AWS Public Sector Blog
A guide to reducing waste and improving efficiency with AWS
The U.S. General Services Administration (GSA) recently announced an innovative OneGov agreement with Amazon Web Services (AWS) to provide up to $1 billion in savings for cloud adoption, modernization, and training for federal agencies. In alignment with this announcement are solutions to reduce waste and improve efficiency.
To help organizations accelerate these initiatives and reduce fraud, waste, and abuse (FWA), this guide outlines practical strategies for optimizing technology spend while improving operational efficiency. We will present a three-tiered approach to optimization:
- Agency-wide digital transformation initiatives
- AWS environment refinements
- Service-level improvements
Each approach includes resources and actionable steps for organizations to start their optimization journey today.
Agency-wide digital transformation initiatives
Large scale IT transformation using OneGov incentives starts with migration and modernization through cloud adoption. In addition to benefiting from the OneGov agreement, The Hackett Group’s study reveals that organizations migrating to AWS achieve an average 20 percent reduction in infrastructure costs, with top performers realizing 47 percent savings. Beyond cost efficiency, agencies experience 66 percent increased infrastructure productivity and redirect 29 percent more time to innovation.
Whether you choose to completely exit data centers or gradually migrate applications over time, a critical first step is collecting application portfolio data. This data must be evaluated against the seven common migration strategies known as the 7 Rs: refactor, replatform, repurchase, rehost, relocate, retain, and retire. Following these strategies, agencies can define the best migration path for each application.
For certain workloads built on commercially licensed technologies, you can also consider the complementary AWS Optimization and Licensing Assessment (AWS OLA) program, which offers up to 60 percent greater license efficiency by pinpointing underused licenses and suggesting more cost-effective alternatives.
Legacy systems consume disproportionate IT budget. With AWS Transform, .NET applications can be modernized up to four times faster while reducing operating costs up to 40 percent through eliminated Windows licensing. VMware migrations eliminate costly third-party licensing and accelerate network conversion by up to 80 times. With the power of agentic AI, mainframe modernization timelines are cut from years to months, reducing ongoing mainframe costs and preserving critical business logic.
Federal agencies can streamline contact center operations and eliminate waste by consolidating their multiple contact center platforms to a unified, cloud-based platform. Amazon Connect offers a pay-as-you-go model and requires no minimum commitments. Built-in AI capabilities create a powerful efficiency flywheel, enabling caller self-service, assisting agents with real-time support and automated summaries, and deploying AI agents for routine tasks. By adopting this modern approach, agencies benefit from a continuous cost-reduction cycle because improved efficiency naturally leads to lower operating expenses. To learn how customers have reduced call volumes 60 percent and reduced agent training time 50 percent, visit Unleash AI to transform every customer interaction.
Agencies are increasingly seeking improvements in staff productivity. Amazon Q Business makes generative AI securely accessible to everyone across the organization. Using your own content, data, and systems, Amazon Q Business makes it easier for users to get fast, relevant answers to pressing questions, solve problems, or generate content. Amazon Q Developer can provide up to 40 percent increase in developer productivity, helping them to write code faster and more securely.
AWS environment refinements
The best way to optimize wasteful spending is to avoid unnecessary costs in the first place. AWS recommends using security best practices in AWS Identity and Access management (IAM) when setting up new users in your AWS environment. This includes regularly reviewing and removing unused privileges, including those privileges that allow resource creation by users. Customers can further refine this by using AWS Service Catalog, providing end users with a portal in which to discover and launch products that comply with organizational policies and budget constraints.
AWS offers a set of solutions to help you with cost management and optimization. This includes services, tools, and resources to organize and track cost and usage data, enhance control through consolidated billing, enable better planning through budgeting and forecasts, and further lower cost with resources and pricing optimizations.
Resources such as backups incur immediate costs upon creation. AWS Backup eliminates the costs of managing backup software and infrastructure and offers powerful cost-saving features through automated lifecycle policies. Customers can reduce expenses by transitioning infrequently accessed backups to cold storage and automatically deleting expired backups. Regular review of backup plans helps you use storage efficiently and maintain compliance and data protection.
You can use Amazon SageMaker Unified Studio, a single data and AI development environment, to find and access data across your organization and take action using the best tools for any use case. These actions range from data analytics to building generative AI applications with Amazon Bedrock.
