How to optimize costs for grant-based research projects with AWS
Amazon Web Services (AWS) continues to be a prominent tool in the research workspace by offering key initiatives, funding programs, and resources for research workloads. Agencies like the National Science Foundation (NSF), National Institutes of Health (NIH), and National Aeronautics and Space Administration (NASA) work with AWS to support cutting-edge research. These alliances help researchers work simply and collaboratively with complex workloads and train open datasets to foster innovation. On the AWS Cloud, advanced research and extensive data lakes can be quickly analyzed using transformative technologies like artificial intelligence (AI), machine learning (ML), and quantum computing.
As the number of research workloads in the cloud grows, researchers face challenges in managing strict budgets with unpredictable workloads and fluctuating needs. Researchers must be frugal with time and money to further their goals in performing cutting-edge and important research. AWS provides many methods for effective cost management that can support organizations in their efforts to optimize the use of grant-based funds.
Cost optimization resources in the AWS Cloud for grant-based research projects
Cost optimization is a key pillar in building well-architected solutions in the cloud. For this reason, AWS offers various cost reduction methods: Amazon Simple Storage Service (Amazon S3) tiered storage, Amazon Elastic Compute Cloud (Amazon EC2) Reserved Instances (RI), Savings Plans (SP), and Advance Pay. The cloud offers on-demand resources and pay-as-you-go pricing, benefitting those with fluctuating workloads, strict timelines, and tight budgets.
According to the National Science Foundation, about 42 percent of basic research projects have government funding with a set expiration date. In this situation, a key to realizing the benefits of AWS pricing is to think holistically about each workload and find discounts for committed and steady-state usage. As mentioned, AWS offers pricing models like Reserved Instances and Savings Plans for dedicated usage commitments that produce significant discounts.
Reserved Instances (RIs) allow customers to make long-term commitments for a minimum amount of resources and receive a pricing discount of up to 72 percent. This covers compute services like Amazon EC2 and database services like Amazon Relational Database Service (Amazon RDS), Amazon Redshift, Amazon ElastiCache, and Amazon DynamoDB. These are beneficial if the research workloads are constantly running and the customer commits for an extended period of time. Customers choose between one- and three-year terms, as well as making no upfront, partial upfront, or all upfront payments for different discount amounts. There are both Standard and Convertible Reserved Instances. Convertible RIs, similar to Standard, allow customers to make one- or three-year commitments for discounts. However, Convertible RIs offer more flexibility; customers can exchange them for other instances with different configurations. Customers with strict procurement or cost allocation processes can consider Reserved Instances, as these can be purchased for one- or three-year terms and can secure a definitive cost.
Similar to Reserved Instances, Savings Plans offer discounts due to committments to a particular amount of usage. This flexible pricing model offers low prices on Amazon EC2, AWS Lambda, AWS Fargate, and Amazon SageMaker services. Researchers can commit to a consistent amount of hourly spend over a one- or three-year term and receive a discount. Savings Plans offer the same discounts as Reserved Instances and still offer the option for paying nothing upfront, partially upfront, or all upfront. Additionally, they are simple for first-time cloud users, as they offer greater flexibility in choices of Regions, tenancy, operating system, and usage.
When research workloads are unpredictable, commitment to a particular service or amount of usage can be constrictive. When grant funds are nearing expiration, Advance Pay is a viable option. Customers have the ability to use Advance Pay to prepay for the cloud while not committing to definite usage or particular services. This option resembles a gift card; the preloaded amount pays bills as needed and won’t expire. Once users allocate funds to Advance Pay, AWS can automatically use them to pay invoices when payment is due, simplifying billing and purchase order responsibilities. Additionally, this allows researchers to use grant funds in advance, before expiration dates.
Procurement processes for grants involve collaboration among multiple departments. Typically, federally funded projects require purchase orders (PO). Customers can manage POs on AWS and configure how they reflect on invoices. This helps streamline invoicing and payment processes as well.
Tiered storage is another flexible option for services like Amazon S3, Amazon Elastic Block Store (Amazon EBS), and Amazon Elastic File System (Amazon EFS). These services offer different storage options, at different price points, depending on how frequently and quickly customers need access to that data.
Additionally, researchers looking to store, query, and analyze genomic, transcriptomic, and other omics data, can use Amazon Omics, which recently launched at AWS re:Invent 2022. Amazon Omics is a fully-managed, efficient, and cost effective service for multimodal analyses at petabyte scale. The service offers omics-aware storage that can automatically transition data between active and lower-cost archival storage tiers, similar to Amazon S3.
When to use Reserved Instances, Savings Plans, or Advance Pay for grant-based research projects?
The complicated answer is that it depends on the particular research workload, planned changes during the grant period, budgetary requirements, and commitment to cost-saving opportunities. If the workload is subject to frequent changes but maximizing savings is crucial, customers may consider using Convertible Reserved Instances and Savings Plans. Both offer different flexibilities in the configurations of Amazon EC2 instances, including instance sizes and operating systems. If the workload has a set architecture and steady usage, customers may consider Standard RIs that support securing instances for a specified time period. Customers may consider Advance Pay when a grant deadline is looming, there is a need to lock in indirect costs, or when research workload designs are not final. Advance Pay offers the flexibility to pay for undetermined usage before the funds are actually utilized.
Optimizing an existing workload for grant-based research projects
Customers can optimize costs for an existing workload with AWS Cost Explorer. Cost Explorer is a data visualization tool that helps customers understand and manage costs and usage for accounts across Regions, services, and utilization types. It provides recommendations for Reserved Instances and Savings Plan options based on historical usage. If future usage will follow similar trends, customers can secure pricing models long-term and apply cost savings immediately. Alongside AWS Cost Explorer, customers can use AWS Cost Anomaly Detection and AWS Budgets to closely track allocated research funds and manage unexpected costs.
Next steps in cost optimization for grant-based research projects
AWS offers a variety of ways to optimize costs, receive discounts, and manage funds. Reserved Instances and Savings Plans allow particular usage commitments in exchange for discounts up to 72 percent—with the main difference being the flexibility of Savings Plans. Advance Pay lets customers load funds before use and gives AWS the ability to manage payments for customers. For those in the research industry, Advance Pay can be beneficial in allocating grant funds before expiration when specific utilization of those funds is still undermined.
To learn more about AWS Cloud Economics, contact your AWS Account Team. Once you plan specific workloads with your AWS Account Team, they can identify opportunities to achieve operational efficiency, and increase agility. Finally, for additional information on registering and adding funds to Advance Pay, see Managing Advance Pay.
Read more about AWS for research on the AWS Public Sector Blog:
- Introducing 10 minute cloud tutorials for research
- Accelerating and democratizing research with the AWS Cloud
- How to set up Galaxy for research on AWS using Amazon Lightsail
- 22 new or updated open datasets on AWS: New polar satellite data, blockchain data, and more
- How to set up MATLAB parallel cloud computing on AWS for researchers
Subscribe to the AWS Public Sector Blog newsletter to get the latest in AWS tools, solutions, and innovations from the public sector delivered to your inbox, or contact us.
Please take a few minutes to share insights regarding your experience with the AWS Public Sector Blog in this survey, and we’ll use feedback from the survey to create more content aligned with the preferences of our readers.