume (1 MiB I/O) 500 250
Max throughput per volume 500 MiB/s 250 MiB/s
Amazon EBS Multi-attach Not supported Not supported
Boot volume Not supported Not supported

Previous generation volume types

The following table describes previous-generation EBS volume types. If you need higher performance or performance consistency than previous-generation volumes can provide, we recommend that you consider using General Purpose SSD (gp2 and gp3) or other current volume types. For more information, see Previous Generation Volumes.

Magnetic
Volume type standard
Use cases Workloads where data is infrequently accessed
Volume size 1 GiB-1 TiB
Max IOPS per volume 40–200
Max throughput per volume 40–90 MiB/s
Boot volume Supported

General Purpose SSD volumes (gp3)

General Purpose SSD (gp3) volumes offer cost-effective storage that is ideal for a broad range of workloads. These volumes deliver a consistent baseline rate of 3,000 IOPS and 125 MiB/s, included with the price of storage. You can provision additional IOPS (up to 16,000) and throughput (up to 1,000 MiB/s) for an additional cost.

The maximum ratio of provisioned IOPS to provisioned volume size is 500 IOPS per GiB. The maximum ratio of provisioned throughput to provisioned IOPS is .25 MiB/s per IOPS. The following volume configurations support provisioning either maximum IOPS or maximum throughput:

  • 32 GiB or larger: 500 IOPS/GiB x 32 GiB = 16,000 IOPS

  • 8 GiB or larger and 4,000 IOPS or higher: 4,000 IOPS x 0.25 MiB/s/IOPS = 1,000 MiB/s

General Purpose SSD volumes (gp2)

General Purpose SSD (gp2) volumes offer cost-effective storage that is ideal for a broad range of workloads. These volumes deliver single-digit millisecond latencies and the ability to burst to 3,000 IOPS for extended periods of time. Between a minimum of 100 IOPS (at 33.33 GiB and below) and a maximum of 16,000 IOPS (at 5,334 GiB and above), baseline performance scales linearly at 3 IOPS per GiB of volume size. AWS designs gp2 volumes to deliver their provisioned performance 99% of the time. A gp2 volume can range in size from 1 GiB to 16 TiB.

I/O Credits and burst performance

The performance of gp2 volumes is tied to volume size, which determines the baseline performance level of the volume and how quickly it accumulates I/O credits; larger volumes have higher baseline performance levels and accumulate I/O credits faster. I/O credits represent the available bandwidth that your gp2 volume can use to burst large amounts of I/O when more than the baseline performance is needed. The more credits your volume has for I/O, the more time it can burst beyond its baseline performance level and the better it performs when more performance is needed. The following diagram shows the burst-bucket behavior for gp2.

gp2 burst bucket

Each volume receives an initial I/O credit balance of 5.4 million I/O credits, which is enough to sustain the maximum burst performance of 3,000 IOPS for at least 30 minutes. This initial credit balance is designed to provide a fast initial boot cycle for boot volumes and to provide a good bootstrapping experience for other applications. Volumes earn I/O credits at the baseline performance rate of 3 IOPS per GiB of volume size. For example, a 100 GiB gp2 volume has a baseline performance of 300 IOPS.

Comparing baseline performance and burst IOPS

When your volume requires more than the baseline performance I/O level, it draws on I/O credits in the credit balance to burst to the required performance level, up to a maximum of 3,000 IOPS. When your volume uses fewer I/O credits than it earns in a second, unused I/O credits are added to the I/O credit balance. The maximum I/O credit balance for a volume is equal to the initial credit balance (5.4 million I/O credits).

When the baseline performance of a volume is higher than maximum burst performance, I/O credits are never spent. If the volume is attached to an instance built on the Nitro System, the burst balance is not reported. For other instances, the reported burst balance is 100%.

The burst duration of a volume is dependent on the size of the volume, the burst IOPS required, and the credit balance when the burst begins. This is shown in the following equation: 

                             (Credit balance)
Burst duration  =  ------------------------------------
                   (Burst IOPS) - 3(Volume size in GiB)

The following table lists several volume sizes and the associated baseline performance of the volume (which is also the rate at which it accumulates I/O credits), the burst duration at the 3,000 IOPS maximum (when starting with a full credit balance), and the time in seconds that the volume would take to refill an empty credit balance.

