Schedule

Please reference the syllabus below. As the semester progresses, updates to the syllabus will be posted here. A few of the assigned papers may change. Please contact the instructor or a TA if you have any questions.

Week Date Topics Readings Notes
1 Tues, Aug 27 Course Syllabus and Intro How to Read a Paper
Eric's lecture
1 Thurs, Aug 29 Datacenter background and intro Jupiter Rising: A Decade of Clos Topologies and Centralized Control in Google’s Datacenter Network
Inside the Social Network’s (Datacenter) Network
(optional) A Scalable, Commodity Data Center Network Architecture
(optional) The Datacenter as a Computer: Designing Warehouse-Scale Machines
Eric's lecture
2 Tues, Sept 3 Datacenter transport Datacenter TCP
TIMELY: RTT-based Congestion Control for the Datacenter
2 Thurs, Sept 5 Datacenter virtualization The Design and Implementation of Open vSwitch
(additional material) OVS website
Andromeda: Performance, Isolation, and Velocity at Scale in Cloud Network Virtualization
3 Tues, Sept 10 Datacenter multipath CONGA: distributed congestion-aware load balancing for datacenters
Presto: Edge-based Load Balancing for Fast Datacenter Networks
3 Thurs, Sept 12 Failures NetBouncer: Active Device and Link Failure Localization in Data Center Networks
Understanding and Mitigating Packet Corruption in Data Center Networks
4 Tues, Sept 17 End-host optimizations The eXpress Data Path: Fast Programmable Packet Processing in the Operating System Kernel
(optional, can review) netmap: A Novel Framework for Fast Packet I/O
(present, no review) A Look at Intel’s Dataplane Development Kit
(more info) Introduction to DPDK: Architecture and Principles
4 Thurs, Sept 19 NIC/OS High Performance Packet Processing with FlexNIC
PicNIC: Predictable Virtualized NIC
5 Tues, Sept 24 Software-defined networking (SDN) intro OpenFlow: Enabling Innovation in Campus Networks
P4: Programming Protocol-Independent Packet Processors
(optional) The Road to SDN
Eric's lecture
5 Thurs, Sept 26 SDN B4: Experience with a Globally-Deployed Software Defined WAN
(might change) B4 and After: Managing Hierarchy, Partitioning, and Asymmetry for Availability and Scale in Google's Software-Defined WAN
6 Tues, Oct 1 P4 NetCache: Balancing Key-Value Stores with Fast In-Network Caching
NetChain: Scale-Free Sub-RTT Coordination
6 Thurs, Oct 3 Intro to NFV OpenNF: Enabling Innovation in Network Function Control
Correctness and Performance for Stateful Chained Network Functions
7 Tues, Oct 8 In-network compute When Should The Network Be The Computer?
Eris: Coordination-Free Consistent Transactions using Network Multi-Sequencing
7 Thurs, Oct 10 Hardware-accelerated datacenter networking Azure Accelerated Networking: SmartNICs in the Public Cloud
Offloading distributed applications onto smartNICs using iPipe
(optional) Floem: A Programming System for NIC-Accelerated Network Applications
8 Tues, Oct 15 Traffic Shaping Carousel: Scalable Traffic Shaping at End Hosts
SENIC: Scalable NIC for End-Host Rate Limiting
8 Thurs, Oct 17 RDMA RDMA over Commodity Ethernet at Scale
Revisitng Network Support for RDMA
9 Tues, Oct 22 Study date for exam No class
9 Thurs, Oct 24 Midterm exam
10 Tues, Oct 29 Container networking Slim: OS Kernel Support for a Low-Overhead Container Overlay Network
FreeFlow: Software-based Virtual RDMA Networking for Containerized Clouds
Project Proposals due
10 Thurs, Oct 31 Low latency design pFabric: Minimal Near-Optimal Datacenter Transport
Re-architecting datacenter networks and stacks for low latency and high performance
11 Tues, Nov 5 Coflows Efficient Coflow Scheduling with Varys
CODA: Toward Automatically Identifying and Scheduling Coflows in the Dark
11 Thurs, Nov 7 OS/Networking Shenango: Achieving High CPU Efficiency for Latency-sensitive Datacenter Workloads
Iron: Isolating Network-based CPU in Container Environments
12 Tues, Nov 12 Isolation EyeQ: Practical Network Performance Isolation at the Edge
Silo: Predictable Message Latency in the Cloud
12 Thurs, Nov 14 Datacenter load balancing Duet: Cloud Scale Load Balancing with Hardware and Software
Maglev: A Fast and Reliable Software Network Load Balancer
13 Tues, Nov 19 Monitoring FlowRadar: A Better NetFlow for Data Centers
Pingmesh: A Large-Scale System for Data Center Network Latency Measurement and Analysis
(optional) Synchronized Network Snapshots
Checkpoint 1 due
13 Thurs, Nov 21 Deep Learning Multi-tenant GPU Clusters for Deep Learning Workloads: Analysis and Implications
Tiresias: A GPU Cluster Manager for Distributed Deep Learning
14 Tues, Nov 26 Fall break No class
14 Thurs, Nov 28 Fall break No class
15 Tues, Dec 3 Verification Validating Networks at Scale
p4v: Practical Verification for Programmable Data Planes
Checkpoint 2 due
15 Thurs, Dec 5 Updates Safely and automatically updating in-network ACL configurations with intent language
Dynamic Scheduling of Network Updates
16 Tues, Dec 10 Presentations
16 Thurs, Dec 12 Presentations Project Reports due