Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage News AWS Launches Massive X1 Instances Targeting High Memory Workloads

AWS Launches Massive X1 Instances Targeting High Memory Workloads

AWS recently added a new instance type with nearly 2 terabytes of memory and 128 virtual CPUs. This is the largest virtual server available today in the public cloud, and is a target for memory-intensive workloads such as SAP HANA.

Plans for the x1 were first revealed last October at the AWS Re:Invent conference where Amazon promised availability in the 1st half of 2016. With this month’s announcement, Amazon comfortably hit their target. The new x1.32xlarge instance has 8x more memory and 3x more vCPUs than any other AWS machines. The instance specifications are:

  • 128 vCPU via four Intel Xeon E7-8880 v3 processors
  • 1,952 GiB of DDR4-based RAM
  • 3.84 TB of local SSD storage, capable of up to 126,000 random-read IOPS
  • 10Gbps of network bandwidth

AWS says that these Intel processors support Turbo Boost 2.0, AVX 2.0, AES NI, and TSX NI, all of which improve performance for memory-intensive applications. The memory itself is capable of “300 GiB/s of sustainable memory-read bandwidth and 140 GiB/s of sustainable memory-write bandwidth.” Only the following Linux operating systems can handle the number of vCPUs present in the x1 instances: Amazon Linux 2016.03, Ubuntu 14.04, Red Hat Enterprise Linux 7.1, SUSE Enterprise Linux 12.

This new instance is ideal for running data processing engines like Apache Spark, high performance computing applications, or SAP systems, according to AWS.

The X1 instances have been certified by SAP for production workloads. They meet the performance bar for SAP OLAP and OLTP workloads backed by SAP HANA.
Many AWS customers are currently running HANA in scale-out fashion across multiple R3 instances. Many of these workloads can now be run on a single X1 instance. This configuration will be simpler to set up and less expensive to run.

The timing of the announcement was likely not a coincidence. At this month’s SAP Sapphire conference, Microsoft and SAP shared the details of a new partnership that included, in part, the ability to run SAP HANA on Azure virtual machines. While Microsoft promised that new virtual machines were on the way to support SAP HANA workloads, AWS was able to beat them to the punch with the x1. TechRepublic reports that SAP is aggressively courting partnerships in order to drive enterprise growth, and cloud providers seem all-too-willing to accommodate.

The x1 joins a dizzying array of cloud server options offered by AWS. There are now forty current-generation instance types, and fifty-five available instances types overall. Customers choose among general purpose, compute optimized, memory optimized, storage optimized, and GPU instance types. AWS now has the most instance types and largest virtual machine in the public cloud.

Cloud Provider Number of Instance Types Maximum RAM Available
AWS EC2 40 (current generation) / 55 (overall) 1,952 GB
Microsoft Azure 40 448 GB
Google Compute Engine 18 208 GB
Digital Ocean 9 64 GB
CenturyLink Cloud configurable 128 GB
IBM Softlayer configurable 64 GB
Joyent 25 110 GB
Rackspace 19 240 GB

The x1 starts at $13.338 per hour for an on-demand instance. Customers can get that down to $3.732 per hour with a three year commitment (“reserved instance”), and a $98,072 upfront payment. For the time being, customers must use an online form to request permission to create x1 instances. The servers can be deployed to AWS locations in the US, Europe, and Asia.


Rate this Article