Reputation: 1
We have been using Standard Hadoop distribution provided by Apache Community. We are working to develop Sales analytics solution .Based on following document https://wiki.apache.org/hadoop/Distributions%20and%20Commercial%20Support
It seems to be many companies providing different distributions and commercial support for Hadoop. In our case we are using only Standard (Debian) Hadoop distribution. We are encountering many problems. Our question is shall we use this Debian Hadoop distribution or shift to other distributions provided by Cloudera, Hortonworks, MapR ...?
Upvotes: 0
Views: 41
Reputation: 97
There are multitude of factors which would go into your decision making, but here are some pointers:
Cloudera - Most expensive of all the vendors. Their "Cloudera Manager" is top notch cluster management software.
Hortonworks - Least expensive and completely open source components.
MapR - Somewhere in the middle in terms of cost but definitely number 1 when it comes to performance. They have gotten rid of NameNode altogether and have some other major changes.
AWS - They allow you to use their infrastructure for a low hourly cost. Support is $15k flat or a variable amount for larger companies. No need to buy hardware if using this option.
There are many other distros like Pivotal, IBM, etc but these three are the market leaders.
Hope this helps.
Upvotes: 0