Volume size (GiB)

Baseline performance (IOPS)

Burst duration when driving sustained 3,000 IOPS (second)

Seconds to fill empty credit balance when driving no IO

1

100

1,802

 54,000

100

300

2,000

 18,000

250

750

 2,400 7,200

334 (Min. size for max throughput)

 1,002

2,703

5,389

500

1,500

3,600

 3,600

750

2,250

7,200

 2,400

1,000

3,000

N/A*

N/A*

5,334 (Min. size for max IOPS)

16,000

N/A*

N/A*

16,384 (16 TiB, max volume size)

16,000

N/A*

N/A*

* The baseline performance of the volume exceeds the maximum burst performance.

What happens if I empty my I/O credit balance?

If your gp2 volume uses all of its I/O credit balance, the maximum IOPS performance of the volume remains at the baseline IOPS performance level (the rate at which your volume earns credits) and the volume's maximum throughput is reduced to the baseline IOPS multiplied by the maximum I/O size. Throughput can never exceed 250 MiB/s. When I/O demand drops below the baseline level and unused credits are added to the I/O credit balance, the maximum IOPS performance of the volume again exceeds the baseline. For example, a 100 GiB gp2 volume with an empty credit balance has a baseline performance of 300 IOPS and a throughput limit of 75 MiB/s (300 I/O operations per second * 256 KiB per I/O operation = 75 MiB/s). The larger a volume is, the greater the baseline performance is and the faster it replenishes the credit balance. For more information about how IOPS are measured, see I/O characteristics and monitoring.

If you notice that your volume performance is frequently limited to the baseline level (due to an empty I/O credit balance), you should consider switching to a gp3 volume.

For information about using CloudWatch metrics and alarms to monitor your burst bucket balance, see Monitor the burst bucket balance for volumes.

Throughput performance

Throughput for a gp2 volume can be calculated using the following formula, up to the throughput limit of 250 MiB/s: 

Throughput in MiB/s = ((Volume size in GiB) × (IOPS per GiB) × (I/O size in KiB))

Assuming V = volume size, I = I/O size, R = I/O rate, and T = throughput, this can be simplified to: 

T = VIR

The smallest volume size that achieves the maximum throughput is given by: 

        T          
V  =  -----  
       I R   

            250 MiB/s
   =  ---------------------
      (256 KiB)(3 IOPS/GiB)


               [(250)(2^20)(Bytes)]/s
   =  ------------------------------------------
      (256)(2^10)(Bytes)([3 IOP/s]/[(2^30)(Bytes)])


      (250)(2^20)(2^30)(Bytes)
   =  ------------------------
           (256)(2^10)(3)

   =  357,913,941,333 Bytes
   
   =  333⅓ GiB (334 GiB in practice because volumes are provisioned in whole gibibytes)

Provisioned IOPS SSD volumes

Provisioned IOPS SSD (io1 and io2) volumes are designed to meet the needs of I/O-intensive workloads, particularly database workloads, that are sensitive to storage performance and consistency. Provisioned IOPS SSD volumes use a consistent IOPS rate, which you specify when you create the volume, and Amazon EBS delivers the provisioned performance 99.9 percent of the time.

io1 volumes are designed to provide 99.8 to 99.9 percent volume durability with an annual failure rate (AFR) no higher than 0.2 percent, which translates to a maximum of twovolume failures per 1,000 running volumes over a one-year period. io2 volumes are designed to provide 99.999 percent volume durability with an AFR no higher than 0.001 percent, which translates to a single volume failure per 100,000 running volumes over a one-year period.

Provisioned IOPS SSD io1 and io2 volumes are available for all Amazon EC2 instance types. Provisioned IOPS SSD io2 volumes attached to R5b instances run on EBS Block Express. For more information, see io2 Block Express volumes.

Considerations for io2 volumes

  • Keep the following in mind when launching instances with io2 volumes:

    • If you launch an R5b instance with an io2 volume, the volume automatically runs on Block Express, regardless of the volume’s size and IOPS.

    • You can't launch an instance type that does not support Block Express with an io2 volume that has a size greater than 16 TiB or IOPS greater than 64,000.

