How MapR Technologies Achieved the Highest Score for Current Offering In Big Data Hadoop Evaluation

MapR announced at the end of last week that we were among the select companies that Forrester Research Inc. invited to participate in its report entitled “The Forrester Wave™: Big Data Hadoop Solutions, Q1 2014.”  In this evaluation, MapR was cited as a Leader and achieved the highest score for Current Offering among all reviewed vendors. Forrester evaluated nine different companies on 32 different criteria in three different divisions. Our score in the “Current Offering” division was 4.25 out of 5, the highest  of all vendors.

Here’s how MapR scored in specific criteria:

Architecture

4.01

Data and Processing

4.00

Setup, Management and Monitoring Tools

5.00

Hadoop Compatibility and Community

4.00

Current Offering Score

4.25

 

Our Response to the Results:

Let’s take a look at what MapR is doing in each of these categories included in the Forrester report.

Architecture

MapR No-NameNode architecture

MapR has a unique and significant architectural advantage over all other Hadoop distributions. At the root, MapR has made essential architectural improvements that boost reliability, performance, and ease of use in all of Hadoop’s functions.

One of the primary architectural enhancements that MapR made was restructuring how Hadoop stores the metadata. MapR provides a no-NameNode distributed architecture that completely eliminates the single point-of-failure that other distributions have to worry about.

Another benefit of the no-NameNode architecture is the elimination of a file bottleneck. Many bottlenecks result in low throughput.

MapR is built on a POSIX compliant file-system so administrators and users can mount the cluster over the network like enterprise network attached storage (NAS). MapR enables Hadoop to support full random reads/writes across multiple readers and writers. By enabling the user to mount the cluster via NFS, MapR has the capability to ingest data directly into the cluster, run analytics on streaming data and gain real-time access to the results.

Data and Processing

MapR provides a wide range of capabilities for ingestion and ETL. As noted above, MapR exposes a standard NFS interface that is highly available and allows for parallel streaming and leverages the aggregate bandwidth of the entire cluster.

MapR supports processing of not only MapReduce applications but also a range of functions including secure, enterprise-grade search, SQL, database and stream processing. MapR innovations have eliminated administrative overheads and performance bottlenecks for NoSQL applications on Hadoop with our M7 Enterprise Database Edition.

Setup, Management and Monitoring Tools

MapR Control System Dashboard

MapR provides a comprehensive suite of capabilities to enable cluster installation and configuration within minutes. MapR also provides enhancements that eliminate the need to install and configure software.

The MapR Control System (MCS) provides an advanced GUI to manage all cluster functions, including the ability to add and remove drives. MCS also provides a REST API, so it’s easy to integrate Hadoop administration with different open source and commercial tools and build custom dashboards.

MapR provides end-to-end monitoring of the Hadoop cluster; hardware-level monitoring is unique to MapR as it can detect and report disk failures. The MapR JobTracker HA ensures jobs once started always complete.

Hadoop Compatibility and Community

MapR provides a complete distribution for Apache Hadoop, including over a dozen Hadoop-related projects, and is completely compatible with Apache Hadoop and the major sub-projects. We are also fully dedicated to the improvement of open source projects and we are a corporate committer on numerous ones, including Apache Drill, Apache Mahout, Apache Storm, Apache ZooKeeper and Apache Solr.

In addition to contributing to ASF projects, MapR provides our M3 edition to the broad user community at no cost. MapR provides additional innovations directly to customers to transform Hadoop into a reliable compute and dependable data store, with ground-breaking performance. MapR also makes it easier to build Hadoop applications by opening up a much broader set of open APIs to use against the data that’s stored in Hadoop.

MapR has made significant effort to optimize our product offering in all of these categories, and our scores speak for themselves. We are proud to have the Top Ranked Current Offering.

MapR will continue to expand the functionality and performance of our product offering. We look forward to maintaining our leadership in the industry, and look forward to providing businesses with a clear competitive advantage through our enterprise-grade Hadoop solution.

no

Streaming Data Architecture:

New Designs Using Apache Kafka and MapR Streams

 

 

 

Download for free