FSISUS07: How do you optimize batch processing components for sustainability?
Because batch processing is often found within many workloads across financial systems, verify that the minimum number of resources are consumed by batching transactions together while meeting your customer SLA and system requirements.
FSISUS07-BP01 Optimize your batch processing systems
Because batch processing is often found within many workloads across financial systems, verify that the minimum number of resources are consumed by batching transactions together while meeting your customer SLA and system requirements.
Prescriptive guidance
-
Queue up several requests together that don't require immediate processing.
-
Increase serialization to flatten utilization across your pipeline.
-
Modify the capacity of individual components to prevent idling resources waiting for input.
-
Create buffers and establish rate limiting to smooth the consumption of external services.
-
Use the most efficient available hardware and services to optimize your software.
-
If possible, schedule jobs during times of day where carbon intensity for power is lowest.
-
Use managed spot training for generative AI model training to utilize spare EC2 capacity efficiently.
-
Implement parameter-efficient fine-tuning (PEFT) techniques like LoRA to reduce computational requirements.
-
Optimize generative AI batch inference jobs using serverless architectures.
