ES or Hadoop?


今天在想 Elasticsearch 可以搜尋資料,做即時分析,Hadoop/Spark 也可以。那這兩個有啥差?所以找到了這篇文章:Elasticsearch vs. Hadoop For Advanced Analytics,文章最後的結論就是: Conclusion: Hadoop Still Wins

在另一個討論找到這樣的說明:

Elastic search is made for searching datasets, it has some visualization tools and some basic analysis functions.
Hadoop and Spark are general pourpose mechanisms for distributed computing (one thing that is very prone to be done in distributed systems), you can implement search in then, if you want.

ES 就是 search, Hadoop/Spark 是一般用途,重點在於分散式運算 (Search 是其中一種運算)。如果需要處理自己設計的演算法,或者商業邏輯,那麼 Hadoop 是比較適合的。ES 不是設計用來『儲存資料』的,所以他沒有認證相關功能 (2.0 有)。

AWS 也有 ES Service 不用自己維護 ES Cluster.


Comments