Schedule and Reading List

Date	Readings	Presenter
Jan 25	Introduction	Marco
Jan 28	Overview	Marco
Feb 1	Background	Marco
Feb 4	Background	Marco
	ML Systems
Feb 8	Scaling Distributed Machine Learning with In-Network Aggregation (Companion) Efficient Sparse Collective Communication and its application to Accelerate Distributed Deep Learning	Marco
Feb 11	A Unified Architecture for Accelerating Distributed DNN Training in Heterogeneous GPU/CPU Clusters (Companion) Semi-Dynamic Load Balancing: Efficient Distributed Learning in Non-Dedicated Environments
Feb 15	KungFu: Making Training in Distributed Machine Learning Adaptive (Companion) Resource Elasticity in Distributed Deep Learning	Ali
Feb 18	Oort: Informed Participant Selection for Scalable Federated Learning (Companion) Towards Federated Learning at Scale: System Design
Feb 22	Serving DNNs like Clockwork: Performance Predictability from the Bottom Up
Feb 25	Balancing Efficiency and Fairness in Heterogeneous GPU Clusters for Deep Learning (Companion) Gandiva: Introspective Cluster Scheduling for Deep Learning	Salma
	Serverless
Mar 1	Serverless in the Wild: Characterizing and Optimizing the Serverless Workload at a Large Cloud Provider
Mar 4	Peeking Behind the Curtains of Serverless Platforms (Companion) Serverless Computing: One Step Forward, Two Steps Back	Ali
Mar 8	Wukong: A Scalable and Locality-Enhanced Framework for Serverless Parallel Computing
Mar 11	Lambada: Interactive Data Analytics on Cold Data using Serverless Cloud Infrastructure (Companion) Occupy the cloud: distributed computing for the 99%	Salma
Mar 15	Spring break
	Performance and failures
Mar 18	Providing SLOs for Resource-Harvesting VMs in Cloud Platforms
Mar 22	Analyzing System Performance with Probabilistic Performance Annotations
Mar 25	Understanding, Detecting and Localizing Partial Failures in Large System Software
Mar 29	Mid-Semester Presentations
	Distributed systems
Apr 1	Building Scalable and Flexible Cluster Managers Using Declarative Programming
Apr 5	Overload Control for µs-scale RPCs with Breakwater
Apr 8	HovercRaft: Achieving Scalability and Fault-tolerance for microsecond-scale Datacenter Services
Apr 12	NetLock: Fast, Centralized Lock Management Using Programmable Switches
	NICs, FPGAs and ASICs
Apr 15	PANIC: A High-Performance Programmable NIC for Multi-tenant Networks
Apr 22	Enabling Programmable Transport Protocols in High-Speed NICs
Apr 26	hXDP: Efficient Software Packet Processing on FPGA NICs
Apr 29	Lyra: A Cross-Platform Language and Compiler for Data Plane Programming on Heterogeneous ASICs
May 3	Wrap Up
May 6	Final Presentations