In partnership with
What happened this week:
Duck Guides to Save You Money with AWS SPs & RIs
AWS AI Cost Optimization Strategies.
Another week with Tons of AWS News, mostly Good 👼
Mention of Honour: FinOps Master Thesis
Job Offers
Professional Spotlight with my partner in FinOps
Dropping great stuff. Enjoy yourself:
Collaborate with FinOps Weekly
AWS Guides
A Duck’s Guide to RIs & SPs
Duckbill releases guides for AWS Reserved Instances & Savings Plans. Here’s a summary of both:
Savings Plans
AWS Savings Plans (SPs) offer discounts for committed hourly compute usage. SPs cover EC2, Lambda, and Fargate services.
SPs are financial tools, not capacity reservations. Think of SPs as committed use discounts.
You can stack multiple SPs. Balance commitments with growth needs. Evaluate quarterly for adjustments.
SPs cannot be resold. No market like Reserved Instances.
Compute and EC2 Instance SPs are two types. Compute SPs are flexible with easy management. EC2 Instance SPs require more specific choices.
Mixing Compute and EC2 Instance SPs can enhance discounts. Be cautious to avoid overspending. Communication is key in planning.
SPs apply with “deepest discount first.” Automatically covers workloads for maximum savings. Think in aggregate, not per instance.
Combine SPs and Reserved Instances for best results. SPs will apply once RIs expire.
Use AWS tools for SP analysis. Coverage and utilization reports guide decisions. Aim for high coverage and utilization.
Reserved Instances
RIs are a way to save money by committing to use certain AWS services. They aren’t reservations about capacity themselves.
They are available for services like EC2, Elasticache, OpenSearch, RDS, and Redshift.
EC2 RIs offer regional and zonal reservation options with specific configurations.
There are Standard and Convertible RIs, each with different flexibility and discount levels.
AWS provides tools to help report and manage your RI usage and coverage.
AWS
Default minimum object size for S3 Lifecycle transition rules
It’s one of the most commented news of the week on LinkedIn. AWS’s new default minimum is affecting one of the core mechanisms of cost optimization in storage, Lifecycle rules. Here’s how:
New Default Minimum Object Size: Amazon S3 will now apply a default minimum object size of 128 KB for S3 Lifecycle transition rules to optimize transition costs.
Customization Options: Users can override this default and set a custom minimum object size for their S3 Lifecycle transition rules.
Impact on Existing Rules: This change will not affect existing S3 Lifecycle rules but will apply to any new or modified configurations.
Deployment: The new default is being deployed across all AWS Regions and will be completed in the coming days.
Events
FinOps NEXUS Virtual Conference 2024
FinOps Nexus Conference brings together thought leaders, practitioners, and innovators from across the globe to explore the latest trends, best practices, and strategies in FinOps.
The conference will feature a series of interactive sessions, panel discussions, and deep dives into topics like cost optimization, cloud financial management, and cross-functional collaboration. Hear from industry experts and gain insights into the challenges and solutions that are shaping the future of cloud finance.
GenAI
AWS AI Cost Optimization Strategies
Are you interested in using generative AI but worried about costs? Here’s a guide to help you optimize expenses effectively.
Choose the Right Model
Define your use case clearly.
Balance performance, accuracy, and cost.
Use smaller models for specific tasks.
Fine-Tuning and Customization
Use pre-trained models and add your own data.
Consider retrieval-augmented generation (RAG) for cost-effective customization.
Fine-tune only when necessary for high accuracy.
Focus on Good Data Management
Quality data is more important than quantity.
Clean data improves model efficiency and accuracy.
Invest in data organization and governance.
Implement Cost-Saving Practices
Foster a culture aware of AI costs.
Use FinOps for better resource management.
Encourage teams to optimize and own their AI costs.
Cloud Architecture
Optimize Distributed Systems: Reduce Latency & Costs
In distributed systems, keeping costs low and performance high can be tricky. Here are some tips to help:
Use Zone Aware Routing: Direct traffic within the same zone. Save on data transfer costs.
