Rebound effect in the public cloud? Are you ready to assume your responsibility?

In 1865, the English economist?William Stanley Jevons?observed that technological improvements that increased the efficiency of coal use led to the increased consumption of coal in a wide range of industries. This so-called "Jevons Paradox" is also commonly known as "rebound effect" when the rebound is greater than the initial improvement.

In?conservation?and?energy economics, the?rebound effect is the reduction in expected gains from new technologies that increase the efficiency of?resource?use, because of behavioral or other systemic responses

Let's focus today on the behavioral response. "Behavior: the way that someone or something?behaves?in a?particular?situation." (Cambridge Dictionary)

The "particular situation" here is affordable, instant and almost unlimited access to the most efficient public cloud services.

How you behave in front of this "dream candy shop" situation will influence the net environmental impact of your AWS use.

How AWS have been helping you identify unnecessary use of its services?

__________________________________________________________________

On the 30th of January 2012, AWS Trusted Advisor was announced, automatically flagging to you unused resources like Elastic IP Addresses that are not attached to an Amazon EC2 instance (https://aws.amazon.com/blogs/aws/new-aws-premium-support-features-third-party-software-support-and-aws-trusted-advisor/).

Adding new resources efficiency automatic flags over the years, AWS Trusted Advisor today assists you today with identifying:

Low Utilization Amazon EC2 Instances, Amazon EBS over-provisioned volumes, Amazon RDS idle DB instances, AWS Lambda over-provisioned functions for memory size, idle Load Balancers, unassociated Elastic IP Addresses, underutilized Amazon EBS Volumes, underutilized Amazon Redshift Clusters, Amazon Comprehend underutilized endpoints and Amazon EC2 instances over-provisioned for Microsoft SQL Server.

__________________________________________________________________

On the 2nd of October 2015, Jeff Bar, Chief Evangelist for AWS, announced (using a reference to Jimmy Hendrix!) the availability of the AWS Well-Architected Framework whitepaper (https://aws.amazon.com/blogs/aws/are-you-well-architected/).

For each pillar, the whitepaper puts forth a series of design principles, and then defines the pillar in detail. Then it outlines a set of best practices for the pillar and proffers a set of questions that will help you to understand where you are with respect to the best practices.

From as early as 2015, the Cost Optimization pillar focused on "the ability to avoid or eliminate unneeded cost or suboptimal resources". Cost being the result of usage x unit cost, all Cost Optimization best practices focusing on usage optimization contribute to reducing the environmental impact of your AWS use.

Looking at this pillar today (https://docs.aws.amazon.com/pdfs/wellarchitected/latest/cost-optimization-pillar/wellarchitected-cost-optimization-pillar.pdf), the 46 best practices help you to only use the right amount of AWS resources needed, but also implement the right organization mechanisms to do it consistently at scale.

__________________________________________________________________

Fast forward to December 2021 at the annual AWS re:Invent conference, the 6th pillar of the Well Architected Framework, "Sustainability", was added.

I routinely invest a disproportionate amount of my professional and personal time to read and study research papers, industry analysis, thought leadership articles, specialized blog posts on IT environmental impacts.

It stills strike me today, almost 1 year after the launch of this Sustainability Pillar, that it does not get the exposure it deserves.?

For the sake of raising your interest and motivating you to have a go and read it (then apply it in your organization!), here are a few excerpts, that might nudge the way you think about lowering your environmental impact from your AWS use (from the 118 items implementation guidance detailed in this document):

"Redefine SLAs to meet business requirements, not exceed them."

"Make trade-offs that significantly reduce sustainability impacts in exchange for acceptable decreases in service levels."

"Evaluate the impact of processes and systems on your device lifecycle, and select solutions that minimize the requirement for device replacement while satisfying business requirements."

"Implement remote management for devices to reduce required business travel."

__________________________________________________________________

Looking at more recent materials, another insightful publication contributing to AWS' effort to help your limiting potential rebound effect is the "Guidance for Optimizing Deep Learning Workloads for Sustainability on AWS" (https://aws.amazon.com/solutions/guidance/optimizing-deep-learning-workloads-for-sustainability-on-aws/). Here are a few excerpts also:

"Try to find simplified versions of algorithms. This will help you use less resources to achieve a similar outcome. For example, DistilBERT, a distilled version of BERT, has 40% fewer parameters, runs 60% faster, and preserves 97% of BERT’s performance."

"Adopt sustainable tuning job strategy. Prefer Bayesian search over random search (and avoid grid search). Bayesian search makes intelligent guesses about the next set of parameters to pick based on the prior set of trials. It typically requires ten times fewer jobs than random search, and therefore ten times less compute resources, to find the best hyperparameters."

"If your users can tolerate some latency, deploy your model on asynchronous endpoints to reduce resources that are idle between tasks and minimize the impact of load spikes."

?

__________________________________________________________________

YES, Environmental sustainability is a shared responsibility between customers and AWS. AWS customers are responsible for Sustainability IN the AWS Cloud - optimizing workloads and resource utilization, and minimizing the total resources required to be deployed for your workloads. (https://docs.aws.amazon.com/wellarchitected/latest/sustainability-pillar/the-shared-responsibility-model.html).

But NO, you are not left alone to assume this responsibility: AWS have been investing continuously for more than 10 years to provide you with guidance, tooling, resources and training enabling your informed decisions.

The opinions expressed in this publication are those of the author. They do not purport to reflect the opinions or views of AWS.