Understanding the bill…
Managing consumption costs is one of the new challenges brought about by cloud adoption. In the past, infrastructure costs were mostly fixed. Organisations bought or leased infrastructure assets and the costs associated with them were relatively static after that point. There was some variability in things like power consumption and data usage, but otherwise the lifetime cost of an asset was relatively predictable and easy to model.
With cloud, cost is tied to consumption and utilisation, the more you use, the more it costs, the less you use, the less it costs. This drives the transformation of infrastructure to take advantage of elasticity and just in time resource provisioning, but this drive to transformation can make the cost of cloud consumption difficult to predict. This is further compounded by the fact that most cloud resources are billed individually, and they are all metered against different consumption units.
For example, modelling the cost of a cloud hosted VM will require input from number of resource meters:
- The hourly cost of the VM based on its CPU and RAM allocation
- The per GB cost of the storage it consumes
- The IOPS cost of the storage IO reserved for it
- The per IO cost of the storage transactions it consumes
- The per GB cost of the network throughput it generates
- The per IP cost of the public IP(s) it consumes
- The per lookup cost of the DNS load balancer
- The per GB cost of any logging you store to the cloud platform
The cloud providers themselves offer detailed output of what resources consumed what meter and what rate over a billing cycle, but understanding that data, and marrying it back to servers, applications and business units becomes a very labour intensive operation. The more complex the deployment, the more difficult these costs become to model.
Addressing these problems of understanding and managing consumption costs needs to be approached from multiple angles. At a platform level, resource tagging offers a partial solution. Cloud platforms allow you to apply key value pairs, or tags, to resources, which show up on billing reports. These tags can be used to logically group multiple resources together to provide a consolidated billing view against a tag. Cloud providers also recommend that cloud environments are deployed under high level billing constructs that align to business units. However, these deployment and management strategies aren’t possible in all scenarios, and they are still far from ideal solution. They require considerable manual handling and processing to turn tag sorted billing data in human readable reports.
Enter Cloud Ctrl
Cloud Ctrl enables organisations to monitor and manage their cloud hosted services across multiple providers. It enables businesses who have stepped into the cloud to build and manage successful cloud implementations.
Cloud Ctrl enables smart resource monitoring. It provides a single dashboard from which all services costs can be viewed. From the monitoring dashboard you can review spend by time, by resource, by tag, or by type.
Cloud providers offer incentives in the form of rate discounts for making longer term commitments to consume resources. However, understanding how consumption is tracking against commitment is often difficult to predict using the provider native tools. Cloud Ctrl uses its spend analysis engine to project daily, monthly and yearly cost projections so the current run cost of a cloud environment can be accurately compared to a spend commitment to determine whether a deployment is tracking ahead of spend schedule or behind. Having this level of visibility can help focus development goals into finding further optimisations to ensure budgets don’t blow out of control, or developing new features to ensure that committed funds don’t to go waste.
You can also use Cloud Ctrl to build resource templates and then compare the deployment cost of that template across different regions, or even different cloud providers. With the increasing focus on finding efficiency within cloud deployments and spreading failure domains across multiple cloud providers, being able to manage and understand the costs of a cloud deployment from a single place where providers and locations can be compared and analysed becomes critical.
Understanding cloud consumption is no longer a monthly battle between spreadsheets and business units when Cloud Ctrl is used to report on and analyse a cloud deployment. With the consumption costs well understood thanks to Cloud Ctrl, focus can be turned back onto improving the deployment itself rather than just understanding it.
In the next blog post we will look at methods of tuning a cloud deployment to deliver more efficient cost consumption.