- Practical Real-time Data Processing and Analytics
- Shilpi Saxena Saurabh Gupta
- 120字
- 2021-07-08 10:23:16
Characteristics
The following are important characteristics of Storm:
- Fast: As per Hortonworks, the benchmark is to process one million 100-bytes of messages per second per node. It is lighting fast.
- Reliable: Storm guarantees message processing either at least once or at most once.
- Scalable: Storm can scale to a massive number of messages per second. You need to add more supervisor nodes along with increasing the parallelism of spouts and bolts in a topology.
- Fault-tolerant: In case any Supervisor dies, then Nimbus node redistributes the task to another Supervisor. If any worker dies, then Supervisor redistributes the task to another worker. The same applies to the Executor and the Task.
- Programing language agnostic: Topology can be written in any language.
推薦閱讀
- OpenNI Cookbook
- 概率成形編碼調(diào)制技術(shù)理論及應(yīng)用
- PhpStorm Cookbook
- Learn React with TypeScript 3
- 組態(tài)軟件技術(shù)與應(yīng)用
- Apache Spark 2.x for Java Developers
- Hands-On Full Stack Development with Go
- C語(yǔ)言程序設(shè)計(jì)
- C++20高級(jí)編程
- Mastering AWS Security
- Deep Learning with R Cookbook
- 軟件工程基礎(chǔ)與實(shí)訓(xùn)教程
- 貫通Tomcat開(kāi)發(fā)
- Mastering ASP.NET Core 2.0
- VMware vSphere Design Essentials