CIS5930 Interconnection Networks, Fall 2015
Syllabus
Lecture 1: Syllabus and introduction
Lecture 2: Basics
Lecture 3: Topologies
- Topology
- Programming assignment
1 (group of 2): Write a program that print all links in a
hypercube (A sample program has about 50 lines of code).
An executable for linprog,
try './hcube0.x 8' and './hcube0.x 16'. Due Sept. 8.
- Homework 1: Describe the connectivity of k-ary n-cube. Due Sept. 8.
Lecture 4, 5: Switching, routing, deadlock, and others
- Switching and Routing
- Programming 2: Extend the program in assignment 1 to compute path(s) for
(a) dimension order single path routing (lowest dimension first),
and (b) all shortest path routing, with the following prototypes:
- int dim_order_routing(int src, int dst, int *path); /* return path length */
- int allpath_routing(int src, int dst, int
allpath[MAX_PATH][MAXZ_PATH_LEN]); /* return number of paths */
Write a driver program. Here is an executable for
linprog;
try './hcube1.x 32 dim' and './hcube1.x 32 all'. Due Sept. 15.
Lecture 6: Dragonfly (Read the papers before the class, write a critic
for each paper)
Presenter: Peyman Faizian
- J. Kim, W.J. Dally, S. Scott, and D. Abts, "Technology-Driven,
Highly-Scalable Dragonfly Topology," ACM ISCA 2008.
- G. Faanes, et. al., "Cray Cascade: A scalable HPC system based on
a Dragonfly network," ACM/IEEE SC, 2012.
Lecture 7, 8: Jellyfish (Read the papers before the class, write a critic
for each paper)
Presenter: Shafayat Rahman
- A. Singla, et. al., "Jellyfish: Networking Data Centers Randomly",
USENIX NSDI'12.
- A. Singla, et. al., "High Throughput Data Center Topology Design,"
USENIX NSDI'14.
Lecture 9, 10: Interconnect Modeling
Lecture 11: Interconnect Simulation
Lecture 12: Flit-level Interconnect Simulation
Presenter: Zhou Tong
- Nan Jiang, et. al., "A Detailed and Flexible Cycle-Accurate Network-on-Chip
Simulator", In Proceedings of the 2013 IEEE International Symposium on
Performance Analysis of Systems and Software, 2013. Check out the software
at http://nocs.stanford.edu/cgi-bin/trac.cgi/wiki/Resources/BookSim.
- Fernando Pardo and Jose A. Boluda, "SimuRed: A Flit-level event-driven
simulator for multicomputer network performance evaluation," Computers and
Electrical Engineering, 35(2009)803-814.
Lecture 13: HPC Interconnect issues
Presenter: Kartik Shankar Achalkar
- Abhinav Bhatele, et. al., "Idenfitying the Culprits behind Network
Congestion," IEEE IPDPS 2015.
- P. Fuentes, et. al., "Contention-based non-minimal adaptive routing
in high-radix networks," IEEE IPDPS, 2015.
Lecture 14: Datacenter network issues
Presenter: Nikhil Lloyd Cordeiro
- A. Agache, R. Deaconescu, and C. Raiciu, "Increasing Datacenter Network
Utilization with GRIN," USENIX NSDI, 2015
- S. Hu, et. al, "Explicit Path Control in Commodity Data Centers:
Design and Applications," USENIX NSDI 2015.
Lecture 15: Interconnect Technology
Lecture 16: Interconnect Technology
Lecture 17: SDN and Openflow
Lecture 18 SDN
Presenter: Kartik Shankar Achalkar
- Ben Pfaff, et. al., "The Design and Implementation of Open vSwitch,",
NSDI'2015.
Term project (in groups): Write a report on
"Large Scale Interconnection Networks: State-of-the-art, Major Recent
Developments, and Future R&D Directions." Due 11:59am, Monday, December 7.
Lecture 19, 20: Mininet/Openflow Labs
Lecture 21: Paper presentation
- Brandon Schlinker, et. al, "Condor: Better Topologies through Declarative
Design," SIGCOMM, 2015.
Lecture 22: Paper presentation
- Keqiang He, et. al., "Presto: Edge-based Load Balancing for Fast
Datacenter Networks," SIGCOMM, 2015