Support Questions
Find answers, ask questions, and share your expertise

Performance Comparison Between Spark and Storm

Solved Go to solution

Performance Comparison Between Spark and Storm

Explorer

Good Morning, guys

Before i will go with the question, first i must to tell you this. I'm college student (IT) in Indonesia. This year is my final year and i want to make a project in Big Data. So, i want to get some data with Apache Nifi dan send it to Output. Afterwards, i process it with Spark and Storm.

1. What Performance Comparison should i do ? Response Time ? Throughput ? Real Time Proccessing Speed ?

2. Is it possible to do that ? Because, i know someone in the internet says benchmark between Spark and Storm inaccurate and there is no paper or journal about this. ( http://www.slideshare.net/ptgoetz/apache-storm-vs-spark-streaming )

3. If you have a better idea about my final project. Let me know it.

Thanks before, Have a nice day.

1 ACCEPTED SOLUTION

Accepted Solutions

Re: Performance Comparison Between Spark and Storm

Super Guru

@Rendiyono Wahyu Saputro I recommend you look at storm vs spark in a different manner. if your stream response can handle some latency (as little as 1/2 a second) then spark may be the way to go. This is just my opinion as spark streaming is so darn easy. Storm is a POWERFUL engine with virtually zero latency. Storm has been clocked on millions of tuples per node per second. So you have to ask yourself if your use case needs zero latency or can you handle micro batch (spark streaming)

View solution in original post

1 REPLY 1

Re: Performance Comparison Between Spark and Storm

Super Guru

@Rendiyono Wahyu Saputro I recommend you look at storm vs spark in a different manner. if your stream response can handle some latency (as little as 1/2 a second) then spark may be the way to go. This is just my opinion as spark streaming is so darn easy. Storm is a POWERFUL engine with virtually zero latency. Storm has been clocked on millions of tuples per node per second. So you have to ask yourself if your use case needs zero latency or can you handle micro batch (spark streaming)

View solution in original post