Ultimate Guide to Storage Optimized EC2 Instance Types

Share this Post

AWS offers EC2 instance types for a variety of compute use cases. They come in different combinations of CPU, memory, storage, and networking capacity that provide the optimum balance of resources for your application needs. Instances are available in a range of sizes to help scale your AWS resources to match your workloads.

You can read our earlier article on EC2 instance types for a general overview. In this article, we will focus on Storage Optimized Instances that you can use to build the right mix of AWS resources for high utilization and savings.

AWS Storage optimized instances are well-suited for what applications?

These instances are ideal for workloads that need high, sequential read/write access to large data sets on local storage. They are optimized for low-latency, high random I/O performance with up to 2 million IOPS.

Instance performance

Instances on the most recent version of Amazon Linux 2 or the Linux Amazon Machine Image (AMI) provide the best disk throughput performance.

SSD I/O performance

When SSD-based instance storage volumes fill up, I/O performance goes down with increased latency. To more efficiently manage the space available for write operations, you can leave 10% of the volume unpartitioned so that the SSD controller can use it for over-provisioning. This reduces the amount of available storage but increases performance even if the disk is close to full capacity.

I3 and I3en Instances

These instances deliver SSD storage optimized for low latency, high random I/O performance, high sequential disk throughput. They can leverage data stores and architectures that manage data redundancy and availability. These instances have the lowest price per GB of SSD storage within Amazon EC2. 

Use Cases

  • NoSQL databases, such as Apache Cassandra, MongoDB Atlas, Redis
  • In-memory databases, such as SAP HANA and Aerospike
  • Scale-out transactional databases
  • High-frequency Online Transaction Processing (OLTP) systems
  • Data warehousing
  • Distributed file systems, such as Hadoop
  • Elasticsearch
  • Analytics workloads

Bare metal instances, i3.metal and i3en.metal, provide applications direct access to physical server resources, such as processors and memory. These instances are designed for non-virtualized workloads or workloads that may have license restrictions.

I3 and I3en instances are EBS optimized. They provide NVMe (Non-Volatile Memory Express) only storage.

What is the difference between i3 and i3en instances?

I3en instances cost up to 50% lower per GB of SSD instance storage. They provide 2.6x greater storage density in GB per vCPU, and up to 2.7x greater ratio of network bandwidth to vCPUs.

How are the latest I3 instances different from I2?

I3 instances have 4x higher I/O performance, 2x higher disk throughput, and 18% more SSD per CPU than I2 instances. I3 instances offer NVMe based SSDs compared to I2 with legacy blkfront storage access. I3 offers the lowest price per read IOPS of all EC2 instances.

I3 instances require running AMIs on a Virtual Private Cloud (VPC). Migrating from I2 to I3 requires additional infrastructure considerations around moving to a VPC environment.

Instance performance

For i3 and i3en instances with NVMe instance store volumes, a Linux AMI with kernel version 4.4 or later is required. Otherwise, instances will not deliver the maximum I/O performance.

Network performance

Increased bandwidth for enhanced networking can be enabled on I3 and I3en instances to deliver lower latencies and network jitter, and higher packet-per-second (PPS) performance.

D2 Instances

D2 instances feature up to 48 TB of HDD-based local storage, deliver high disk throughput with the lowest cost on Amazon EC2. They are EBS optimized by default.These instances offer one of the lowest price per GiB of RAM among Amazon EC2 instance types.

Use Cases

  • Massively Parallel Processing (MPP) data warehousing
  • MapReduce and Hadoop distributed computing
  • Distributed file systems
  • Network file systems
  • Log or data-processing applications

H1 instances

These instances are designed for applications that require low cost, high disk throughput, and high sequential disk I/O access to very large data sets. They provide the best price/performance in the magnetic disk storage in EC2 instances. They are EBS optimized by default.

Use Cases

These instances have the largest memory of any EC2 instance and are designed for the following:
  • Data-intensive workloads such as MapReduce-based workloads
  • Distributed file systems, such as Hadoop DFS and MapR-FS
  • Network file systems
  • Log or data processing applications such as Apache Kafka
  • Big data workload clusters

Network performance

Enhanced networking can be enabled on h1 instances to deliver lower latencies and network jitter, and higher packet-per-second (PPS) performance.

Managing AWS Spend for Storage Optimized EC2 Instances

Cost management is key to optimizing your AWS architecture. As your storage requirements increase your costs can quickly add up. To control your cloud spend you can leverage AWS pricing options that provide significant discounts in return for volume purchases.

Reserved Instances

Production environments with steadily increasing workloads will benefit from Reserved Instances. A commitment to purchase a set number of storage optimized instances for 1 to 3 years will yield savings of up to 60% over on-demand rates. Reserved instances need to be continuously monitored to ensure that they are allocated to match usage. Critical decisions, such as migrating from I2 to the latest generation I3 instances, require a thorough knowledge of RIs to choose the right option, such as convertible vs standard RIs. In addition, you may need to sell unused instances in the AWS Marketplace to recover costs. A cost management tool like Parquantix will automatically make these adjustments based on your application needs and usage trends. It will migrate to the latest versions as they are released, without any involvement required from you.

Savings Plans

Savings Plans will provide discounted rates when you commit to a specific dollar amount of instance usage over a 1 or 3 year period.

The table below illustrates how much you can save on storage optimized instances through volume based discount plans on AWS. An i3.xlarge running on Linux in US East (N. Virginia) can provide significant savings over on-demand rates.

Automated Cost Optimization

Managing your cloud architecture with the right storage optimized instances through precise pricing options requires considerable time and effort. Parquantix develops a cost optimization strategy based on your application needs and constantly monitors usage to maximize instance utilization and performance. Our AI-driven cost management tool ensures that your instances deliver the best value.

Are you ready to start optimizing your storage-intensive applications?

Contact us to schedule a 30-minute consultation:
  • This field is for validation purposes and should be left unchanged.