Big Data, Sensor Data and Business Analytics are accelerating the already rapid growth of unstructured data. To make matters worse these data sets often need to be retained in an online format for very long periods of time. The capacity and density of new hard drive technologies like the HGST 6TB Helium drive promises to allow enterprises to keep pace with this rapid data growth. But these high capacity drives are causing new challenges in terms of performance and recovery from errors. In the latest release of its software and hardware, Panasas may have the solution.
Panasas has been shipping scale-out NAS solutions since 2004. From the beginning, their focus has been on storing large, unstructured datasets. Initially this meant an emphasis on the high performance computing market, but now unstructured data is a problem for data centers of all sizes. As a result Panasas has matured their product offering so that it provides the enterprise features those data centers want.
In the latest release of their ActiveStor platform and PanFS file system Panasas has made its most aggressive move to address the needs of the more traditional enterprise. And with the influx of Big Data and Business Analytics these enterprises are looking for these kinds of solutions.
New ActiveStor Appliance for Fast, Big Data
Panasas’ ActiveStor scale-out NAS are now hybrid systems. But Panasas does more than simply ‘throw flash storage’ into the system, they use it intelligently for their target market. The flash drives are only used to store small files and metadata; it puts large files on very high capacity hard disks. Panasas is also one of the few companies shipping the HGST 6TB Helium drive in a system. Using this combination of flash and high capacity disk, allows them to deliver 50% greater density than previous systems, while offering better performance.
The ActiveStor hardware architecture is built on three components that work hand in hand to deliver optimal performance and reliability. The system has integrated 10Gb Ethernet switch modules to provide enterprise class connectivity. The storage blades are the hardware foundation of the system, storing files as objects. Each blade has two high capacity, cost effective SATA drives for high speed sharing. And as mentioned, ActiveStor 14 & 16 storage blades include an enterprise SSD for rapid access to metadata and small files. The Director blade orchestrates the file system activity outside of the data path, allowing I/O to occur in parallel, directly between compute clients and specific storage blades.
PanFS is the operating system for the ActiveStor systems. It enables direct parallel access to the high performance pool of storage. PanFS manages the scale-out cluster, automatically load balances data and provides data protection. The latest version of the Panasas File System now includes RAID 6+ data protection. This implementation is done on a per-file basis and is distributed across the cluster. This means there’s no data loss with two simultaneous storage blade failures and nearly instantaneous repair from most drive failures.
The growth of unstructured data is unrelenting and no CIO needs convincing of the enormity of the problem. The challenge is how to address that problem while still delivering high performance access to data. Higher capacity drives like the HGST 6TB Helium drive are ready for the enterprise, but data protection schemes like RAID 5 and RAID 6 are not. The rebuild time in these instances would be excruciatingly long.
Panasas with its file level driven, triple parity design is one of the first companies to deliver a file system that delivers the right balance of high capacity, high performance and high reliability that Big Data environments need.