Nimbix and Mangstor have announced a joint technology offering designed to accelerate high performance data analysis for seismic processing, bioinformatics, and other big data analytics use cases in the Nimbix Cloud.
Today’s large-scale distributed applications demand tight integration between high performance compute and storage in order to process massive amounts of data quickly and cost effectively. To address this challenge, Nimbix has integrated the industry leading performance of Mangstor’s NMX-Series Flash Storage Array into its purpose built high performance computing cloud. The solution is powered by JARVICE, Nimbix’s Platform-as-a-Service for big data and HPC applications, and leverages an InfiniBand RDMA fabric to maximize throughput and minimize latency.
JARVICE delivers point and click simplicity to both application developers and end users. It is ideally suited for both parallel and distributed applications and specializes in workflow automation at scale. JARVICE incorporates Nimbix’s patented Reconfigurable Cloud Computing to achieve industry-leading performance and cost effectiveness for applications benefiting from bare-metal, heterogeneous environments. Applications can leverage the newly integrated Mangstor flash storage as high performance block, object, or file, and can scale into the petabytes. While targeted at systems based on Hadoop, Cassandra, and other distributed platforms, the new solution can also accelerate traditional HPC use cases such as high-end simulation and rendering.
Mangstor’s NMX-Series uses RDMA network access to bring the low latency benefits of NVMe to a shared storage environment. The arrays are based on Mangstor’s Software Configurable MX6300-Series SSDs in combination with Mangstor’s TITAN storage stack. NVMe over RDMA read and write access to the array provides servers with data at nearly identical latencies as if they were accessing local PCIe SSDs. This allows compute clusters to realize the low latency application benefits of PCIe SSDs without having to put an SSD in every server. Typical configurations deliver millions of IOPS and throughput starting at 14 gigabytes per second, with near linear scaling as arrays grow.
“Optimizing high performance computing workflows is all about reducing overhead and eliminating bottlenecks,” said Nimbix CTO Leo Reiter. “With Mangstor’s NMX-Series, big data applications can finally unlock their true potential while taking advantage of the automation and ease of use of the JARVICE platform.”
”Mangstor’s mission of delivering unparalleled storage performance, at Web scale, to the most demanding of high performance use cases, is perfectly complemented by Nimbix’s innovative and scalable high-performance workload orchestration and management capabilities,” said Trevor Smith, Founder and CEO of Mangstor. “I am excited about the potential these two great technologies have to uniquely serve the market for powerful, flexible and easy-to-use high performance computing on-demand.”
The new solution is available on the Nimbix cloud as a technology preview, and is scheduled for general availability in late Q3 of 2015. Pricing plans will depend on scale and be offered as pay-per-use or subscription. The solution also supports hybrid and private cloud delivery models, including JARVICE-based on-premises deployments.
Bigdata and data center