Software-defined storage products: 2017 Products of the Year finalists
The 11 software-defined storage finalists in this year's competition highlight the most recent developments in flash, cloud and container technologies.
The software-defined storage products selected as finalists in Storage magazine and SearchStorage's 2017 Products of the Year competition reflect the latest trends in flash, cloud and container technologies.
The finalists include a mix of new and updated software-defined storage products, and many support performance-boosting nonvolatile memory express (NVMe) flash technology and extended cloud and container storage options.
To be eligible in the Products of the Year competition, a software-defined storage product must run on standard servers with no dependencies on the underlying hardware. Finalists in the category include software that pools and centrally manages storage, file systems, object stores and software powering hyper-converged infrastructure.
Software-defined storage products that made it through the finalist cut include the following:
DataCore Software SANsymphony PSP6 Update
The PSP6 upgrade of DataCore Software's SANsymphony storage virtualization product enhanced REST API integration with more than 200 new operation methods. It also lets customers use the Kubernetes PersistentVolume API to orchestrate deployment, operations and scaling of containerized applications. The update optimized the product's parallel I/O capabilities to improve performance.
Elastifile Cloud File System
The distributed Elastifile Cloud File System (ECFS) can pool server-based flash storage and present it to applications through a global namespace, enabling any server in the cluster to directly access files. ECFS uses a distributed metadata model and enables linear performance scalability. The product supports global deduplication, compression and snapshot shipping, and its CloudConnect feature lets customers tier colder data to cloud-based object storage.
Excelero NVMesh 1.1
Excelero's NVMesh software virtualizes NVMe-based storage, pools the capacity of the SSDs and uses Remote Direct Memory Access networking to connect them. Any Linux-based host server running the NVMesh block client can access the virtual block volumes to enable applications to get the latency, throughput and IOPS of local NVMe. Excelero's Remote Direct Drive Access technology disaggregates storage from applications to save server CPU for the applications.
Hedvig Distributed Storage Platform 3.0
The Hedvig Distributed Storage Platform supports scale-out block, file and object storage on premises and in public clouds, and manages the storage as a single pool. New FlashFabric technology adds tiering capabilities between two different classes of SSDs, whether SAS or SATA, NVMe-based PCI Express (PCIe) or 3D XPoint. Other new features include AES 256-bit software-based encryption for data in use, in flight and at rest, and CloudScale plugins for Red Hat products, Veritas OpenStorage Technology and VMware vSphere.
IBM Spectrum Scale 4.2.3
IBM enhanced its Spectrum Scale clustered file system -- based on the company's General Parallel File System -- in areas such as cloud data sharing, NVMe, audit logging and general usability. IBM also beefed up support for iSCSI, Hadoop Distributed File System and OpenStack object storage.
Pivot3 Acuity HCI Platform 2.1.1
Pivot3's Acuity hyper-converged infrastructure platform added policy-based management with quality of service to enable customers to provide more resources for high-priority applications. The software update also enables the system to take greater advantage of NVMe-based PCIe flash storage to reduce latency and improve performance.
Qumulo File Fabric
Qumulo File Fabric (QF2) runs on premises and in the public cloud, and it claims to provide billion-file capacity. QF2 gives customers programmatic access to features and administrative settings through a programmable REST API and includes cloud-based monitoring to proactively detect disk failures. Additional capabilities include real-time analytics down to the file level, snapshots to the directory level and continuous replication.
Red Hat Ceph Storage 2.3
Red Hat's commercially supported open source Ceph distribution added a lightweight NFS interface to the Ceph Object Gateway, letting users access the same object storage data via NFS or Amazon's Simple Storage Service (S3) API. Compatibility with the Hadoop S3A file system client enables big data analytics applications -- such as Apache Hadoop MapReduce, Hive, Presto and Spark -- to read and write data from the Ceph object store. Red Hat Ceph Storage 2.3 also introduced an option to deploy Ceph in containers.
Scality Ring7
Scality stood out among the software-defined storage products for steps it took to unite its scale-out file system and object store, and enable users to access the same data from either protocol. Ring7 supports versioning and replication for object and file storage; data encryption; secure write once, read many (WORM) change control; and cross-region replication to Scality Ring and Amazon's S3 cloud object storage for disaster recovery.
SUSE Enterprise Storage 4
SUSE's Enterprise Storage 4, the company's commercially supported Ceph distribution, added support for the Ceph file system, or CephFS, to round out its unified block, object and file capabilities. The software update enabled long-distance replication for block storage and multisite object storage replication to improve disaster recovery. SUSE also enhanced cluster orchestration through open source Salt configuration management software, improved the graphical user interface to ease management and expanded hardware choice with support for 64-bit ARM processors.
WekaIO Matrix
The WekaIO Matrix scale-out parallel file system is designed to pool server-based flash SSDs for high-performance workloads and tier cold data to object storage in public and private clouds. Matrix presents the local flash storage as a single global namespace to host applications, and the software extends the namespace beyond the flash pool to offload colder data to less expensive S3-compliant and OpenStack Swift-compliant object storage.
