Rootconf 2017

On service reliability

Ganesh Hegde


Cost Optimizations on AWS: Spend where it really matters

Submitted Jan 27, 2017

Modern startups need to focus on infrastructure optimization without compromising on performance, availability and redundancy to make sure they are efficiently capitalizing on their limited budgets. This talk aims to provide a brief rundown on this based on AWS infrastructure.



With infrastructure budgets being a critical component of modern web companies and most of the firms going cloud completely or partially, cost optimization with minimal production impact gains more importance than ever before.
The talk will explore:

  • Some less explored paths for cost savings on AWS without spending additionally
  • Case studies on problems faced when engineering for lower costs

Cost On Optimization on AWS : A practical guide

  • Cost is an afterthought. Shouldn’t be
  • Cloud cost overruns - 5 symptoms
  • Know your enemies
    • Know where exactly are you burning money
    • Using AWS Billing Dashboard to breakdown resource usage
    • Limitations of AWS instance types
  • Docker! Docker! Docker!
    • Impact of Dockerization on your budget
    • Using Docker to optimize resource usage of your hosts
  • Shift to Cheaper regions!
    • Deciding between Latency vs Cost for cheaper regions
    • Headaches to avoid when handling multiple regions
    • Benefits of using a denser region
  • Can you use Spot Fleet?
    • Deciding payloads that can be on Spot Fleet
    • Spot Fleet’s and their limitations
  • Leveraging your CDN to reduce your infra
    • Caching API responses to reduce your infrastructure load
    • Serving assets from S3 wherever possible
    • Caveats
  • Do you really need third party providers?
    • Deciding the right answer for third party cost optimization services
  • Test setup shouldn’t be burning a hole in your pocket
    • Do you really need HA in your test bed 24/7
    • “Production Parity in Test” Greatest fallacies of our generation
  • Reserved Instances Strategies
    • If all fails….

Case Studies

  • Leverage AWS Lambda
    • Use AWS Lambda to intelligently handle compute resources for random, asymmetric workloads using cloudwatch, SQS and boto3
  • Handling resources on non business hours
    • Save on your resources without tampering your developer’s peace of mind using Python, boto3 and Google Sheets API


Intermediate to Advanced knowledge of AWS or equivalent concepts in GCP/Azure
Basic to Intermediate knowledge of Linux
Intermediate knowledge of SRE/DevOps concepts

Speaker bio

Senior Software Engineer at Urban Ladder



{{ gettext('Login to leave a comment') }}

{{ gettext('Post a comment…') }}
{{ gettext('New comment') }}
{{ formTitle }}

{{ errorMsg }}

{{ gettext('No comments posted yet') }}

Hosted by

We care about site reliability, cloud costs, security and data privacy