BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
DataCore Software enhanced its storage virtualization software to support hyper-converged systems and OpenStack cloud storage with the release of SANsymphony-V10.
SANsymphony-V10 also adds support for virtual desktops and makes available a production-ready version of its Random Write Accelerator (RWA) software that reorganizes how writes occur on disk.
DataCore SANsymphony-V is designed mainly for storage in Microsoft Windows Server environments. The software presents itself to storage arrays as a standard Windows driver. DataCore runs either on x86 servers in a storage area network or as a software-only instance for virtualizing shared storage, which the vendor licenses as DataCore Virtual SAN.
Do-it-yourself features for hyper-convergence with Virtual SAN
SANsymphony-V 10 includes templates for customers to cobble together a hyper-converged system using their existing storage hardware and a two-node configuration of DataCore virtual storage appliances. The DataCore software analyzes a customer's hardware and software stack to automate networking, high-availability clusters, redundant failover copies, storage pools and scale-out file systems, based on the required service level.
Augie Gonzalez, DataCore Software director of product marketing, said DataCore SANsymphony-V10 provides an alternative to purchasing integrated hardware from hyper-convergence vendors.
"We're offering a different way to fashion a hyper-converged system with your choice of storage server hardware, rather than having to buy the whole kit. Most hyper-convergence vendors require you to buy three of their nodes and often recommend a minimum of four. We can do hyper-convergence with two DataCore nodes," Gonzalez said.
There are other software-only hyper-converged options, including VMware's Virtual SAN (VSAN), hardware pre-certified by the software vendor.
OpenStack Cinder now supported, RWA makes it to production
Gonzalez said DataCore's OpenStack integration enables customers to build cloud storage without buying OpenStack-compatible storage gear. DataCore software serves as an interface between Cinder drivers and back-end storage to automate self-service provisioning and management based on a user's quality of service requirements.
DataCore introduced the RWA for experimental use in May 2014. The technology converts random writes on hard disk drives (HDDs) into sequential writes to boost performance for write-heavy workloads, including databases, ERP and online transaction processing systems.
Rather than wait for the actuator head to rotate back to the original write sector, RWA writes updates to available disk sectors and places a pointer to the file, which it subsequently uses to clean the original sector for reuse.
"We're doing the I/O sequentalization of those random writes to make it very easy for HDDs to absorb the load," Gonzales said.
Jon Toigo, CEO and managing principal at Toigo Partners International LLC, said the RWA's novel architecture has the potential to reduce the need for log restructuring as a method for managing randomized I/O requests in write-intensive applications.
"I like the sanity they bring to the whole architecture. This the first real application besides log restructuring to solve random writes," Toigo said.
Persistent VDI for Hyper-V, 'universal' deduplication and Veeam backup integration
DataCore SANsymphony-V10 features automated tools to support persistent virtual desktop infrastructure (VDI) applications running in Microsoft Hyper-V. DataCore provides server-side RAM cache to reduce backend I/O traffic and makes use of native Hyper-V features for failover clusters and disk differencing. The vendor said can support up to 500 virtual desktop configurations.
Gonzalez said customer demand for space reduction spurred DataCore to incorporate tools for deduplication and compression of virtual disks. DataCore's Universal Disk Deduplication and Compression Services extend data reduction to other vendors' storage arrays. DataCore performs parallel post-processing deduplication at the block level and compresses a single data instance for synchronous mirroring between primary and secondary storage devices.
Previous versions of DataCore SANsymphony-V supported Commvault Simpana 10 and Microsoft Data Protection Manager backup software. Version 10 adds APIs to Veeam Backup and Replication software for VMware vSphere snapshots on DataCore nodes. Two simultaneous VM snapshots are taken: A server-based snapshot by Veeam and a snapshot by DataCore on one of its virtual nodes. Veeam snapshot requests are directed to DataCore to give applications full use of host resources.
DataCore chairman says SANsymphony canboost SSD performance
DataCore optimizes its SANsymphony-V with flash
Electric company solves performance problems with SANsymphony-V