Dan's Take

MapR-XD Converged Data Platform Offers Cloud-Scale Data Fabric

A place to put all your "data stuff."

• By Dan Kusnetzky
• 06/08/2017

MapR Technologies, Inc.  announced MapR-XD, describing it as "a cloud-scale data store  to manage files and containers." The company went on to explain that MapR-XD is "part of the MapR Converged Data Platform." This product is designed to support "any data type from the edge to the data center and multiple cloud environments with automatic policy driven tiering from hot, warm or cold data."

Here is how MapR describes some of MapR-XD's capabilities:

• Files, Container Support. MapR-XD is designed to simplify management across many different types of files and containers. It unifies management for security and data protection while also supporting high availability. The company points out that the same underlying data can be accessed using a number of standard APIs including: NFS, POSIX, HDFS and S3 to simplify development, administration, and eliminate data  sprawl.
• Global exabyte scale. Scaleability is one of the attributes of MapR-XD that the company touts. MapR-XD scales to support trillions of files, exabytes of data, on thousands of commodity servers or cloud instances, all accessible through a single global namespace.
• Cloud-grade reliability. MapR-XD delivers high availability, data protection and disaster recovery with no single points of failure, fully distributed metadata, point-in-time snapshots and high-performance, distributed mirroring.
• Speed at Scale with Flash. The MapR-XD utilizes the full power of network interconnects and takes advantage of the available performance of the underlying heterogeneous hardware such as disk, flash, GPU.
• Stateful Persistence for Containerized Applications. MapR-XD includes a secure optimized container client for providing containers with access to persisted data. The client supports both legacy and new containerized event based microservices applications; multiple data types of files, containers, database, event streams; works with multiple schedulers such as kubernetes, mesos, docker swarm; and across any infrastructure such as on-premise, multiple clouds and edge.   

Before I get too far into commenting on MapR's announcement of MapR-XD, I'm reminded of a joke by Steven Wright: "You can't have everything, where would you put it?"

MapR's answer to Wright's joke is MapR-XD. They believe you can put nearly everything there and you'll be able to find it and use it, regardless of where is it physically placed or what type of media it ends up on.

One of the more challenging areas of IT in the world in which everyone carries around a mobile device, increasingly-intelligent vehicles, and huge masses of data being generated everywhere is where to put the data. It needs to be effectively and safely stored, made quickly available for analysis and still keep storage and communications costs under control. MapR believes that MapR-XD is an answer to those questions, and questions that haven't yet been asked.

Consider the example of needing to maintain data locally in a moving vehicle that gathers enormous amounts of operational data and has only a limited amount of storage. This scenario also requires the data to made available, as needed, on a global basis, without requiring developers and users to become experts in networking, caching, storage, auto-tiering and several other IT specialties. It appears that I'm not the only one who sees the possibilities for situations similar to this.

Maybe that's why MapR has an impressive list of customers using its technology in nearly every vertical market worldwide.