    • You can't launch an R5b instance with an encrypted io2 volume that has a size greater than 16 TiB or IOPS greater than 64,000 from an unencrypted AMI or a shared encrypted AMI. In this case, you must first create an encrypted AMI in your account and then use that AMI to launch the instance.

  • Keep the following in mind when creating io2 volumes:

    • If you create an io2 volume with a size greater than 16 TiB or IOPS greater than 64,000 in a Region where Block Express is supported, the volume automatically runs on Block Express.

    • You can't create an io2 volume with a size greater than 16 TiB or IOPS greater than 64,000 in a Region where Block Express is not supported.

    • If you create an io2 volume with a size of 16 TiB or less and IOPS of 64,000 or less in a Region where Block Express is supported, the volume does not run on Block Express.

    • You can't create an encrypted io2 volume that has a size greater than 16 TiB or IOPS greater than 64,000 from an unencrypted snapshot or a shared encrypted snapshot. In this case, you must first create an encrypted snapshot in your account and then use that snapshot to create the volume.

  • Keep the following in mind when attaching io2 volumes to instances:

    • If you attach an io2 volume to an R5b instance, the volume automatically runs on Block Express. It can take up to 48 hours to optimize the volume for Block Express. During this time, the volume provides io2 latency. After the volume has been optimized, it provides the sub-millisecond latency supported by Block Express.

    • You can't attach an io2 volume with a size greater than 16 TiB or IOPS greater than 64,000 to an instance type that does not support Block Express.

    • If you detach an io2 volume with a size of 16 TiB or less and IOPS of 64,000 or less from an R5b instance and attach it to an instance type that does not support Block Express, the volume no longer runs on Block Express and it provides io2 latency.

  • Keep the following in mind when modifying io2 volumes:

    • You can't modify an io2 volume and increase its size beyond 16 TiB or its IOPS beyond 64,000 while it is attached to an instance type that does not support Block Express.

    • You can't modify the size or provisioned IOPS of an io2 volume that is attached to an R5b instance.

Performance

Provisioned IOPS SSD volumes can range in size from 4 GiB to 16 TiB and you can provision from 100 IOPS up to 64,000 IOPS per volume. You can achieve up to 64,000 IOPS only on Instances built on the Nitro System. On other instance families you can achieve performance up to 32,000 IOPS. The maximum ratio of provisioned IOPS to requested volume size (in GiB) is 50:1 for io1 volumes, and 500:1 for io2 volumes. For example, a 100 GiB io1 volume can be provisioned with up to 5,000 IOPS, while a 100 GiB io2 volume can be provisioned with up to 50,000 IOPS. On a supported instance type, the following volume sizes allow provisioning up to the 64,000 IOPS maximum:

  • io1 volume 1,280 GiB in size or greater (50 × 1,280 GiB = 64,000 IOPS)

  • io2 volume 128 GiB in size or greater (500 × 128 GiB = 64,000 IOPS)

Provisioned IOPS SSD volumes provisioned with up to 32,000 IOPS support a maximum I/O size of 256 KiB and yield as much as 500 MiB/s of throughput. With the I/O size at the maximum, peak throughput is reached at 2,000 IOPS. Volumes provisioned with more than 32,000 IOPS (up to the maximum of 64,000 IOPS) yield a linear increase in throughput at a rate of 16 KiB per provisioned IOPS. For example, a volume provisioned with 48,000 IOPS can support up to 750 MiB/s of throughput (16 KiB per provisioned IOPS x 48,000 provisioned IOPS = 750 MiB/s). To achieve the maximum throughput of 1,000 MiB/s, a volume must be provisioned with 64,000 IOPS (16 KiB per provisioned IOPS x 64,000 provisioned IOPS = 1,000 MiB/s). The following graph illustrates these performance characteristics:

Throughput limits for io1 volumes

Your per-I/O latency experience depends on the provisioned IOPS and on your workload profile. For the best I/O latency experience, ensure that you provision IOPS to meet the I/O profile of your workload.

io2 Block Express volumes

Note

io2 Block Express volumes are supported with R5b instances only.

io2 Block Express volumes is the next generation of Amazon EBS storage server architecture. It has been built for the purpose of meeting the performance requirements of the most demanding I/O intensive applications that run on Nitro-based Amazon EC2 instances.

