Talk to Jack Norris, MapR’s newly appointed senior vice president of data and applications, and he’ll insist that MapR is the data platform of the future.
“No one else has what we have for running mission-critical applications” he told CMSWire, referring to MapR’s newly patented Converged Data Platform (CDP), which has been architected to bring together Hadoop with Spark, web-scale storage, NoSQL and streaming capabilities into one unified cluster, enabling customers to deploy real-time global data applications with unique scale and flexibility.
While it may sound like a bunch of boasting, in its latest Magic Quadrant for Data Management and Data Warehouse Solutions for Analytics, Gartner analysts Roxane Edjlali and Mark A. Beyer wrote that “MapR‘s support for streaming, operational and analytical use cases, all from the same platform, with multi-model support and SQL capabilities across all models,” is one of its primary strengths.
Norris said that while competitors might try to confuse buyers by calling their offerings things like “connected data platform” or claiming to have the flexibility to run a variety of workloads — batch, interactive SQL, enterprise search, and advanced analytics simultaneously — that the products aren’t at all alike.
The others are inherently slower as a result of their design, he said, inferring that the companies, who use them, and not MapR, are at a disadvantage. And Norris wasn’t talking only about Hadoop-based data management solutions, mind you, he said MapR has won deals where Oracle was the competition.
MapR will also begin shipping CDP, which has been further optimized for performance, extended unified security and data governance. More on this later.
And starting today, MapR also offers converged data services across Docker workloads including persistent storage, YARN/Mesos integration and a converged data layer.
Breaking it Down
Both security and governance are must-have’s when considering enterprise-grade data warehouse and data management platforms. One of the ways MapR has sought to separate itself out from the crowd is by being a step ahead of the competition.
So it goes to follow that CDP now offers file and stream access control expressions (ACEs), making it easier to grant permissions to users and groups across data files and directories using Boolean expressions, making security administration more scalable and manageable.
To be more specific: Boolean expressions allow for fine-grain permission (I.e. let everyone but Jack and Diane have access); Intuitive Inheritance provides for subdirectories and files to inherit permissions from parent directories; Whole-Volume ACEs offer volume-level filter – useful in multi-tenant environments so that one group can’t even accidentally view another group’s data; Roles, an arbitrary grouping of users according to business needs.
“All of this granularity is available without a performance hit,” Dale Kim, director, product marketing at MapR, told CMSWire.
Robin Bloor, lead analyst at the Bloor Group explained, that MapR is “a step” ahead of its competitors, partly because of the way CDP was built from the get-go and partly because MapR keeps delivering capabilities that are “sensible” and speak to the concerns of Enterprise customers.
That’s not to say that competitors won’t catch-up and offer the similar capabilities down the road, he added.
CDP is also now optimized for containerized environments.
Closing the Container Gap
More and more enterprises want to run big data applications leveraging container technology, but it has been a challenge because the connection between the data and the application is lost when servers or containers crash.
That’s a problem that MapR claims to have solved for its customers by making CDP act like a comprehensive data services layer for Docker containers. Not only does the platform provide distributed, resilient storage for these containers, but it also includes the database and messaging/streaming capabilities that many containerized operational applications require, according to Kim.
This is accomplished via MapR’s POSIX Client, which Norris said was unmatched when it comes to a fully distributed, secure, reliable, read-write file system to Docker containers for resilient Docker deployments on commodity hardware.
It’s supposed to let organizations deploy data-oriented applications in Docker with the assurance that critical data will be persisted across application or server failures or container movement across servers with no manual intervention.
MapR is also leveraging Apache Myriad, a resource manager that bridges Hadoop YARN and Apache Mesos, to provide for a unified application ecosystem that delivers shared, persistent, high performance storage for all apps with choice of YARN + Services per tenant and/or single, shared YARN for all tenants. Myriad makes it possible for resources to be reallocated across clusters for efficiency.
Norris and Kim are pitching CDP within the context of the Zeta Architecture (similar to Google’s) which is supposed to give enterprises flexibility and scale to deploy an interoperating network of computing technologies including a distributed file system, real-time storage, a pluggable compute model, as well as dynamic and global resource management.
The idea is that by tying the MapR platform with Docker and Myriad, MapR will be able to support a variety of solution architectures and enterprise applications that work together, enabling environments that treat data as a shared pool between enterprise applications and analytics, eliminating data latency, duplication and complex pipelines.
Ready for Big Data Prime Time
In case you haven’t noticed, MapR has been making a great number of product announcements of late, so CMSWire had to ask Norris why such a flurry. “Because they’re ready,” was one answer, “because the market is ready,” was another, but there’s also the sense that MapR believes that this is their time to win.
“Companies have been experimenting with Hadoop using other distributions,” said Norris, “when they’re ready for production, they come to us.”