alphaspirit - Fotolia
In the midst of launching new services and features at re:Invent in October, AWS chief technology officer Werner Vogels announced two new instance types during his keynote: t2.nano and X1.
Both Elastic Compute Cloud (EC2) instance types satisfy different usage trends, leading to approval from the Amazon Web Services (AWS) community. X1 instances can be used when there's a need for large in-memory workloads or real-time analytics; the t2.nano can be used to run modest loads on websites or microservices. In terms of computing resources, each EC2 instance type is quite different. The X1 is loaded with a large amount of memory, while t2.nanos offer bursting capabilities and processing power.
Go big with X1
AWS X1 instances come with up to 2 TB of memory and can be powered by up to four Intel Xeon Processor E7-8880 v3 (Haswell) processors; they support high-memory bandwidth and large L3 caches. X1 instances have more than 100 vCPUs, which allow them to handle large, in-memory workloads and concurrent connections. Prior to the debut of the X1 EC2 instance type, the maximum available memory instances were r3.8xlarge, i2.8xlarge and d2.8xlarge, all of which only support 244 GB of memory.
X1 instances are specific to mission-critical enterprise workloads that require high performance and a lot of in-memory processing. These types of workloads could include SAP HANA, real-time analytics, Apache Spark, Presto and Microsoft SQL Server, in addition to applications that generate enormous amounts of cache.
"The X1 is the first real use industry-wide of the Xeon e7 microprocessor in an infrastructure as a service offering," said Diane Bryant, the senior vice president and general manager of Intel's Data Center Group.
X1 instances will be available during the first half of 2016, with pricing details to follow.
As Bryant stated, no other public cloud providers offer instances that match the X1. The maximum memory available with Google Compute Engine instances is 208 GB for the n1-highmem-32, which offers 32 vCPU. Microsoft Azure offers G-series machines, which have 32 vCPU and 448 GB of memory and run on Intel Xeon E5 V3 family processors.
Burst up with the t2.nano
In 2014, AWS launched the T2 family of instances, low-cost VMs that provide assured baseline performance with the capability to burst up to a full core if high compute power is required. The burst duration depends on the CPU credits earned during idle usage periods. The T2 family includes the micro (t2.micro), small (t2.small) and medium (t2.medium) instance types. Earlier this year, AWS launched the t2.large instance.
During this year's re:Invent, AWS launched another EC2 instance type in same family: t2.nano with 1vCPU and 512 MB of memory. The t2.nano can run at full core performance for over an hour on a full credit balance, according to the AWS blog. Credits for the new instance expire within 24 hours of launching them.
T2.nano instances seem like a great fit for dynamic websites that require short bursts of processing power, microservices or monitoring systems. T2.nano instances will be available for general use later this year; pricing will also be released around that time.
From a competitor's perspective, Google Compute Engine offers f1.micro instances with 1vCPU and 0.60 GB of memory, and they work on the same bursting concept as AWS. That is, a customer is allowed to use additional physical CPUs for short periods of time. Another AWS competitor, Microsoft Azure offers the A0 VM with 1 CPU core and 768 MB of memory.
Get to know all the AWS EC2 instances
Spot fleets lead to cost savings
AWS adds T2 large instance type