Block Express architecture increases performance and scale. Block Express servers communicate with Nitro-based instances using the Scalable Reliable Datagram (SRD) networking protocol. This interface is implemented in the Nitro Card dedicated for Amazon EBS I/O function on the host hardware of the instance. It minimizes I/O delay and latency variation (network jitter), which provides faster and more consistent performance for your applications. For more information, see io2 Block Express volumes.

io2 Block Express volumes are suited for workloads that benefit from a single volume that provides sub-millisecond latency, and supports higher IOPS, higher throughput, and larger capacity than io2 volumes.

io2 Block Express volumes support the same features as io2 volumes, including Multi-Attach, Elastic Volume operations, and encryption.

Considerations

  • io2 Block Express volumes are currently supported with R5b instances only.

  • io2 Block Express volumes are currently available in all Regions where R5b instances are available, including us-east-1, us-east-2, us-west-2, ap-southeast-1, ap-northeast-1, and eu-central-1. R5b instance availability might vary by Availability Zone. For more information about R5b availability, see Find an Amazon EC2 instance type.

  • io2 Block Express volumes do not support fast snapshot restore. We recommend that you initialize these volumes to ensure that they deliver full performance. For more information, see Initialize Amazon EBS volumes.

Performance

With io2 Block Express volumes, you can provision volumes with:

  • Sub-millisecond average latency

  • Storage capacity up to 64 TiB (65,536 GiB)

  • Provisioned IOPS up to 256,000, with an IOPS:GiB ratio of 1,000:1. Maximum IOPS can be provisioned with volumes 256 GiB in size and larger (1,000 IOPS x 256 GiB = 256,000 IOPS).

  • Volume throughput up to 4,000 MiB/s. Throughput scales proportionally up to 0.256 MiB/s per provisioned IOPS. Maximum throughput can be achieved at 16,000 IOPS or higher.

Throughput limits for io2 Block Express volumes

Quotas

io2 Block Express volumes adhere to the same service quotas as io2 volumes. For more information, see Amazon EBS quotas.

Pricing and billing

io2 volumes and io2 Block Express volumes are billed at the same rate. For more information, see Amazon EBS pricing.

Usage reports do not distinguish between io2 Block Express volumes and io2 volumes. We recommend that you use tags to help you identify costs associated with io2 Block Express volumes.

Throughput Optimized HDD volumes

Throughput Optimized HDD (st1) volumes provide low-cost magnetic storage that defines performance in terms of throughput rather than IOPS. This volume type is a good fit for large, sequential workloads such as Amazon EMR, ETL, data warehouses, and log processing. Bootable st1 volumes are not supported.

Throughput Optimized HDD (st1) volumes, though similar to Cold HDD (sc1) volumes, are designed to support frequently accessed data.

This volume type is optimized for workloads involving large, sequential I/O, and we recommend that customers with workloads performing small, random I/O use gp2. For more information, see Inefficiency of small read/writes on HDD.

Throughput credits and burst performance

Like gp2, st1 uses a burst-bucket model for performance. Volume size determines the baseline throughput of your volume, which is the rate at which the volume accumulates throughput credits. Volume size also determines the burst throughput of your volume, which is the rate at which you can spend credits when they are available. Larger volumes have higher baseline and burst throughput. The more credits your volume has, the longer it can drive I/O at the burst level.

The following diagram shows the burst-bucket behavior for st1.

st1 burst bucket

Subject to throughput and throughput-credit caps, the available throughput of an st1 volume is expressed by the following formula: 

(Volume size) x (Credit accumulation rate per TiB) = Throughput

For a 1-TiB st1 volume, burst throughput is limited to 250 MiB/s, the bucket fills with credits at 40 MiB/s, and it can hold up to 1 TiB-worth of credits.

Larger volumes scale these limits linearly, with throughput capped at a maximum of 500 MiB/s. After the bucket is depleted, throughput is limited to the baseline rate of 40 MiB/s per TiB.

