HPE Ezmeral Data Fabric (formerly MapR, acquired by HPE in 2019) is a software-defined datastore and file system that simplifies data management and analytics by unifying data across core, edge, and multicloud sources into a single platform. Just as a loom weaves multiple threads into a single piece of fabric, HPE Ezmeral Data Fabric weaves distributed data into a single enterprise-wide data layer that ingests, processes, and stores data once and then makes it available for reuse across multiple…
Well suited: To most of the local run of datasets and non-prod systems - scalability is not a problem at all. Including data from multiple types of data sources is an added advantage. MLlib is a decently nice built-in library that can be used for most of the ML tasks. Less appropriate: We had to work on a RecSys where the music dataset that we used was around 300+Gb in size. We faced memory-based issues. Few times we also got memory errors. Also the MLlib library does not have support for advanced analytics and deep-learning frameworks support. Understanding the internals of the working of Apache Spark for beginners is highly not possible.
MapR had very fast I/O throughput. The write speed was several times faster than what we could achieve with the other Hadoop vendors (Cloudera and Hortonworks). This is because MapR does not use HDFS, which is essentially a "meta filesystem". HDFS is built on top of the filesystem provided by the OS. MapR has their filesystem called MapR-FS, which is a true filesystem and accesses the raw disk drives.
The MapR filesystem is very easy to integrate with other Linux filesystems. When working with HDFS from Apache Hadoop, you usually have to use either the HDFS API or various Hadoop/HDFS command line utilities to interact with HDFS. You cannot use command line utilities native to the host operation system, which is usually Linux. At least, it is not easily done without setting up NFS, gateways, etc. With MapR-FS, you can mount the filesystem within Linux and use the standard Unix commands to manipulate files.
The HBase distribution provided by MapR is very similar to the Apache HBase distribution. Cloudera and Hortonworks add GUIs and other various tools on top of their HBase distributions. The MapR HBase distribution is very similar to the Apache distribution, which is nice if you are more accustomed to using Apache HBase.
The only thing I dislike about spark's usability is the learning curve, there are many actions and transformations, however, its wide-range of uses for ETL processing, facility to integrate and it's multi-language support make this library a powerhouse for your data science solutions. It has especially aided us with its lightning-fast processing times.
1. It integrates very well with scala or python. 2. It's very easy to understand SQL interoperability. 3. Apache is way faster than the other competitive technologies. 4. The support from the Apache community is very huge for Spark. 5. Execution times are faster as compared to others. 6. There are a large number of forums available for Apache Spark. 7. The code availability for Apache Spark is simpler and easy to gain access to. 8. Many organizations use Apache Spark, so many solutions are available for existing applications.
Spark in comparison to similar technologies ends up being a one stop shop. You can achieve so much with this one framework instead of having to stitch and weave multiple technologies from the Hadoop stack, all while getting incredibility performance, minimal boilerplate, and getting the ability to write your application in the language of your choosing.
Increased employee efficiency for sure. Our clients have various levels of expertise in their deployment and user teams, and we never receive complaints about MapR.
MapR is used by one of our financial services clients who uses it for fraud detection and user pattern analysis. They are able to turn around data much faster than they previously had with in-house applications
