Category Archives: Amazon Elastic Compute Cloud

Contrarian: Building, Colocating Your Own Servers, No Cloud Involved

Jeff Atwood has a great post at Coding Horror talking about his penchant for building his own servers to rack at a colo. It tries to compare to the Amazon Ec2 alternative, all the while admitting it’s pretty much apples and oranges.

I want to make it clear that building and colocating your own servers isn’t (always) crazy, it isn’t scary, heck, it isn’t even particularly hard. In some situations it can make sense to build and rack your own servers, provided …

  • you want absolute top of the line server performance without paying thousands of dollars per month for the privilege
  • you are willing to invest the time in building, racking, and configuring your servers
  • you have the capital to invest up front
  • you desire total control over the hardware
  • you aren’t worried about the flexibility of quickly provisioning new servers to handle unanticipated load
  • you don’t need the redundancy, geographical backup, and flexibility that comes with cloud virtualization

It’s worth reading in its own right, but also because he does a pretty good job of outlining the pros and cons of cloud versus self-hosting. It’s also good thing to remember that no matter how “virtual” we get there’s still gotta be a bunch of hardware somewhere to make it all go.


Amazon Web Services Launches High Performance Storage Option for Amazon Elastic Block Store

Image representing Amazon as depicted in Crunc...

Amazon Web Services today announced new features for customers looking to run high performance databases in the cloud with the launch of Amazon Elastic Block Store (Amazon EBS) Provisioned IOPS. Provisioned IOPS (input/output operations per second) are a new EBS volume type designed to deliver predictable, high performance for I/O intensive workloads, such as database applications, that rely on consistent and fast response times. With Provisioned IOPS, customers can flexibly specify both volume size and volume performance, and Amazon EBS will consistently deliver the desired performance over the lifetime of the volume. To get started with Amazon EBS, visit http://aws.amazon.com/ebs.

Provisioned IOPS volumes are engineered to allow customers to develop, test, and deploy production applications and be confident that they will receive their desired performance. With a few clicks in the AWS Management Console, customers can create an EBS volume provisioned with the storage and IOPS they need and attach it to their Amazon EC2 instance. Amazon EBS currently supports up to 1,000 IOPS per Provisioned IOPS volume, with plans to deliver higher limits soon. Customers can attach multiple Amazon EBS volumes to an Amazon EC2 instance and stripe across them to deliver thousands of IOPS to their application.

To enable Amazon EC2 instances to fully utilize the IOPS provisioned on an EBS volume, Amazon EC2 is introducing the ability to launch selected Amazon EC2 instance types as EBS-Optimized instances. EBS-Optimized instances deliver dedicated throughput between Amazon EC2 and Amazon EBS, with options between 500 Megabits per second and 1,000 Megabits per second depending on the instance type used. The combination of EBS Provisioned IOPS and EBS-Optimized instances allows customers to run their most performance-sensitive applications on Amazon EC2, giving them predictable scaling with the same ease of use, durability, and flexibility of provisioning benefits they expect from Amazon EC2 and Amazon EBS.

“AWS introduced Amazon EBS in 2008 to provide a highly scalable virtual storage service and now, four years later, our customers are running applications on Amazon EC2 using EBS volumes at tremendous scale,” said Peter De Santis, Vice President of Amazon EC2. “Customers have been asking for the ability to set their performance rate to achieve consistently high performance. With EBS Provisioned IOPS volumes, EBS-Optimized instances and the recently launched High I/O SSD-based EC2 instances, customers have a range of choices for running their most demanding applications and databases on AWS while achieving peak performance in a predictable manner.”

At NASA’s Jet Propulsion Laboratory, Amazon EBS is used to support various missions and research programs. Consistent performance of I/O is a major requirement for numerous use cases across NASA ranging from scientific computing to large scale database deployments. JPL now routinely provisions cloud compute capacity in an elastic manner but database latencies have proven difficult. To help meet this challenge, JPL’s missions and its Office of the CIO prototyped the new EBS Provisioned IOPS capability to provision flexible compute capacity and overcome database latency restrictions. The results were highly successful and the release of EBS Provisioned IOPS, coupled with Amazon EC2 High I/O SSD-based instances, will introduce a whole new realm of I/O intensive scientific applications for JPL from radar data processing to the quest of black holes.

Stratalux is a leader in building and managing tailored cloud solutions for customers of all sizes. “A common request we see from both our large and small customers is the need to support high performance database applications. Throughput consistency is critical for these workloads,” said Jeremy Przygode, CEO at Stratalux. “Based on positive results in our early testing, the combination of EBS Provisioned IOPS and EBS-Optimized instances will enable our customers to consistently scale their database applications to thousands of IOPS, enabling us to increase the number of I/O intensive workloads we support.”

Amazon EBS Provisioned IOPS volumes are currently available in the US-East (N. Virginia), US-West (N. California), US-West (Oregon), EU-West (Ireland), Asia Pacific (Singapore), and Asia Pacific (Japan) regions with additional Region launches coming soon.


Google Compute Engine and Cloud Video Transcoding — How Does it Compare?

Zencoder performed some initial comparison tests of Google Compute Engine versus Amazon EC2 for transfering and transcoding video.

“On EC2, we use Cluster Compute instances, which are fast dual-CPU machines in two sizes: 4XL and 8XL. We compared these with the fastest GCE instance type, which is currently a single-CPU 8-core server.”

Here’s one of their resulting charts:

Read the full post for details.