Paper Insights #3 - Memory Resource Management in VMware ESX Server

Paper Insights #25 - CliqueMap: Productionizing an RMA-Based Distributed Caching System

Memcached is a popular in-memory cache, but I'd like to discuss CliqueMap, Google's caching solution. Having worked closely with CliqueMap, I have a deep understanding of its architecture. One major difference from Memcached is CliqueMap's use of RMA for reads. We'll also take a closer look at RDMA, a crucial cloud technology that emerged in the 2010s. Paper Link Let's begin with some basic concepts. Network Interface Card (NIC) The NIC facilitates data reception and transmission. Understanding its operation requires examining the fundamental interaction between the CPU and memory. CPU <-> Memory Communication In a Von Neumann Architecture , the CPU and memory are core components, enabling Turing computation. Their communication relies on the system bus (e.g. PCIe ), a set of electrical pathways connecting the CPU, memory, and I/O devices. The system bus comprises three primary logical components: Data Bus : Bidirectional, carrying the actual data being tran...

Paper Insights #26 - Don't Settle for Eventual: Scalable Causal Consistency for Wide-Area Storage with COPS

This work provides a strong foundation for understanding causality , both within distributed systems and more broadly. Its principles underpin systems achieving causal consistency, a powerful form of consistency that ensures high availability. Presented at SOSP 2011, this paper features contributions from prominent distributed systems researchers Wyatt Lloyd and Michael Freedman . Paper Link Let's begin with some basic concepts. Causal Ordering In 1978, Leslie Lamport published Time, Clocks, and the Ordering of Events in a Distributed System , a seminal paper that significantly impacted distributed system design. This work, alongside Paxos and TLA+ , stands as one of Lamport's most influential contributions. A fundamental challenge in distributed systems is clock synchronization . Perfect synchronization is unattainable, a fact rooted in both computer science and physics. However, the goal isn't perfect synchronization itself, but rather the ability to totally order even...

Paper Insights #24 - Spanner: Google's Globally-Distributed Database

This landmark paper, presented at ODSI '12, has become one of Google's most significant contributions to distributed computing. It didn't solve the long-standing core problem of scalability of 2PC in distributed systems, rather, it introduced TrueTime that revolutionized system assumptions. Authored by J.C. Corbett , with contributions from pioneers like Jeff Dean and Sanjay Ghemawat , this paper effectively ended my exploration of distributed SQL databases. It represents the leading edge of the field. Paper Link I would highly recommend reading the following before jumping into this article: 1. Practical Uses of Synchronized Clocks in Distributed Systems where I introduced why clock synchronization is necessary but not sufficient for external consistency. 2. A New Presumed Commit Optimization for Two Phase Commit where I introduced two-phase commits (2PC) and how it is solved in a distributed system. 3. Amazon Aurora: Design Considerations for High Th...

Pico

Search This Blog