Distributed Systems (15-440) Mohammad Hammoud December 4th, 2013 Course Objectives The course aims at providing an indepth and hands-on understanding on Principles on which distributed systems are based Principles on which distributed systems are optimized Distributed system programming models and analytics engines How modern distributed systems meet the demands of contemporary distributed applications List of Topics Considered: a reasonably critical and comprehensive understanding. Thoughtful: Fluent, flexible and efficient understanding. Masterful: a powerful and illuminating understanding. .1. Architectures and Communications .2. Naming .3. Synchronization .4. Consistency and Replication .5. Fault Tolerance .6. Programming Models .7. Distributed File Systems .8. Virtualization Course Content Course Overview and Introduction (2 Lectures): – – – – Why distributed systems? Defining distributed systems Course overview and intended learning outcomes Trends in distributed systems • • • • High performance platforms Mobile and ubiquities computing Cloud computing Etc., – Challenges in designing distributed systems • Heterogeneity, openness, security, scalability, reliability, concurrency, transparency and quality of service Course Content Architectural Models (1 Lecture): – Client-server, peer-to-peer, tiered and layered architectures Networking (1 Lecture): – Types of networks – Networking principles: • Packet transmission • Network Layers (Physical, data-link, network and transport layers) • Congestion control Course Content Communication Paradigms (1 Lecture): – Socket communication • TCP and UDP sockets – Remote invocation • RPC and RMI – Indirect communication • Message-queuing, publish-subscribe, and group communication systems Course Content Naming (2 Lectures): – Flat naming • Broadcasting, forwarding pointers, home-based naming, and distributed hash tables – Structured naming • Hierarchical name spaces, name resolution, linking and mounting – Attribute-based naming • LDAP and RDF Course Content Synchronization (3 Lectures): – Time synchronization • Physical clocks (UTC, Cristian & Berkeley Algorithms and Network Time Protocol) • Logical clocks (Lamport and vector clocks) – Distributed Mutual Exclusion • Permission-based • Token-based – Election Algorithms • Bully and Ring algorithms Course Content Consistency and Replication (3 Lectures): – Data-Centric Consistency Models: • Continuous, Sequential and Causal Models – Client-Centric Consistency Models: • Eventual consistency and client consistency guarantees – Replica Management: • Server and content replication and placement strategies – Consistency Protocols: • Primary-based, replicated-write and cache coherence protocols Course Content Distributed Programming Models (4 Lectures): – Classical programming models • Shared-memory and message-passing models – MPI Library • Point-to-point and group communication routines – Hadoop MapReduce, Google’s Pregel and CMU’s GraphLab • • • • • • The parallelism models The programming models The architectural models The computational models Task/Vertex/Job scheduling Distributed application suitability Course Content Fault-Tolerance (3 Lectures) – Failure models • Crash, omission, timing, response and byzantine models – Process resilience and agreement protocols • Lamport’s agreement protocol – Reliable communication • Request-reply reliable communication (Request-reply call semantics) • Group reliable communication (Virtual synchrony and atomic multicasting) – Recovery (Checkpointing and message-logging) Course Content Distributed File Systems (2 Lectures): – DFS Aspects: • Architectures (Client-server, cluster-based, and • • • • • • symmetric architectures) Processes (Stateless vs. state-full processes) Communication (RPC2) Naming (Constructing global name spaces) Synchronization (Semantics of file sharing) Consistency and replication (Client-side caching, serverside replication and versioning) Fault-tolerance (Quorum-based mechanisms) Course Content Virtualization (1 Lecture): – Why Virtualization in distributed systems? – Virtualization types • Full virtualization vs. para-virtualization – Virtual machine types • Process VMs vs. system VMs