CSC 2231: Internet Systems and Services
Fall 2005

Course Description

Instructor: Stefan Saroiu
Lectures: Mondays and Thursdays 1-2
Location: BA4290 (new location!!!)

Internet systems and services have grown to become indispensable tools; we use them to buy books online, to schedule flights, to chat with people world-wide, or to download multimedia content, such as songs and movies. Despite the Internet's apparent maturity, we are far from understanding how to build fast, scalable, fault-tolerant, and, secure Internet-scale systems. Exploring current solutions and understanding their limitations is an important prerequisite in designing the next generation of Internet systems and services. The need for highly scalable, available, and secure large networked systems is likely to be exacerbated as the Internet becomes as ubiquitous as the telephone and utility networks.

This course is intended to be a tour through modern Internet systems. Topics include peer-to-peer systems, clusters, Web systems, content delivery networks, and Internet security. The course's goals are:

• To be familiar with a number of influential and interesting papers on these topics;
• To understand the state-of-the-art in modern Internet systems;
• To understand how Internet systems research is done and evaluated;
• To investigate novel ideas over a short-term project.

Prerequisite: CSC458 and CSC469 or permission of the instructor.

Course Deliverables

1. The students will read and review research papers. There will be one or two papers assigned as reading for each lecture. The students will submit their paper reviews two hours before lecture. Reviews submitted later than two hours before the lecture will not be read. David Wetherall has a great list of suggestions on how to review papers (here).

2. The students will participate in class. Participation includes but is not limited to asking questions, raising a point, presenting an argument, offering an alternative point of view.

3. The students will complete a group research report on a topic relevant to this course. The instructor will provide a list of suggestions, but the students are free to choose any topic. The research projects will be peer-reviewed by the students and the instructor. More details here.

4. The students will undertake a group reserch project during the term. The goal is to undertake a research idea that could eventually lead to a publication. The instructor will provide a list of suggestions, but the students are free to choose any topic. If you take CSC2228 and this course simultaneously, you can use the same project for both courses. Come see us (Eyal and I) for this. More details here.

The final grade will depend on:

• research project (50%)
• research report (25%)
• paper reviews (15%)
• class participation (10%)

Syllabus and Schedule

Note: All papers linked from this page are copyrighted by their respective authors or publishers.

Date	Topic	Papers	Lectures
M 09/12/05	Introduction	None	Lecture 1
R 09/15/05	NOW -- Network of Workstations Forming report teams due at noon.	NOW (Check reviews)	Lecture 2
M 09/19/05	Projects Discussion Forming project teams due at noon. Choosing report topics due at noon.
R 09/22/05	Building big servers Project topics due at noon.	Lessons (Check reviews)	Lecture 3
M 9/26/05	Building fast servers	LARD (Check reviews)	Lecture 4
R 9/29/05	Building fault-tolerant servers	Failure study (Check reviews), Beyond Availability	Lecture 5
M 10/3/05	DNS	DNS (Check reviews)	Lecture 6
R 10/6/05	Akamai Project proposals due at noon.	Akamai (Check reviews)	Lecture 7
M 10/10/05	Thanksgiving Day
R 10/13/05	Web workloads	Zipf (Check reviews)	Lecture 8
M 10/17/05	Cascon - No lecture
W 10/19/05	Reports due at 11:59pm EST.
R 10/20/05	Cooperative Web Caching	Cooperative Web caching (Check reviews)	Lecture 9
M 10/24/05	SOSP - No lecture
R 10/27/05	SOSP - No lecture Report reviews due at noon.
M 10/31/05	Network measurements	TCP congestion control (Check reviews), Sting (Check reviews)	Lecture 10
R 11/03/05 2-hour Lecture (start at 12pm)	Latency on the Internet Research report introspection Project progress reports due at noon.	Internet Path Selection (Check reviews), King (Check reviews)	Lecture 11
M 11/07/05	Peer-to-Peer intro	Chord (Check reviews), Pastry (Check reviews)	Lecture 12
R 11/10/05	DNS as DHT	Chord DNS (Check reviews), CoDoNs (Check reviews)	Lecture 13
M 11/14/05	P2P Churn	Understanding Availability, (Check reviews), Pick Two (Check reviews)	Lecture 14
R 11/17/05	P2P Security	P2P Security, (Check reviews), Sybil (Check reviews)	Lecture 15
M 11/21/05	DHT Geometries	DHT geometries (Check reviews)	Lecture cancelled
R 11/24/05	P2P Replication	TotalRecall, (Check reviews), Glacier (Check reviews)	Lecture 16
M 11/28/05	File-sharing workloads	The Long Tail, (Check reviews), P2P Workloads (Check reviews)	Lecture 17
R 12/1/05	Video file-servers	Tiger (Check reviews)	Lecture 18
W 12/7/05 (4-6:30pm)	Project Presentations Presentation schedule
R 12/8/05	State of the Course Address

