CS 425

Protocols, specification techniques, global states and their determination, reliable broadcast, transactions and commitment, security, and real-time systems. Course Information: Same as ECE 428. 3 or 4 undergraduate hours. 3 or 4 graduate hours. Prerequisite: Credit or concurrent enrollment in one of CS 240, CS 241, CS 340, CS 341 or ECE 391.

No required textbook. Webnotes: https://courses.engr.illinois.edu/cs425/fa2018/lectures.html

Design solutions ­both algorithmic and systems to new problems that arise in today¹s distributed systems, and reason about their practicality, i.e., complexities and tradeoffs involved. (1, 2, 6)
Identify classical distributed computing concepts in the context of today¹s distributed computing systems, e.g., in the context of clouds, peer-to-peer systems, mobile systems, etc. Apply known concepts where applicable, or design new ones where new solutions are required. (1, 2, 6)
Know the fundamental limitations of what is possible in asynchronous distributed systems and what is not. Apply these in today's distributed systems. (1, 2, 6)
Analyze a given distributed algorithm in terms of its computation and communication complexity, scalability, and fault-tolerance. (1, 2, 6)
Build, debug, deploy and demo a fully working distributed system from scratch and in progressive stages, given merely specifications for each stage, e.g., specifications about scalability and fault-tolerance. Example programming assignments have include cloud computing systems like key-value stores, Mapreduce, distributed hash tables, peer-to-peer wireless systemsm stream processing systems, and graph processing systems. (1, 2, 3, 4, 5, 6)
Be able to design and implement working distributed systems in project groups. Pair programming is highly encouraged. (1, 2, 3, 4, 5, 6)
Know and be able to apply basics of popular industry distributed systems today (e.g., Google¹s and Yahoo¹s Hadoop), and basics of working with distributed systems in industry today (e.g., unit testing, good logging, etc.). (1, 2, 6)

Distributed Computing Theory Fundamentals: Failure detectors, Time synchronization, Lamport timestamps, Global snapshots, Multicast, Leader Election, Mutual Exclusion,
Distributed Computing Systems Fundamentals: RPCs and RMIs, Peer to peer systems, Distributed hash tables,
Consensus: Impossibility, Paxos, Two-phase commit

Cloud Computing: Basics, Mapreduce, Key-value stores, Graph processing, Stream Processing
Distributed Concurrency and Replication: Concurrency control, Replication control, Quorums, Gossiping, Self-stabilization
Security: Basics, Byzantine fault-tolerance
Miscellaneous: Distributed file systems, Distributed shared memory,

Revisions in last 6 years	Approximately when revision was done	Reason for revision	Data or documentation available?	Documentation provided?
Introduced new topics such as cloud computing (e.g., mapreduce, key-value stores) as well as classical topics such as self-stabilization, graph processing, and wireless systems.	Fall 2010-Present	To keep the course current with evolving industry trends.	informal discussions and coordination among Gupta, Nahrstedt, Mitra over several semesters	informal discussions
Reduced focus on transactions	Fall 2010-Present	To make space for new topics (see above)	informal discussions and coordination among Gupta, Nahrstedt, Mitra over several semesters	informal discussions
Programming assignments oriented towards cloud computing industry	Fall 2012-Present	To keep the course programming assignments current with evolving industry trends.	informal discussions	informal discussions

Selected Elective.

2/12/2019by Indranil Gupta