On volume sizes ranging from 0.125 TiB to 16 TiB, baseline throughput varies from 5 MiB/s to a cap of 500 MiB/s, which is reached at 12.5 TiB as follows: 

            40 MiB/s
12.5 TiB x ---------- = 500 MiB/s
             1 TiB

Burst throughput varies from 31 MiB/s to a cap of 500 MiB/s, which is reached at 2 TiB as follows: 

         250 MiB/s
2 TiB x ---------- = 500 MiB/s
          1 TiB

The following table states the full range of base and burst throughput values for st1:

Volume size (TiB) ST1 base throughput (MiB/s) ST1 burst throughput (MiB/s)
0.125 5 31
0.5 20 125
1 40 250
2 80 500
3 120 500
4 160 500
5 200 500
6 240 500
7 280 500
8 320 500
9 360 500
10 400 500
11 440 500
12 480 500
12.5 500 500
13 500 500
14 500 500
15 500 500
16 500 500

The following diagram plots the table values:

Comparing st1 base and burst throughput
Note

When you create a snapshot of a Throughput Optimized HDD (st1) volume, performance may drop as far as the volume's baseline value while the snapshot is in progress.

For information about using CloudWatch metrics and alarms to monitor your burst bucket balance, see Monitor the burst bucket balance for volumes.

Cold HDD volumes

Cold HDD (sc1) volumes provide low-cost magnetic storage that defines performance in terms of throughput rather than IOPS. With a lower throughput limit than st1, sc1 is a good fit for large, sequential cold-data workloads. If you require infrequent access to your data and are looking to save costs, sc1 provides inexpensive block storage. Bootable sc1 volumes are not supported.

Cold HDD (sc1) volumes, though similar to Throughput Optimized HDD (st1) volumes, are designed to support infrequently accessed data.

Note

This volume type is optimized for workloads involving large, sequential I/O, and we recommend that customers with workloads performing small, random I/O use gp2. For more information, see Inefficiency of small read/writes on HDD.

Throughput credits and burst performance

Like gp2, sc1 uses a burst-bucket model for performance. Volume size determines the baseline throughput of your volume, which is the rate at which the volume accumulates throughput credits. Volume size also determines the burst throughput of your volume, which is the rate at which you can spend credits when they are available. Larger volumes have higher baseline and burst throughput. The more credits your volume has, the longer it can drive I/O at the burst level.

sc1 burst bucket

Subject to throughput and throughput-credit caps, the available throughput of an sc1 volume is expressed by the following formula: 

(Volume size) x (Credit accumulation rate per TiB) = Throughput

For a 1-TiB sc1 volume, burst throughput is limited to 80 MiB/s, the bucket fills with credits at 12 MiB/s, and it can hold up to 1 TiB-worth of credits.

Larger volumes scale these limits linearly, with throughput capped at a maximum of 250 MiB/s. After the bucket is depleted, throughput is limited to the baseline rate of 12 MiB/s per TiB.

On volume sizes ranging from 0.125 TiB to 16 TiB, baseline throughput varies from 1.5 MiB/s to a maximum of 192 MiB/s, which is reached at 16 TiB as follows: 

           12 MiB/s
16 TiB x ---------- = 192 MiB/s
            1 TiB

Burst throughput varies from 10 MiB/s to a cap of 250 MiB/s, which is reached at 3.125 TiB as follows: 

             80 MiB/s
3.125 TiB x ----------- = 250 MiB/s
              1 TiB

The following table states the full range of base and burst throughput values for sc1:

Volume Size (TiB) SC1 Base Throughput (MiB/s) SC1 Burst Throughput (MiB/s)
0.125 1.5 10
0.5 6 40
1 12 80
2 24 160
3 36 240
3.125 37.5 250
4 48 250
5 60 250
6 72 250
7 84 250
8 96 250
9 108 250
10 120 250
11 132 250
12 144 250
13 156 250
14 168 250
15 180 250
16 192 250

The following diagram plots the table values:

Comparing sc1 base and burst throughput
Note

When you create a snapshot of a Cold HDD (sc1) volume, performance may drop as far as the volume's baseline value while the snapshot is in progress.

For information about using CloudWatch metrics and alarms to monitor your burst bucket balance, see Monitor the burst bucket balance for volumes.

Magnetic volumes

Magnetic volumes are backed by magnetic drives and are suited for workloads where data is accessed infrequently, and scenarios where low-cost storage for small volume sizes is important. These volumes deliver approximately 100 IOPS on average, with burst capability of up to hundreds of IOPS, and they can range in size from 1 GiB to 1 TiB.

Note

Magnetic is a previous generation volume type. For new applications, we recommend using one of the newer volume types. For more information, see Previous Generation Volumes.

