File based storage technology has long been a popular option for the storage of unstructured data. File storage builds on top of a block storage layer and stores data as files organized logically as a hierarchical structure of directories. There are different ways of storing files, and the hierarchical interface does not necessarily need to match the actual storage layout, however it is this logical interface that creates the basis of file storage. Access of file data can be either to a local filesystem or to a filesystem mounted over a network in the case of NAS systems.
Usage of file-based storage includes a wide range of workloads. It has traditionally been a popular choice for tier 2 workloads, backup targets, and content repositories. As unstructured data has become more prevalent in tier 1 and performance demanding workloads, file storage has become a popular solution for these cases as well.
With a broad range of workload requirements, file storage solutions have also developed a broad range of capabilities for areas such as capacity, performance, and scalability. Two common file storage technologies are Network Attached Storage (NAS) systems, which provide remote file access over NFS or SMB, and Scale-Out File Systems which provide highly scalable, parallel file access via a POSIX compliant client. While the two technologies are closely related, they have distinct characteristics, capabilities, and use cases. This paper looks to distinguish the two technologies and identify scenarios in which they are best utilized.
Network Attached Storage, or NAS, refers to storage appliances that offer remote file storage accessible over a network. NAS systems typically involve both the hardware and software elements necessary to provide remote file storage – including storage controllers, network interfaces, storage devices, native file systems, and customized operating systems. NAS systems are connected over Ethernet networks with file access provided by a standard file protocol such as NFS or CIFS/SMB.
NAS systems are often dual controller systems with an active/active architecture for reliability and scale up with expansion capacity. Scale-out NAS systems are also available with data distributed across nodes and accessible through a global namespace. Scale-out systems are capable of scaling both performance and capacity simultaneously, and typically provide data availability through an N+1 failover scheme. In most cases there is some limitation to their scalability. Examples of scale-out NAS systems include Dell PowerScale, NetApp AFF, Pure Storage FlashBlade, Quantum Myriad, Qumulo, and VAST Data.
Download this free Technical Insight report to continue reading!