Adopt Advanced Tools: Use Istio or Kubernetes for zone awareness. Tools help manage cross-zone data transfer.
Choose the Right Databases: Some DBs can read data locally to save costs. Consider Kafka or Redis for zone-friendly routing.
Be Mindful of Cluster Hotspots: Uneven data can overload specific zones. Balance and scale services efficiently.
Optimizing distributed systems is vital for performance and cost.
AWS
Estimate your AWS Network Costs with a New Tool
AWS Networking Costs Calculator helps you estimate AWS networking service costs.
It’s open-source, self-hosted, and web-based.
How It Works: Choose your AWS Region and services.
Enter details like data transfer and endpoints.
Cost Visualization: Displays diagrams for traffic flow. Shows estimated charges for each service.
Tool Architecture: Combines a serverless backend and a ReactJS frontend.
Uses services like AWS Lambda and DynamoDB.
Getting Started: Prerequisites include Linux OS, NodeJS, and AWS CDK.
Using It: Access via a web URL provided during setup. Adjust inputs to see cost projections
Limitations and Considerations: Cost estimates are approximate. Not all AWS services are included yet.
Future Plans: Continual updates and more services are planned.
The calculator simplifies understanding of AWS networking costs.
Mention of Honour presented by FinOpsly
End-to-End FinOps.
Data at the speed of thought.
FinOps Empowered by AI.
FinOpsly’s AI-powered platform helps optimize your cloud cost and performance by understanding the unique complexities of your cloud infrastructure.
🎖️ Mention of Honour
FinOps Master Thesis
Carlos made a complete Master’s thesis on a FinOps case, and it’s publicly available!
Using FinOps to Efficiently Manage Infrastructure and Operational Costs: A Case Study of Massive Data.
Quoting the author’s abstract:
This study examines the intersection of FinOps with AI deployments in the cloud, focusing on the cost optimization challenges and strategies. Utilizing a mixed-methods approach, the research integrates quantitative data analysis of performance metrics with qualitative case studies from multiple organizations.
The study contributes to the existing literature by providing empirical evidence of the benefits of cloud computing and offers a framework for organizations to adopt these practices effectively.
A great read, and deeply recommended for anyone interested in FinOps!
🏆 Quick Wins
Announcements
Amazon EC2 X8g instances Release + Introducing Amazon EC2 C8g and M8g Instances
Karpenter & AWS CUDOS included in ThoughtWorks Technology Radar
Articles & Podcasts
What’s new in Cloud FinOps? Interview with Udam Dewaraja – The story of FOCUS
Handbook to AWS Cost Optimization
Inventory of Azure Resource Tagging with Graph Explorer and KQL
Professional Spotlight
Damian Munafo
My New Partner in FinOps
As some of you may know, we launched a Podcast for FinOps in Spanish together with Damian. A guy who is extremely easy to work with, and we’ll launch several things together. Stay tuned and check his content!
👩🏭 FinOps Job Offers
Permanent links (i.e. FinOps Job site) and FinOps Job Offers online.
💬 FinOps Connect Community
Join the FinOps Connect Slack Channel and let’s grow the community!
🗳️ Poll of the Week
That’s all for this week. See you next Sunday!
Let’s Collaborate!
Connect on LinkedIn: I share valuable free content there. I am open to chatting with anybody interested in Cloud & FinOps.
Sponsor this newsletter: Promote your company in this newsletter and reach the Cloud audience that wants to maximize business value and reduce costs.
Collaborate with FinOps Weekly: Our brand expands to more than just the newsletters. Podcasts, Posts, Webinars and any collaboration related to FinOps is available.
P.S.1: Know someone who’ll enjoy this newsletter? Please forward it. You’ll make them Stay Ahead in FinOps.
P.S.2: Was this email forwarded to you? Subscribe and get all editions!
P.S.3: I’ll do my best to provide you with the best content. However, some of it may slip. If you think something should have been included, please REPLY to this email using the link above.