Service-level agreement (SLA)-based computing is a futuristic and modernized concept for IT resource cloud providers....
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
Whether it's for an instance, an underlying infrastructure resource or software as a service resources, an SLA provides a level of flexibility unlike anything else in the cloud. If users are willing to compromise on service quality and SLAs, they will be able to enjoy lower cost service tiers.
A great example of this is Amazon Web Services (AWS) EC2 Spot Instances. While AWS users can save up to 80% off On-Demand Instance prices, they compromise on resource availability. Spot Instances provide no continuity guarantee; Amazon takes back the allocated spare capacity for Spot Instances to fulfill the AWS SLA commitment to On-Demand and Reserved Instance users. Administrators use Spot Instances to run non-mission-critical environments, such as test and development, and when there is a need for a great pool of resources, such as analytics or high-performance computing.
So why doesn't AWS extend its SLA-based resource catalog to provide users with the option to purchase a resource with an uptime of 80%? It could be because AWS prefers to simplify its business model. In addition, doing so might require significant changes to how cloud vendor billing works. Billing currently is based on the size and the number of hours a resource uses. An AWS SLA can complicate things, especially if customers don't require it.
If you look carefully, however, you can see that the seeds for SLA-based computing have been planted and development is in the works. For example, Google Cloud Platform recently introduced its version of Spot Instances, which are referred to as Preemptible Instances. In addition, Amazon acquired ClusterK, a startup that helps AWS users wrap a single Spot Instance with a better SLA, using On-Demand and Reserved Instances. New AWS capabilities, including Spot Fleets and alerts that inform you before a spot instance terminates, are good examples of next steps for public cloud.
Wrangle AWS Spot Instances with Spot Fleets and save
AWS cloud cost management: The buck stops with you
Dig Deeper on AWS pricing, cost and ROI
Related Q&A from Ofir Nachmani
Security should be a top priority in public cloud deployment, but breaches are still possible. If your AWS account has been hacked, follow these ...continue reading
AWS users in the US-East-1 region have seen a string of outages in the last few years. But does that mean they shouldn't deploy workloads there?continue reading
Reserved Instances save IT teams money on compute resources but lock them into a particular type of instance. How do Convertible RIs differ, and what...continue reading
Have a question for an expert?
Please add a title for your question
Get answers from a TechTarget expert on whatever's puzzling you.