While the product's name may need a little work, DataDirect Network's new SFA12K series storage array represents a new high-water mark for networked storage performance, according to analysts.
With up to 160 6Gbps serial-attached SCSI (SAS) parallel I/O lanes attached to either SAS, SATA or solid-state drives (SSDs), the appliance can offer up to 40GB/sec performance, or 1.4 million I/Os per second (IOPS).
"This is a highly parallelized storage engine that talks through 160 channels at the same time," said Jeff Denworth, vice president of marketing for DataDirect. "That means you can use cost-effective SSDs versus enterprise-class SSDs or more expensive PCIe cards."
The arrays come with 64GB of memory, and with InfiniBand and Fibre Channel connectivity the SFA12K-40 is designed to scale file storage systems to 1TB/sec and beyond with as few as 25 disk arrays. It is DataDirect's highest performing array.
DataDirect calls its SFA12K a "Big Data" appliance because of its ability to create a single domain name space file system that holds up to 6.7 petabytes of data in two racks - that's 60 drives per 4U (3.5-in. high) enclosure. It can also host applications on the array, which allows it to run virtual machines on the storage array's controller or run an instance of Apache Hadoop.
"You can serve up 16 virtual machines running at 96% utilization on this appliance," Denworth said.
The SFA12K can hold up to 84 drives in a 4U (7-in. high) enclosure, 840 drives in a rack, and up to 20 enclosures can be managed behind a single interface.
"So, we get to dilute the cost of up to 1,600 drives, which means you can manage them for well under $1 per gigabyte," Denworth said. "It's a big system, and we recognize we're not competing with smaller storage vendors. Our first two orders were for a 16PB [array] and a 35PB [array] configured as single file system deployments."
EMC is also planning to announce that its arrays can run virtual machines and applications .
There are two versions of the new array, the SFA12K-40 and the SFA12K-20E. Both use DataDirect's SATAassure RAID scheme, which offers double parity drive production, meaning up to two drives can fail at the same time without data loss.
The SFA12K-20E uses DataDirect's In-Storage Processing capability, first announced in 2009, to host the company's ExaScaler and GridScaler parallel file systems and/or customer applications natively within the storage array. ExaScaler is for clustering up to 10,000 Linux-based clients; GridScaler supports the Network File System (NFS), or Common Internet File System (CIFS), and and connects up to 2,000 clients.
The SFA12K-20E delivers 20GB/sec of bandwidth and 700,000 IOPS when populated with SSDs. Because Windows and Linux applications can run directly on the arrays, they can lower networking and server costs, DataDirect said.
Denworth said a typical configuration of the SFA12K appliance costs about $300,000, but the products can retail for more than $1 million if loaded with the maximum of 1,600 SSDs.
Lucas Mearian covers storage, disaster recovery and business continuity, financial services infrastructure and health care IT for Computerworld. Follow Lucas on Twitter at @lucasmearian or subscribe to Lucas's RSS feed . His e-mail address is [email protected] .
Read more about virtualization in Computerworld's Virtualization Topic Center.