Apache Distribution
Apache Hadoop refers to an open source framework that provides the users with a distributed platform for processing of large sets of data across collections of machines using simple programming paradigms (Hortonwork, 2011). The distribution allows scalability from one server to several servers, with each server rendering resident processing and storage. One of the features that make apache preferred by several organizations is its capability to hold, manage and analyze huge volumes of both structured and unstructured data quickly and at low cost. Some of the features include:
Reliability-Apache distribution is resilient, and when one node fails, then processing can be redirected to other remaining nodes within the system.
Scalability and performance- data is processed on a distributed scale thus improving the performance
Low-cost- unlike the commercial software models, apache is open source and can be installed on the cheap hardware.
Cloudera
Cloudera was founded in 2008. The Cloudera distribution enhances the open source Hadoop software by integrating certain features to it with a view of improving security, governance, availability and software administration. It provides the major elements of Hadoop specifically distributed processing and scalable storage together with a web-based user interface (Reifer, 2016). Moreover, it also has unified batch processing capabilities, access controls on the basis of roles, and interactive SQL. Strengths include:
Flexibility- this Hadoop type can handle any data type and conduct any type of computation on the same data.
Security- it can process highly sensitive data without compromising security.
High availability- it can conduct mission-critical tasks with a lot of confidence
Scalability-it can be expanded to as the requirements increase.
MapR
MapR has taken a varied approach from the two above. Although they make use of core Hadoop, they don’t put more focus on open source compared to the other two distributions discussed above. They have the majority of proprietary software in their collection, usually a custom file system (Butler, 2016). However, they don’t use HDFS. MapR also does not employ the namecode architecture in its implementations. The features and strength of apache include:
High-availability capabilities prevents job restarts
Quick recovery from errors with atomic snapshots
Less expensive due to NFS direct access
Use of native authentication mechanism and Kerberos for security.
References
Butler, B. (2016).The top 5 Hadoop distributions, according to Forrester. Network World. Retrieved from < http://www.networkworld.com/article/3024812/big-data-business- intelligence/the-top-5-hadoop-distributions-according-to-forrester.html >
Hortonwork. (2011). Apache Hadoop. Retrieved from < http://hortonworks.com/apache/hadoop/>
Reifer, A. (2016).Learn More About The Cloudera Hadoop Distribution. Tech Target. Retrieved from < http://searchdatamanagement.techtarget.com/feature/Learn-more-about-the- Cloudera-Hadoop-distribution >
