Cloud Auto-Scaling

Cloud Autoscaling Allows Businesses To Rapidly Scale With Demand

Posted on February 3rd, 2014 by William Hayles

One of the major benefits of building a cloud-based infrastructure is the ability to scale quickly. There are three basic scaling scenarios that businesses face.

  • Scaling Out As A Business Grows — Because deploying new compute and storage resources is almost trivially easy on IaaS platforms, businesses can be confident that as their needs grow, their infrastructure will always be able to grow to meet new demands.
  • Scaling Down As Requirements Change — With physical hardware this is much more difficult. Businesses often find themselves in a position where they’ve invested heavily in servers and other hardware that becomes surplus to requirements. Idle hardware is money wasted. Cloud platforms enable businesses to smoothly reduce their deployments, making more efficient use of their budgets.
  • Reactive Scaling As Resource Demands Change — Many businesses have highly variable demand for resources. The paradigmatic example is Netflix, who need to be able to massively scale up at times of peak demand and scale down when fewer users are streaming video from their services. If they were forced to use physical hardware deployments, then Netflix would have a huge number of servers sitting idle for much of the time, significantly driving up their costs.

Netflix is an extreme example of the need for reactive scaling, but the same principle applies to many businesses. The need to scale rapidly in both directions over short timescales is a difficult problem, but it’s a problem that is much more soluble with IaaS infrastructure deployment.

The rapid scaling capabilities of IaaS platforms have many different use cases, including:

  • - Deployment of development and testing environment
  • - Software rollouts
  • - Traffic spikes
  • - Variable demand for compute resources in data processing

On its own, the ability to scale quickly is a strong driver of enterprise cloud adoption, but real power lies in the programmatic potential of IaaS. Because cloud deployments are basically software-driven and abstracted from the physical hardware, almost every aspect of a cloud platform can be exposed via an API. The API gives businesses the ability to develop applications and scripts to tailor their cloud’s behavior, including its scaling behavior.

Automated scaling allows a company’s cloud to expand or contract as demand changes. Companies can leverage metrics to track usage patterns within their cloud and, via the API, dynamically adjust their resource deployments.

Being able to autoscale in accordance with chosen metrics can have a powerful influence on the cost and efficiency of infrastructure deployment. Obviously, if deployment can track usage in real time, the risk of having idle infrastructure is significantly reduced, and because on cloud platforms users only pay for resources they use, costs can more accurately track business operations than is possible with physical hardware deployments. The flip side of that coin is that without sufficiently speedy infrastructure deployment, business opportunities can be lost because of performance degradation and poor availability. Being able to automatically scale up when a traffic spike hits or user demand increases unexpectedly is equally important.

In the modern business world, it’s crucial that companies are able to modify their infrastructure deployments in line with demand. Automatic scaling and the application programming interfaces that enable it are a killer feature for enterprise cloud users.

First Class Support. EC2 Compatibility.

Learn more about our state-of-the-art Cloud platform.
William Hayles

Author William Hayles

About William - Will Hayles is a technical writer and blogger for Outscale, a leading cloud hosting provider in the USA and France.

Share this post