On Thursday Amazon Web Services announced High I/O instances, a new EC2 instance designed for applications requiring low latency access to high-speed storage. The new instances are backed by solid-state disks (SSDs), and Amazon says they can deliver in excess of 120,000 read input and output operations (IOPS) and over 80,000 write IOPS, making them ideally suited for transaction processing, time series analysis, and mobile and streaming applications that require low latency access to storage systems that can deliver tens of thousands of IOPS. “With the launch of High I/O instances, customers can take advantage of SSD-based instances to run their most demanding applications on AWS, whether it’s running databases that support high-transaction enterprise applications or powering massively popular social, mobile or gaming apps for consumers,” said Peter De Santis, Vice President of Amazon EC2. "
The announcement prompted discussion in the cloud computing community. Here's a roundup of notable commentary and analysis from around the web:
AWS says disk is to cloud as tape is to disk - From The Register: "Amazon Web Services (AWS) says magnetic disks cannot handle modern NoSQL-powered applications, which have such high throughputs and generate such weirdly bursty traffic that spinning rust looks and feels as slow and awkward as tape."
High capacity Amazon update targets app bottlenecks - From Network World: "AWS makes weekly, if not daily, updates to its cloud offering but Paul Burns, an analyst at Neovise, says this is one of AWS's more significant updates. "There are all kinds of great use cases for something like this," he says. Applications are often bottlenecked by the their access to databases, he says, and having high capacity compute power, particularly with 2TB of SSD-backed disks, could help speed database processing for customers."
Benchmarking I/O with SSD for Cassandra on AWS - At the Netflix Tech Blog, Adrian Cockroft has a detailed post discussing their results with the new SSD instances: "With this announcement AWS makes it easy to provision extremely high I/O capacity with consistently low latency. AWS has been competitive in instance memory capacity for a long time and is leading the industry in CPU performance along with 10GBit networks. Now that extremely IO intensive applications can be deployed, a commonly cited obstacle to running in the cloud has been removed."
I/O Performance (no longer) Sucks in the Cloud - From James Hamilton of Amazon: "Yesterday an ideal EC2 instance type was announced. It’s the screamer needed by these workloads. The new EC2 High I/O Instance type is a born database machine. Whether you are running Relational or NoSQL, if the workload is I/O intense and difficult to cost effectively scale-out without bound, this instance type is the solution. It will deliver a booming 120,000 4k reads per second and between 10,000 and 85,000 4k random writes per second."
Expanding The Cloud – High Performance I/O Instances for Amazon EC2 - From Amazon CTO Werner Vogels: "Earlier this year I attended a panel on “Scaling to Infinity” with the top engineers from Netflix, Facebook, Tumblr, etc. In unison they proclaimed that in all of their systems the scaling bottleneck had been the database. These bottlenecks can often be attributed to constraints in the I/O system and the challenges of providing consistent I/O performance in systems that have not been designed for high performance I/O. The fast growing popularity of Amazon DynamoDB, which provides consistent read/write performance through an I/O provisioning interface, demonstrates that if the database can be configured such that it no longer is a bottleneck, applications can become much simple, and thus more reliable and scalable. It is my expectation that with the increase of data-centric applications we will see more and more I/O hungry systems being built that require this type of rock solid High Performance I/O that the hi1.4xlarge can give you. "