Menu
Amazon Elastic Compute Cloud
User Guide for Linux (API Version 2015-04-15)

HS1 Instances

HS1 instances (hs1.8xlarge) provide very high storage density and high sequential read and write performance per instance. They are well suited for the following scenarios:

  • Data warehousing

  • Hadoop/MapReduce

  • Parallel file systems

You can cluster HS1 instances in a placement group. For more information, see Placement Groups.

By default, you can run up to two HS1 instances. If you need more than two HS1 instances, you can request more using the Amazon EC2 Instance Request Form.

Hardware Specifications

HS1 instances support both Amazon Elastic Block Store (Amazon EBS)-backed and instance store-backed Amazon Machine Images (AMIs). HS1 instances support both paravirtual (PV) and hardware virtual machine (HVM) AMIs.

HS1 instances provide high bandwidth networking and can also be used with Provisioned IOPS (SSD) volumes for improved consistency and performance.

For more information about the hardware specifications for each Amazon EC2 instance type, see Amazon EC2 Instances.

Instance Store

HS1 instances support both instance store and Amazon EBS root device volumes. However, even when using an Amazon EBS-backed instance, primary data storage is provided by the hard disk drives in the instance store. Like other instance store volumes, these instance store volumes persist only for the life of the instance. For more information about instance store volumes, see Amazon EC2 Instance Store.

Disk Initialization

If you plan to run an HS1 instance in a steady state for long periods of time, we recommend that you zero the hard disks first for improved performance. This process can take as long as six hours to complete.

Setting the Memory Limit

Many Linux-based AMIs come with CONFIG_XEN_MAX_DOMAIN_MEMORY set to 70. We recommend that you set this as appropriate for 117 GiB of memory.

Setting the User Limit (ulimit)

Many Linux-based AMIs come with a default ulimit of 1024. We recommend that you increase the ulimit to 2048.