MapR Distributed File and Object Store (MapR XD) has the capability to handle trillions of files, thousands of nodes, and hundreds of petabytes across thousands of client hosts, clusters, and racks and across different geographical locations. MapR XD has the unique ability to extend across on- premises enterprise datacenters, edge clusters, and the public cloud.
End users get a unified view and access method to the files without having to be aware of the physical location of the files. This allows for easy data management, very minimal overhead, and distributed scale since the files can now be spread across edge, on-premises, and the cloud.
With MapR XD, you get the unique benefit of deploying NFS-based applications and Container based applications. MapR XD allows you to consolidate across different environments to keep your operating costs down. With MapR XD, there is no need to stand up different clusters for different types of data.
MapR XD maintains different temperatures on data based on their access frequency—hot, warm, and cold. Using topology awareness, MapR XD keeps track of nodes and racks across the datacenter. Data volumes are then placed on the appropriate nodes depending on their temperature. Volumes are also automatically moved across tiers as the temperature of data changes. This automated way of tiering data ensures invested capacity is efficiently managed and used at all times.
MapR XD supports different kinds of infrastructures. You can build an on-premises cluster with commodity hardware or extend your datacenter to the cloud, or build edge clusters if your business requires it. Whatever the environment, MapR XD offers a solution. With MapR XD, you can deftly migrate data from on-premises to the cloud. MapR XD also offers a simple management tool that allows you to manage all your data on MapR across different environments.
The foundational framework for MapR XD ensures that multiple copies of data are spread across nodes, clusters, and racks. This ensures that no data is lost at any given failure scenario. Access to data is always achieved by accessing any of the copies. During failure scenarios, data is reconstructed at a fast pace wherein every node pitches in towards the reconstruction. The end result is a highly reliable and available cluster powered by MapR XD.
MapR XD has multi-tenancy characteristics built in throughout its stack. Volumes are units of management within MapR XD, and customers can create isolation at volume level and add quotas per tenant level as well. In addition to industry standard authentication, authorization, wire-level encryption and auditing features, MapR XD offers a unique feature called Access Control Expressions (ACE), which allows end users to be given or denied access to volumes. This combination of unique features makes MapR XD the ideal platform for multi-tenant environments.
With MapR XD the complex decision around what kind of storage and how much you need is removed. MapR XD supports HDDs and flash and can be hosted on-premises, the cloud, and edge clusters. This allows you the freedom of choice with less pressure upfront since you can easily scale to different types and sizes at a later stage. MapR XD is flash optimized for those applications that are performance sensitive and supports disks for less accessed data. You can migrate data from a MapR on-premises cluster to the cloud for long-term repository then brings it back for analytical purposes.
MapR XD built-in analytics allows you to analyze the data without having to move the data out of the cluster once the analytics is complete. This reduces time to deploy and unlike traditional analytical environments, reduces costs in having to setup dedicated clusters. In situations where you simply want to create an analytical environment, MapR XD allows you to achieve the same results with no extra time or cost.
MapR XD exposes an NFS interface to the external clients. You can bring data into the MapR cluster using the simple NFS and/or Posix access methods and analyze the same data using Hadoop, Spark, or industry-specific data processing apps. NFS is a well established industry standard open protocol that allows you to deploy both legacy as well as newer applications on the same MapR XD platform. Using NFS to bring in data also eliminates the need to find and manage ETL tools just to ingest data, which minimizes the overall management aspect. And because MapR XD is radically simple, it requires less specialized skill set to setup, configure, and manage thereby reducing the operation complexity and cost.