Papers

• A Case for NOW, Tom Anderson, David Culler, David Patterson, IEEE Micro, February 1995.
• Lessons from Giant-Scale Services, Eric Brewer, IEEE Internet Computing 2001.
• Locality-Aware Request Distribution in Cluster-based Network Servers Vivek S. Pai, Mohit Aron, Gaurav Banga, Michael Svendsen, Peter Druschel, Willy Zwaenepoel and Erich Nahum, ASPLOS 1998.
• Why do Internet services fail, and what can be done about it? David Oppenheimer, Archana Ganapathi, and David A. Patterson. 4th USENIX Symposium on Internet Technologies and Systems, March 2003.
• Beyond Availability: Towards a Deeper Understanding of Machine Failure Characteristics in Large Distributed Systems Praveen Yalagandula, Suman Nath, Haifeng Yu, Philip B. Gibbons, Srinivasan Seshan. WORLDS 2004.
• Development of the Domain Name System Paul Mockapetris and Kevin Dunlap. Sigcomm 1988.
• Globally distributed content delivery. J. Dilley, B. Maggs, J. Parikh, H. Prokop, R. Sitaraman, and B. Weihl, IEEE Internet Computing, September/October 2002, pp. 50-58.
• TCP Congestion Control with a Misbehaving Receiver Stefan Savage, Neal Cardwell, David Wetherall and Tom Anderson. ACM Computer Communications Review, 29(5):71-78, October 1999.
• Sting: a TCP-based Network Measurement Tool Stefan Savage. USITS 1999.
• The End-to-end Effects of Internet Path Selection Stefan Savage, Andy Collins, Eric Hoffman, John Snell and Tom Anderson. Sigcomm 1999.
• King: Estimating Latency between Arbitrary Internet End Hosts Krishna P. Gummadi, Stefan Saroiu, and Steven D. Gribble. IMW 2002.
• Web Caching and Zipf-like Distributions: Evidence and Implications Lee Breslau, Pei Cue, Li Fan, Graham Phillips, Scott Shenker, Infocom 1999.
• On the scale and performance of cooperative Web proxy caching. Alec Wolman, Geoffrey M. Voelker, Nitin Sharma, Neal Cardwell, Anna Karlin, and Henry M. Levy. In Proceedings of the Seventeenth ACM Symposium on Operating Systems Principles (SOSP), Kiawah Island, SC, December 1999.
• Chord: A Scalable Peer-to-peer Lookup Protocol for Internet Applications Ion Stoica, Robert Morris, David Liben-Nowell, David R. Karger, M. Frans Kaashoek, Frank Dabek, Hari Balakrishnan. IEEE/ACM Transactions on Networking.
• Pastry: Scalable, distributed object location and routing for large-scale peer-to-peer systems Antony Rowstron and Miguel Castro. IFIP/ACM International Conference on Distributed Systems Platforms (Middleware). 2001.
• The Impact of DHT Routing Geometry on Resilience and Proximity Krishna P. Gummadi, Ramakrishna Gummadi, Steven D. Gribble, Sylvia Ratnasamy, Scott Shenker, and Ion Stoica. Proceedings of the ACM SIGCOMM 2003.
• Serving DNS using a Peer-to-Peer Lookup ServiceRuss Cox, Athicha Muthitacharoen, Robert Morris. IPTPS 2002.
• The Design and Implementation of a Next Generation Name Service for the Internet. Venugopalan Ramasubramanian and Emin Gun Sirer. Proceedings of the ACM SIGCOMM 2004.
• Understanding Availability Ranjita Bhagwan, Stefan Savage and Geoffrey Voelker. IPTPS 2003.
• High Availability, Scalable Storage, Dynamic Peer Networks: Pick Two Charles Blake and Rodrigo Rodrigues. HOTOS 2003.
• Security for structured peer-to-peer overlay networks M. Castro, P. Druschel, A. Ganesh, A. Rowstron, and D. S. Wallach. OSDI 2002.
• The Sybil Attack John R. Douceur. IPTPS 2002.
• TotalRecall: System Support for Automated Availability Management Ranjita Bhagwan, Kiran Tati, Yuchung Cheng, Stefan Savage and Geoffrey M. Voelker. NSDI 2004.
• Glacier: Highly Durable, Decentralized Storage Despite Massive Correlated Failures Andreas Haeberlen, Alan Mislove, and Peter Druschel. NSDI 2005.
• Measurement, Modeling, and Analysis of a Peer-to-Peer File-Sharing Workload Krishna P. Gummadi, Richard J. Dunn, Stefan Saroiu, Steven D. Gribble, Henry M. Levy, John Zahorjan. SOSP 2003.
• Distributed Schedule Management in the Tiger Video File Server William J. Bolosky, Robert P. Fitzgerald, and John R. Douceur. SOSP 1997.