MapR Technologies announced Tuesday the launch of the MapR App Gallery. The gallery, the first of its kind in the industry, includes ready-made applications for administrators, analysts, and developers from a range of Hadoop ecosystem partners. It’s designed to help customers derive greater value from their data as they scale out their enterprise data architectures.
The company made the announcement at the 7th annual Hadoop summit in San Jose, California.
“2014, in our minds, is the year of big data apps on Hadoop,” said Steve Wooledge, vice president of product marketing at MapR. “And we just see so many use cases of people doing really interesting things on MapR…We have customers today that have over 50 applications just on a single MapR cluster.”
Applications in the gallery include a real-time streaming application from Data Torrent that uses the YARN framework within Hadoop; HP Vertica Analytics Platform Community Edition, which provides real-time interactive SQL analytics; and Dataguise DgSecure, a data security, governance, and compliance app.
For end users, the gallery includes support information, installation instructions, documentation, and release notes.
“Based on what we are seeing with customers partners integrating with the MapR platform, we wanted to create a one-stop shop where people trying to get educated about how to use Hadoop in their enterprise can go and see examples and get ideas for different applications,” Wooledge said. “But also for developers and analysts, it’s a place to download different apps from partners that are certified against MapR so they can start to test out some of these use cases on our platform.”
Jack Norris, MapR Chief Marketing Officer, said that the user-friendliness of the gallery allows end users to locate apps by category and speeds time to value.
“People who might not be as familiar with Hadoop or the Hadoop ecosystem are able to take advantage of these and get implemented much faster,” he said.
“Hadoop is a wonderful platform for doing large scale analytics on all different types of data, as long as you have got the right people running it that know what to do with it,” said John Webster, Senior Partner at Evaluator Group. “And sometimes those people can cost a lot of money. So there has been desire from the enterprise side to say, ‘Look, can you give us something easier to use to manipulate and get value from Hadoop other than going out and hiring the expertise?’ So this app gallery starts to fill that hole.”
The app gallery also makes it easy for developers to submit their applications.
“We expect a lot of developers and other partners to be able to self-service and build their apps and make them available,” Wooledge said. “We have created documentation for how a partner can certify against MapR. We get those submissions and verify that everything works before it goes up into the app gallery.”
Jeffrey F. Kelly, Principal Research Contributor, The Wikibon Project, called the app gallery a good idea, but said that it needs to be adopted by partners who actually put useful apps in the gallery.
“Sounds like there are only a handful of apps as of the launch,” he said, “but if they can gain traction and populate the gallery with high-value analytics apps for data scientists and analysts, it could prove quite valuable.”
About 30 applications are currently available in the gallery, Wooledge said. “We have seen tremendous uptake,” he said. “I think there will be a lot of people who will want to port to it.”
In addition to the app gallery, MapR also announced Tuesday a partnership with big data software provider Syncsort that will enable data and processing to be offloaded from legacy systems into a secure, enterprise-ready Hadoop platform. As part of the new partnership, MapR certified Syncsort’s Hadoop ETL software, Syncsort DMX-H, with the MapR Distribution for Apache Hadoop.
“Syncsort does ETL work, data transformation, integration, but they have got some expertise in mainframes specifically,” Wooledge said. “They grew up in that world and understand the language that’s required, and they do a lot of specific sorting algorithms that help people process data out of the mainframe and turn it into formats that other systems can use.”
“One of the things that the enterprise wants to do with Hadoop is use it as this repository for lots of different kinds of data that exist in the enterprise,” said Evaluator Group’s Webster. “It’s just locked up in different data stores. They are all in current use in production, but the enterprise would like to be able to extract that data and load it into Hadoop, and that’s where Syncsort comes in.
“This basically builds a bridge to Hadoop that didn’t exist before.”
Moving workloads to a more modern data architecture allows organizations to avoid an expertise gap with older data management systems.
“Some of the concerns I have heard from our customers is they have got a lot of investment in the mainframe, but they want to get a lot of that data and those workloads onto a system that people are wanting to get current on, like Hadoop,” Wooledge said. “I don’t think the mainframe is going anywhere anytime soon. But there are certain jobs that can be easily run in Hadoop and that will free up cycles and get some of those apps updated and modernized.”
“Data is continuing to grow, and if there are upgrades that need to happen on the mainframe, maybe those can be deferred until later to bring the cost down but also to start to modernize the architecture.”
Norris pointed out additional advantages of moving data and workloads off mainframes.
“You have got a cost advantage,” he said. “You have got a scale advantage in terms of how that scales more easily, and you have a variety advantage in terms of once it’s on Hadoop, you can combine it with other data sources and use it more effectively.”
“The Syncsort-MapR partnership fits with the larger trend of offloading workloads to Hadoop,” said The Wikibon Project’s Kelly. “There are a number of workloads that reside in expensive data warehouses that can be offloaded to Hadoop, which results in two benefits: One, these workloads are much less expensive to run in Hadoop than an EDW [enterprise data warehouse] and two, this frees up capacity in the EDW for higher value analytic workloads. Syncsort has solid tools for moving workloads from the EDW/mainframe to Hadoop, and this partnership will help joint Syncsort-MapR customers accomplish this task faster, resulting in shorter time to insight.”