Reading List

• Fault-tolerant and transactional stateful serverless workflows

• Milvus: A Purpose-Built Vector Data Management System

• INFaaS: A Model-less and Managed Inference Serving System

• Serving DNNs like Clockwork: Performance Predictability from the Bottom Up

• Dynamic Query Re-Planning using QOOP

• InferLine: ML Prediction Pipeline Provisioning and Management for Tight Latency Objectives

• Semeru: A Memory-Disaggregated Managed Runtime

• Serverless Computing: One Step Forward, Two Steps Back

• HyperSched: Dynamic Resource Reallocation for Model Development on a Deadline

• BPF for storage: an exokernel-inspired approach

• Reducto: On-Camera Filtering for Resource-Efficient Real-Time Video Analytics

• Machine Learning on Graphs: A Model and Comprehensive Taxonomy

• Rearchitecting Linux Storage Stack for μs Latency and High Throughput

• MLSys: The New Frontier of Machine Learning Systems