Sustainability
The sustainability pillar provides you with the discipline to address the long-term environmental, economic, and societal impact of your business activities. You can find extensive prescriptive guidance on your implementation's sustainability in the Well-Architected Sustainability Pillar whitepaper.
Financial institutions must focus on sustainability within their cloud operating model to reduce their impact on the environment and to encourage sustainable practices. Focusing on these areas helps financial institutions adapt their workloads to financial services industry sustainability best practices, to adopt new environmentally friendly technology trends, and to plan for the business impacts of potential future regulatory requirements.
Sustainability topics
The sustainability pillar includes the following topics on AWS cloud-based architectures. Keep these topics in mind when developing your workload and also when assessing the sustainability performance of your workloads.
-
Cloud sustainability: Compare cloud capabilities against on-premises or hosted sustainability performance.
-
Shared responsibility model: You and AWS are responsible for your cloud sustainability performance. AWS is responsible for providing the most sustainable infrastructure possible while you are responsible for judiciously developing workloads that take advantage of the most sustainable options provided by AWS.
-
Sustainability of the cloud: AWS' responsibility to you
-
Sustainability in the cloud: Your responsibility
-
-
Design principles for sustainability in the cloud: This lens pillar and the sustainability whitepaper offer an extensive set of design principles and best practices for achieving the best possible outcomes.
-
Improvement processes: After performing a Well-Architected Framework review, a number of improvement recommendations are provided by the AWS Well-Architected Tool. You can use this process to implement sustainability improvements to your workloads.
-
Best practices for sustainability: The sustainability pillar and this lens pillar provide recommendations for sustainable practices across six areas of your critical workload infrastructure.
Design principles
The following section defines a set of design principles for financial services sustainability best practices. Typically, some areas of financial services workload performance must be very low latency (for example, trading and investing where microseconds can cost $1MMs) while other areas have no concerns about speed. In many areas of financial services, data retention for at least seven years is critical. However, maintaining low latency storage strategies for six-year-old insurance data is very wasteful.
-
Implement low-latency workloads only for time-critical performance. Trading generally requires high-performance compute, networks, and storage, while banking and insurance typically do not.
-
Use tiered storage for data requiring long-term archive storage. Financial records typically must be stored for at least seven years. Amazon S3 storage classes can better align resource usage to retrieval needs. Amazon S3 Standard is not recommended for long-term storage without any requirement for timely data retrieval. Amazon Glacier storage classes offer long- term, secure, durable storage classes for data archiving.
-
Region selection is a complex factor for implementing financial workloads. While selection of low-carbon Regions is generally recommended for processing of financial data, sometimes data residency requirements stipulate the use of higher carbon storage. Also, some financial data's low latency requirements drive the choice of Regions with a higher carbon footprint due to greater network latency requirements. The selection of the best Region might be driven by considering a variety of reasons.
-
Back up data only when it's difficult to recreate. Too often data is backed up in multiple locations and in the wrong tiers of storage. Use smart backup and storage disciplines to reduce your workload's overall carbon footprint.
-
Agent resource consumption: Monitor and optimize computational resources used by autonomous agents
-
Agent task prioritization: Implement sustainability-aware task scheduling for non-time-critical agent operations.
-
Agent efficiency metrics: Define sustainability KPIs specific to agent operations.
-
Sustainable agent design patterns: Develop patterns that minimize environmental impact of agent operations.
-
Agent hardware selection: When possible, select energy-efficient infrastructure for agent workloads.
Definitions
