You need 10 reasons for Spark

Top 10 Reasons You Need Spark: 1. Spark is the only current replacement for revolutionary Hadoop that does everything Hadoop does and is more than 100 times faster than Hadoop: Logistic regression in Hadoop and Spark can be seen in Spark's special field of expertise More than 120 times faster than Hadoop!

2, the original support for Hadoop's four major business organizations have announced support for Spark, including the well-known Hadoop solution provider Cloudera and the well-known Hadoop provider MapR;

3, Spark is the next generation cloud computing big data core technology instead of Hadoop after Hadoop. At present, SPARK has built its own entire big data processing ecosystem such as stream processing, graph technology, machine learning, NoSQL query and so on Has its own technology, and is Apache's top-level project, it is estimated that 3721.html "> from the second half of 2014 to 2015 there will be explosive growth in community and business applications.

4, some large foreign Internet companies have deployed Spark. Even Yahoo, the early major contributor to Hadoop, is now deploying Spark across multiple projects; domestic companies such as Taobao, Youku Tudou, Netease, Baidu, Tencent and others have used Spark for their own commercial production systems, both domestically and internationally Starting wider and wider. Spark is gradually maturing and playing a more important role in this area.

One thing I must mention is Spark's "Onestacktorulethemall" feature. One of the characteristics of Spark is that it uses a technology stack to solve all the problems of cloud computing big data stream processing, graph technology, machine learning, interactive query, error inquiries, At this point we only need a technical team through Spark can get everything, and if based on Hadoop need to build a real-time flow processing team, data statistical analysis team, data mining team, etc., and these teams both in code and experience are Can not learn from each other, will form a huge cost, but the use of Spark does not exist this problem;

6, Mahout said in the previous phase that from now on they will no longer accept any form of implementation in the form of MapReduce; on the other hand, Mahout announces that the new algorithm is based on Spark;

If you already use Hadoop, Spark is even more needed. The previous phase of Mahout said that from now on they will no longer accept any form of implementation in the form of MapReduce. On the other hand, Mahout announced that the new algorithm is based on Spark. In the meantime, Hadoop improvements have been basically at the code level for years. That is, tinkering, which led to Hadoop now has a depth of "technical debt", a heavy load;

8, with the popularization and promotion of Spark technology, the demand for professionals is increasing. Spark professionals are also hot in the future, easily get millions of remuneration;

9, tens of billions of dollars market, professors resign, students drop out of school, the trend of the times!

10, Lifeisshort.