MapR Platform Offers Persistent Data Access for Containerized Applications
This is the container age. The growing use of services like Docker is transforming the way that software is being handled within enterprises. However, this rise in container utilization does throw present problems for enterprise CIOs when it comes to rolling out applications in production.
One of the biggest issues is scaling applications to meet business demands. While, in theory, containers are able to handle enterprise applications, in practice, they are often hampered by day-to-day disruptions: network failures, server breakdowns or even scheduled maintenance, consequently, organizations have tended to play safe and use containers for stateless web applications, rather than try to overcome these storage issues, indeed some analysts have warned companies to be wary of using stateful application.
This is a problem that data analytics specialist, MapR has been struggling with. For companies to enable truly powerful analytics, this separation between storage and computing needs to be bridged, enabling containers to be deployed.
Dale Kim, senior director, product marketing at data analysis platform company MapR Technologies, said that companies have been looking at ways to introduce containers within enterprise infrastructure without running into scalability issues. “There are two approaches. First is to use existing storage and add containers. If there’s a small amount of data, this approach is OK but if you have large volumes then you’re going to run into difficulties. The second approach is to add an orchestration layer on top of the storage, which doesn’t address the complexity.”
MapR claimed to have cracked this issue by enabling instant access of data from any resource within the infrastructure. According to Jack Norris, MapR’s senior vice president, the system would enable a single, converged system for data analytics. He said that it would mean that companies would no longer need to have clusters talking to SANs and then moving the data to another system to handle the analytics.
The new MapR Converged Data Platform for Docker enables fast access of any data from any underlying infrastructure resource.
“With the introduction of MapR Persistent Application Client Containers (PACCs), containerized applications can easily leverage all the MapR platform services (MapR-FS, MapR-DB, MapR Streams) as a persistent data store,” the company’s website asserts.
This container-optimized environment includes a range of features to help companies deploy Docker more easily: these include a pre-built Docker container and secure authentication at the container level.
One of Map R’s customers who is looking forward to the new platform is data analytics company, Quantium, which has 500 analysts running Docker for their customers’ time-critical applications. According to Gerard Paulke, an enterprise architect for Quantium, the use of the MapR data platform will maintain a service for the customers. “At the moment, if we lose a container then we’d have to replicate our Docker volumes and start up nodes again.”
He said that the new offering will help the company to ‘Dockerize’ components enabling to spin up clusters more quickly. It will also help Quantium improve its services to its customers by providing a greater range of options. “We have a lot of external clients who want to use our infrastructure, the new platform will enable us to offer them a space with a fully containerized file system or, if they want, to spin up containers in the cloud.
Norris said the introduction of the platform could transform organizations. “The conversations will be completely different: there will no longer be a discussion on whether something is a cloud or an on-premise app. The conversations will no longer be about where an application sits but will be around issues like cost, latency and regulation.”
He added that it will also help companies plan better. “They’d be able to start off with a few containers and expand with the data. It’s what people have been waiting for.”
More info here: