<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Paper Note</title><link>https://blog.shuaizhang.cc/en-us/category/paper-note/</link><description>Posts gathered under this term.</description><generator>Hugo</generator><language>en-US</language><lastBuildDate>Mon, 28 Feb 2022 22:36:11 +0000</lastBuildDate><atom:link href="https://blog.shuaizhang.cc/en-us/category/paper-note/index.xml" rel="self" type="application/rss+xml"/><item><title>Paper Notes: [SOSP'01] SEDA: An Architecture for Well-Conditioned, Scalable Internet Services</title><link>https://blog.shuaizhang.cc/en-us/posts/seda/</link><pubDate>Mon, 28 Feb 2022 00:00:00 +0000</pubDate><guid>https://blog.shuaizhang.cc/en-us/posts/seda/</guid><description>This article reviews how the SEDA architecture improves scalability and stability under high load through a staged event-driven design.</description></item><item><title>Paper Note: [EDBT'16] Designing Access Methods: The RUM Conjecture</title><link>https://blog.shuaizhang.cc/en-us/posts/rum-conjecture/</link><pubDate>Thu, 07 May 2020 00:00:00 +0000</pubDate><guid>https://blog.shuaizhang.cc/en-us/posts/rum-conjecture/</guid><description>This post introduces the RUM Conjecture's view on the trade-offs among read amplification, write amplification, and memory overhead, and discusses its implications for storage-structure design.</description></item><item><title>Paper Notes: [OSDI'16] Slicer: Auto-Sharding for Datacenter Applications</title><link>https://blog.shuaizhang.cc/en-us/posts/google-slicer/</link><pubDate>Sun, 13 Oct 2019 00:00:00 +0000</pubDate><guid>https://blog.shuaizhang.cc/en-us/posts/google-slicer/</guid><description>This article introduces the control plane, client-side routing, and load-balancing design of Google Slicer, an auto-sharding component.</description></item><item><title>Paper Notes: [FTNDB'07] Architecture of a Database System</title><link>https://blog.shuaizhang.cc/en-us/posts/architecture-of-a-database-system/</link><pubDate>Sun, 07 Jul 2019 00:00:00 +0000</pubDate><guid>https://blog.shuaizhang.cc/en-us/posts/architecture-of-a-database-system/</guid><description>This article excerpts and organizes the overall architecture of relational database systems, including core components such as query processing, the executor, and storage management.</description></item><item><title>Paper Notes: [OSDI'10] Large-scale Incremental Processing Using Distributed Transactions and Notifications</title><link>https://blog.shuaizhang.cc/en-us/posts/percolator/</link><pubDate>Sun, 09 Jun 2019 00:00:00 +0000</pubDate><guid>https://blog.shuaizhang.cc/en-us/posts/percolator/</guid><description>This article outlines how Google Percolator implements cross-row transactions and reliable notifications on top of Bigtable, with a focus on the transaction component.</description></item><item><title>Paper Notes: [CONCUR'15] A Framework for Transactional Consistency Models with Atomic Visibility</title><link>https://blog.shuaizhang.cc/en-us/posts/transactional-consistency-models/</link><pubDate>Sun, 26 May 2019 00:00:00 +0000</pubDate><guid>https://blog.shuaizhang.cc/en-us/posts/transactional-consistency-models/</guid><description>This article introduces the formal framework of a paper on transactional consistency models and explains the intuitive meaning of concepts such as atomic visibility.</description></item><item><title>Paper Notes: [FAST'03] ARC: A Self-Tuning, Low Overhead Replacement Cache</title><link>https://blog.shuaizhang.cc/en-us/posts/arc-one-up-on-lru/</link><pubDate>Sat, 30 Mar 2019 00:00:00 +0000</pubDate><guid>https://blog.shuaizhang.cc/en-us/posts/arc-one-up-on-lru/</guid><description>This article summarizes how the ARC cache replacement algorithm adaptively balances recency and frequency, and explains its advantages over LRU.</description></item><item><title>Paper Notes: [Operating Systems Review 2007] Autopilot: Automatic Data Center Management</title><link>https://blog.shuaizhang.cc/en-us/posts/microsoft-autopilot/</link><pubDate>Sun, 15 Jul 2018 00:00:00 +0000</pubDate><guid>https://blog.shuaizhang.cc/en-us/posts/microsoft-autopilot/</guid><description>This article outlines the design goals of Microsoft's Autopilot cluster management system, as well as its mechanisms for machine lifecycle management, application deployment, and automated operations.</description></item><item><title>Paper Note: [ICDE'18] Anna: A KVS for any scale</title><link>https://blog.shuaizhang.cc/en-us/posts/anna-kvs/</link><pubDate>Sun, 08 Apr 2018 00:00:00 +0000</pubDate><guid>https://blog.shuaizhang.cc/en-us/posts/anna-kvs/</guid><description>This post explains how the Anna key-value store uses the actor model and lattice-based conflict resolution to achieve high performance and tunable consistency.</description></item><item><title>Paper Note: [OSDI'14] F4: Facebook's Warm BLOB Storage System</title><link>https://blog.shuaizhang.cc/en-us/posts/facebook-f4/</link><pubDate>Mon, 02 Apr 2018 00:00:00 +0000</pubDate><guid>https://blog.shuaizhang.cc/en-us/posts/facebook-f4/</guid><description>This article summarizes the architectural design of Facebook F4 for warm data object storage, along with its erasure coding and cross-data-center fault-tolerance scheme.</description></item><item><title>Paper Note: [OSDI'10] Finding a Needle in Haystack: Facebook's Photo Storage</title><link>https://blog.shuaizhang.cc/en-us/posts/facebook-haystack/</link><pubDate>Sun, 01 Apr 2018 00:00:00 +0000</pubDate><guid>https://blog.shuaizhang.cc/en-us/posts/facebook-haystack/</guid><description>This post summarizes the design and optimization ideas behind Facebook Haystack's single-machine object storage engine for hot photo storage scenarios.</description></item><item><title>Paper Note: [USENIX ATC'14] In Search of an Understandable Consensus Algorithm (Raft)</title><link>https://blog.shuaizhang.cc/en-us/posts/raft/</link><pubDate>Thu, 29 Mar 2018 00:00:00 +0000</pubDate><guid>https://blog.shuaizhang.cc/en-us/posts/raft/</guid><description>This article introduces Raft's leader election and log replication mechanisms, and discusses how it differs from Paxos in understandability and implementation.</description></item><item><title>Paper Notes: Paxos Made Simple</title><link>https://blog.shuaizhang.cc/en-us/posts/paxos-made-simple/</link><pubDate>Fri, 16 Mar 2018 00:00:00 +0000</pubDate><guid>https://blog.shuaizhang.cc/en-us/posts/paxos-made-simple/</guid><description>This article summarizes Paxos's core safety properties, proposal process, and the basic derivation of the single-decree consensus problem.</description></item><item><title>Paper Note: [SOSP 2007] Dynamo: Amazon's Highly Available Key-value Store</title><link>https://blog.shuaizhang.cc/en-us/posts/amazon-dynamo/</link><pubDate>Mon, 12 Mar 2018 00:00:00 +0000</pubDate><guid>https://blog.shuaizhang.cc/en-us/posts/amazon-dynamo/</guid><description>Reviews the consistency, replication, conflict handling, and routing designs in the Amazon Dynamo paper, and summarizes the key engineering trade-offs it made for highly available key-value storage.</description></item></channel></rss>