Test Your Distributed Systems Knowledge with this Quiz!

Distributed Systems Lecture 1 Summary

• Backrub, which became Google, was founded in 1997.
• The Cloud is not amorphous, but consists of companies such as Google, Facebook, and Microsoft.
• Facebook's Prineville data center has $250M physical infrastructure and $1B IT infrastructure.
• Distributed systems provide higher-level abstractions to hide complexity, scalability, reliability, well-defined semantics, and security.
• The goal of distributed systems is to appear as a single coherent system to its users.
• A distributed system is characterized by no shared memory, message-based communication, each running its own local OS, and heterogeneity.
• To present a single-system image, a distributed system hides internal organization and communication details while providing a uniform interface.
• A distributed system is organized as middleware, running on all machines and offering a uniform interface to the system.
• The course aims to teach the principles and techniques behind the design of modern, reliable, and high-performance systems, with a focus on distributed systems.
• The course references books such as The Go Programming Language and Distributed Systems: Principles and Paradigms.
• The course prohibits copying another person's programs, comments, README description, or any part of submitted assignments, and publishing code on Github during/after the course.
• MapReduce is a case study in data-parallel programming at scale, where partial aggregation is used to compute word counts from individual files in parallel, merge intermediate output, and compute word count on merged intermediates.