Upload
noel-bell
View
216
Download
2
Embed Size (px)
Citation preview
Hadoop Distribution Comparison
©2013 OpalSoft Big Data
©2013 OpalSoft Big Data
The following slides would compare Hadoop distribution for 5 prominent big data companies in the market today, based on various aspects:
• Cloudera• HortonWorks• MapR• Amazon EMR• Intel Hadoop
©2013 OpalSoft Big Data
Setup Flexibility:
Cloudera Easy Setup using Cloudera Manager
Horton Works HDP Installer (Ambari). Requires a few different databases.
MapR Easy Setup Installer
Amazon EMR Easy, Management Console
Intel Hadoop Yes, Intel Manager
©2013 OpalSoft Big Data
Security Infrastructure:
ClouderaRole Based authorizationNo info about data at rest security. Hadoop itself supports encryption.
Horton WorksSSL based authentication between client and server machines. No role based access control or data classification or compliance management
MapRNo mention about the data classification or compliance managementRestrictions can be applied at volume level to prevent unauthorized access to files
Amazon EMR Has Amazon virtual private cloudIAM tools by users, roles.
Intel Hadoop Encryption, decryption uusing Inten AES-NI
©2013 OpalSoft Big Data
Learning support:
Cloudera Well Documented and provides training and certifications
Horton Works Well Documented and provides training and certifications
MapR Well documented
Amazon EMR Well documented with tutorials
Intel Hadoop Training provided by Intel
©2013 OpalSoft Big Data
Operation Tools:
ClouderaWeb Based GUI, Cloudera Navigator (Enterprise Subscription),Cloudera manager (Enterprise subscription)
Horton Works Ganglia, Nagios
MapR Yes
Amazon EMR Ganglia and other Amazon cloud monitoring facilities
Intel Hadoop Intel Manager
©2013 OpalSoft Big Data
Operation Support:
ClouderaProfessional support provided for production environment Cloudera Enterprise Support. POC, dev support is available too
Horton Works Professional support provided. 3 levels, Developer, standard, enterprise
MapR 24/7 support and also professional services for POC, implementation
Amazon EMR 24/7
Intel Hadoop 24/7
©2013 OpalSoft Big Data
Market Share:
Cloudera Used by many leading organisation
Horton Works Used by a few leading companies. Not as many as Cloudera
MapRUsed by many companies in commercial, finance and government sectors.Tested partner with Amazon AWS and Google compute engine
Amazon EMR Widely used
Intel Hadoop Not much info about Customers using. # customers mentioned in the website
©2013 OpalSoft Big Data
Developer/People availability:
Cloudera Good
Horton Works Fair
MapR No Info
Amazon EMR No Info
Intel Hadoop No Info
©2013 OpalSoft Big Data
Editions available:
Cloudera Standard, Enterprise
Horton Works Windows, HDP2, SandboxComes with 3 support level
MapR M3, M5, M7
Amazon EMR EC2, may options of data store is available
Intel Hadoop One
©2013 OpalSoft Big Data
Integration with BI tools:
ClouderaCloudera developed connectors for BI toolsMicrostrategy, Netezza, Oracle, Qlikview, Tableau, Teradata
Horton Works Basic ODBC drivers for BI integration
MapRWell integrated with BI tools using JDBC, ODBC, NFS based interfaces, Hadoop interfaces
Amazon EMR Can be used with BI tools
Intel Hadoop No Info
©2013 OpalSoft Big Data
Performance:
Cloudera No particular performance advantage
Horton Works No particular performance advantage
MapRFaster than any other Hadoop distribution because of their Native NFS based file system. Leads to lesser costHA for job tracker
Amazon EMR HA for Job tracker, faster
Intel Hadoop Faster than normal Hadoop setup as hardware and, storage all tuned for Hadoop
©2013 OpalSoft Big Data
Supported OS:
Cloudera RHEL, CentOS, SLES, Debian, Ubuntu, Oracle Enterprise Linux
Horton Works RHEL, CentOS, SLES, Ubuntu, Windows (Only distribution available for windows)
MapR No Info published
Amazon EMR Amazon OS
Intel Hadoop No Info published
©2013 OpalSoft Big Data
Professional Services:
Cloudera Cluster Certification, ETL pilot, analytics pilot, production readiness
Horton Works Yes
MapR Yes
Amazon EMR Yes
Intel Hadoop Yes