For information about using CloudWatch metrics and alarms to monitor your burst bucket balance, see Monitor the burst bucket balance for volumes.

Performance considerations when using HDD volumes

For optimal throughput results using HDD volumes, plan your workloads with the following considerations in mind.

Comparing Throughput Optimized HDD and Cold HDD

The st1 and sc1 bucket sizes vary according to volume size, and a full bucket contains enough tokens for a full volume scan. However, larger st1 and sc1 volumes take longer for the volume scan to complete due to per-instance and per-volume throughput limits. Volumes attached to smaller instances are limited to the per-instance throughput rather than the st1 or sc1 throughput limits.

Both st1 and sc1 are designed for performance consistency of 90% of burst throughput 99% of the time. Non-compliant periods are approximately uniformly distributed, targeting 99% of expected total throughput each hour.

In general, scan times are expressed by this formula:

 Volume size
------------ = Scan time
 Throughput

For example, taking the performance consistency guarantees and other optimizations into account, an st1 customer with a 5-TiB volume can expect to complete a full volume scan in 2.91 to 3.27 hours.

  • Optimal scan time

       5 TiB            5 TiB
----------- = ------------------ = 10,486 seconds = 2.91 hours 
 500 MiB/s     0.00047684 TiB/s

  • Maximum scan time

      2.91 hours
-------------- = 3.27 hours
 (0.90)(0.99) <-- From expected performance of 90% of burst 99% of the time

Similarly, an sc1 customer with a 5-TiB volume can expect to complete a full volume scan in 5.83 to 6.54 hours.

  • Optimal scan time

    
   5 TiB             5 TiB
----------- = ------------------- = 20972 seconds = 5.83 hours 
 250 MiB/s     0.000238418 TiB/s

  • Maximum scan time

      5.83 hours
-------------- = 6.54 hours
 (0.90)(0.99)

The following table shows ideal scan times for volumes of various size, assuming full buckets and sufficient instance throughput.

Volume size (TiB) ST1 scan time with burst (hours)* SC1 scan time with burst (hours)*
1 1.17 3.64
2 1.17 3.64
3 1.75 3.64
4 2.33 4.66
5 2.91 5.83
6 3.50 6.99
7 4.08 8.16
8 4.66 9.32
9 5.24 10.49
10 5.83 11.65
11 6.41 12.82
12 6.99 13.98
13 7.57 15.15
14 8.16 16.31
15 8.74 17.48
16 9.32 18.64

* These scan times assume an average queue depth (rounded to the nearest whole number) of four or more when performing 1 MiB of sequential I/O.

Therefore if you have a throughput-oriented workload that needs to complete scans quickly (up to 500 MiB/s), or requires several full volume scans a day, use st1. If you are optimizing for cost, your data is relatively infrequently accessed, and you don’t need more than 250 MiB/s of scanning performance, then use sc1.

Inefficiency of small read/writes on HDD

The performance model for st1 and sc1 volumes is optimized for sequential I/Os, favoring high-throughput workloads, offering acceptable performance on workloads with mixed IOPS and throughput, and discouraging workloads with small, random I/O.

For example, an I/O request of 1 MiB or less counts as a 1 MiB I/O credit. However, if the I/Os are sequential, they are merged into 1 MiB I/O blocks and count only as a 1 MiB I/O credit.

Limitations on per-instance throughput

Throughput for st1 and sc1 volumes is always determined by the smaller of the following:

  • Throughput limits of the volume

  • Throughput limits of the instance

As for all Amazon EBS volumes, we recommend that you select an appropriate EBS-optimized EC2 instance in order to avoid network bottlenecks. For more information, see Amazon EBS–optimized instances.

Monitor the burst bucket balance for volumes

You can monitor the burst-bucket level for gp2, st1, and sc1 volumes using the EBS BurstBalance metric available in Amazon CloudWatch. This metric shows the percentage of I/O credits (for gp2) or throughput credits (for st1 and sc1) remaining in the burst bucket. For more information about the BurstBalance metric and other metrics related to I/O, see I/O characteristics and monitoring. CloudWatch also allows you to set an alarm that notifies you when the BurstBalance value falls to a certain level. For more information, see Creating Amazon CloudWatch Alarms.