Amazon sets sights on cloud cost sprawl


Amazon(s amzn) is making it easier for companies to track and price out the cloud services they’re deploying with a new cost allocation process.

The fact that Amazon Web Services are so inexpensive and easy to spin up is both a blessing and a curse for companies. A blessing because internal developers can try out new stuff fast and cheap; a curse because it leads to cloud sprawl where companies find it difficult to track and monitor cloud usage and the costs of which — let’s face it — add up. Even cheap services cost money. A post on the Amazon Web Services blog outlines how corporate users can tag those services to make billing less of an, um, adventure.

It does still require some manual input. The Amazon resources you use have to be tagged  as one of the following:

  • S3 buckets
  • EC2 Instances
  • EBS volumes
  • Reserved Instances
  • Spot Instance requests
  • VPN connections
  • Amazon RDS DB Instances
  • AWS CloudFormation Stacks

The tags — 10 of which can be applied for each resource — are then entered via the AWS Management Console, the service APIs, the command line or Auto Scaling.

For many big AWS users, this is a good thing. Because there are so many services offered by the hour, it is easy to leave something running unnecessarily. “For us, the underlying mechanism is very valuable because it means we can do a ‘one-stop-shopping’ for our clients and re-bill them for a combination of the AWS charges plus our support and engineering charges,” said Bob Shear, president of Greystone Solutions, a Boston company that sets up ecommerce sites for customers.

A variety of third parties including Rightscale, Cloudability, CloudynNewvem and UptimeCloud already offer services to help customers better track their cloud spending. And Rightscale just announced its PlanforCloud service for forecasting purposes.

Amazon clearly sees a need to offer a better cost tracking service of its own.

Feature photo courtesy of Flickr user Will Merydith




It would be very helpful to your non-insider readership, to explain (briefly) what exactly each of those 8 or 10 Amazon resources actually *do*.

The labels are pretty opaque for any other than the IT Illuminati.


A quick overview of the AWS services:

S3 buckets – Redundant data storage infrastructure
EC2 Instances – Computing capacity
EBS volumes – Block level storage volumes
Reserved Instances – Upfront payments for EC2
Spot Instance requests – Bid on unused Amazon EC2 capacity
VPN connections – Hardware Virtual Private Network
Amazon RDS DB Instances – Relational Database Service
AWS CloudFormation Stacks – AWS management framework for AWS resources.

Check out for more info regarding reserve instances, RDS, EBS, EC2 and S3 control and optimization

Comments are closed.