Hadoop Essentials: A Quantitative Approach (Paperback)

This textbook adopts a unique approach to helping developers and CS students learn Hadoop MapReduce programming fast. Rather than filled with disjointed, piecemeal code snippets to show Hadoop MapReduce programming features one at a time, it is designed to place your total Hadoop MapReduce programming learning process in a common application context of mining customer spending patterns ensconced in large volumes of credit card transaction record data. Precise, end-to-end procedures are given to help you set up your Hadoop MapReduce development environment quickly on Eclipse with Maven on Mac OS X or Linux. Step-by-step procedures are also given on how to set up a four-node Linux cluster at minimum so that you can run your MapReduce programs not only in local mode but also in fully distributed mode on a real cluster. In fact, all MapReduce programs presented in the book have been tested and verified in local mode and on such a Linux cluster.

This textbook mainly focuses on teaching Hadoop MapReduce programming in a scientific, objective, quantitative approach. Rather than heavily relying on subjective, verbose (and sometimes even pompous) textual descriptions with sparse code snippets, this textbook uses Hadoop Java APIs, Hadoop configuration parameters, complete MapReduce programs and their execution logs and outputs to demonstrate how Hadoop MapReduce framework works and how to write MapReduce programs. Specifically, this text covers the following subjects:

* Introduction to Hadoop
* Setting up a Linux Hadoop Cluster
* The Hadoop Distributed FileSystem
* MapReduce Job Orchestration and Workflows
* Basic MapReduce Programming
* Advanced MapReduce Programming
* Hadoop Streaming
* Hadoop Administration

No matter what role you play on your team, this text can help you gain truly applicable Hadoop skills in a most effective and efficient manner. The book can also be used as a supplementary textbook for a distributed computing or Hadoop course offered to upper-division college CS students.

這本教科書採用獨特的方法，幫助開發人員和計算機科學學生快速學習Hadoop MapReduce編程。與其填滿零散、零碎的程式碼片段，一次展示Hadoop MapReduce編程的特性，這本書設計成將你的整個Hadoop MapReduce編程學習過程置於一個常見的應用背景中，該背景是在大量信用卡交易記錄數據中挖掘客戶消費模式。書中提供了精確的端到端程序，以幫助你在Mac OS X或Linux上快速設置你的Hadoop MapReduce開發環境，使用Eclipse和Maven。書中還提供了逐步的程序，教你如何設置一個至少包含四個節點的Linux集群，這樣你不僅可以在本地模式下運行MapReduce程序，還可以在真實集群上以完全分佈式模式運行。事實上，書中介紹的所有MapReduce程序都在本地模式和這樣的Linux集群上經過測試和驗證。

這本教科書主要以科學、客觀、定量的方法教授Hadoop MapReduce編程。它不依賴於主觀、冗長（有時甚至是浮誇的）文字描述和稀疏的程式碼片段，而是使用Hadoop Java API、Hadoop配置參數、完整的MapReduce程序以及它們的執行日誌和輸出來演示Hadoop MapReduce框架的工作原理和如何編寫MapReduce程序。具體而言，本書涵蓋以下主題：

* Hadoop介紹
* 設置Linux Hadoop集群
* Hadoop分佈式文件系統
* MapReduce作業編排和工作流程
* 基本的MapReduce編程
* 高級的MapReduce編程
* Hadoop Streaming
* Hadoop管理

無論你在團隊中擔任什麼角色，這本書都可以幫助你以最有效和高效的方式獲得真正實用的Hadoop技能。這本書也可以作為一門分佈式計算或Hadoop課程的補充教材，提供給高年級大學計算機科學學生使用。