In situations where you prefer to stay with your existing tools, you can significantly reduce data infrastructure costs by implementing a data mesh architecture that uses services such as AWS Lake Formation for decentralized data governance, eliminating redundant data silos across departments.
Service-level cost optimization
Organizations are adopting serverless architectures to drive down the total cost of ownership. The serverless pay-for-value model eliminates over-provisioning waste, and it’s ideal for workloads with variable demand patterns like sporadic usage or seasonal spikes. The AWS Well-Architected Framework for serverless applications provides best practices for cost optimization.
For applications that rely on virtual machines, AWS recommends considering AWS Graviton processors, which provide up to 40 percent better price performance than comparable Amazon Elastic Compute Cloud (Amazon EC2) instances not based on Graviton. The AWS Graviton Savings Dashboard is a visualization tool that helps you evaluate and find the potential savings of moving workloads to AWS Graviton. When using EC2 instances, consider AWS Auto Scaling, which monitors your applications and automatically adjusts capacity to maintain steady, predictable performance at the lowest possible cost.
By migrating from commercial databases to Amazon Aurora, you can achieve the performance and availability of commercial-grade databases at down to one-tenth the cost. AWS Database Migration Service (AWS DMS) helps a secure and seamless transition to these cost-effective solutions.
For both databases as well as compute instances, upgrading your block storage from Amazon EBS gp2 to gp3 delivers better performance at lower costs, saving up to 20 percent compared to older gp2 volumes.
Amazon CloudFront caches content closer to your end users, lowering latency for a better end user experience as well as providing potential improvements in cost and resiliency. For a deeper exploration of cost optimization with CloudFront, refer to Cost-Optimizing your AWS architectures by utilizing Amazon CloudFront features.
After you’ve properly architected your application, cleaned up your environment, and deleted resources that are no longer needed, AWS Compute Optimizer provides rightsizing recommendations for Amazon EC2 and Amazon Relational Database Service (Amazon RDS) instances tailored to your workload requirements. AWS Compute Optimizer can also generate license recommendations for commercial software that runs on Amazon EC2. Savings Plans can further accelerate savings, potentially saving you up to 72 percent compared to On-Demand prices. When using AWS Organizations as part of your AWS environment, discounts can be shared across AWS accounts, further optimizing your overall cloud spend. Using Savings Plans Purchase Analyzer, you can model your Savings Plans purchases and evaluate the impact on cost, coverage, and utilization in the AWS Billing and Cost Management Console. Customers with AWS Enterprise Support can also work with their Technical Account Manager to dive deeper on Savings Plans and cost optimization.
Some workloads such as batch, analytics, or generative AI model pre-training are well-suited to Amazon EC2 Spot Instances, which can save customers up to 90 percent compared to On-Demand pricing. Amazon EC2 Spot Instances are integrated into a number of AWS services such as AWS Batch and some third-party services as well.
Amazon Simple Storage Service (Amazon S3) helps you cost-effectively store objects throughout their lifecycle by transitioning them to lower-cost storage classes or deleting expired objects on your behalf. The most seamless way to optimize cost with Amazon S3 is to use the S3 Intelligent-Tiering storage class, which automates storage cost savings by moving data when access patterns change.
To optimize costs while adopting generative AI, organizations can first consider AWS pre-built solutions such as Amazon Q Business, Amazon Q Developer, and Amazon Q in QuickSight for general-purpose use cases rather than building custom applications. For custom solutions, carefully evaluate model selection based on price-performance requirements. Smaller, specialized models such as Amazon Nova Micro can be 70 percent more cost-effective than Amazon Nova Lite but still meet accuracy needs for many scenarios. Another option is model distillation. Distilled models in Amazon Bedrock are up to 500 percent faster and up to 75 percent less expensive than original models, in some cases with less than 2 percent accuracy loss for use cases like Retrieval Augmented Generation (RAG). For RAG applications with large vector databases, consider using Amazon S3 Vectors, reducing total costs by up to 90 percent. Finally, implement prompt caching to reduce inference costs by up to 90 percent for frequently used contexts.
Conclusion
The recommendations in this post will help you reduce waste and optimize costs wherever you are on your cloud journey. To dive even deeper into cost optimization, we recommend following the guidance in the Cost Optimization Pillar of the AWS Well-Architected Framework.
Contact your account team to learn more about how AWS can help to meet your specific goals for optimizing cost and reducing waste, including how you can maximize your benefits of the OneGov agreement.