File Storage and Block Storage Explained


File storage and block storage are the two main types of data storage commonly found on network attached storage systems. The terms “file” and “block” refer to how data is stored, managed, and accessed on storage media, such as hard drives, solid-state drives, or tape. File and block are access protocols that use different methods to save and retrieve data.

File storage is associated with unstructured data. Unstructured data is data that can be of varying lengths, with files that do not necessarily have common formats or characteristics that allow the data to be organized in a meaningful, cohesive, and consistent way. Some examples of file data include word processing documents, presentations, email messages, videos, and graphics.

Block data is not file-oriented, but rather consists of blocks of data that include databases and other forms of data that have specific structures that define each segment of data by its constituents. Block data is often represented in columns and rows. A row represents a single entity, while a column defines the content of that entity. The rows are generally called “records” and the columns, “fields”.


How does file storage work?

File storage and block storage are forms of data virtualization because they provide an organization and management layer that operates on top of the storage medium’s native management system.

File storage devices manage the data they hold centrally, using a file system interface, such as Windows File Explorer or Apple’s Finder utility on the Mac. The data that makes up a single file as recognized by a word-processing application, for example, may actually consist of multiple pieces spread across one or more drives. The file system maintains pointers that indicate the locations of these parts. The system can then allow the assembly of these parts in the correct order to present the complete file.

The file system also stores metadata for each file. Metadata is basic information that helps identify the file and includes file name, file size, date the file was created, and date it was last modified. Files are listed hierarchically in multiple levels of folders.

In addition to how data is stored on individual PCs, file storage is the main type of storage used for shared storage called network attached storage or NAS. NAS systems allow multiple servers and the users they support to access a defined share of a centralized storage pool.

File storage on NAS systems generally supports file access protocols such as Network File System (NFS), which is native to Linux and Linux applications, and Server Message Block (SMB) – formerly called Common Internet File System (CIFS) – for Windows servers and the applications they host.

How does block storage work?

Block storage does not overlay the inner workings of a drive to the extent that file storage does. It is more functionally tied to the underlying native disk management system that controls the storage medium. As such, block storage is not file-oriented and retains far less metadata for the storage it manages.

With block storage, users define volumes with blocks of storage capacity. To the operating system and server applications, a volume looks like a single disk. Blocks have a fixed size and are sometimes difficult to adjust once used. Therefore, when allocating a volume to an application, users should anticipate data growth and make appropriate allocations.

Block is the type of storage used in storage area networks (SANs), which are shared storage resources accessible by many servers and applications. SANs are similar in concept to NAS systems, but typically host only block storage. Originally, SANs required their own specialized network protocol, Fiber Channel (FC), which was developed specifically for network attached storage. FC networks use different network interface cards, switches, and other equipment than Ethernet networks, so SANs supporting block storage were usually built around their own specialized infrastructure.

In 2004, this restriction eased somewhat with the release of the Internet Small Computer Systems Interface (iSCSI) standard. ISCSI enabled shared block storage systems to use conventional TCP/IP networks so that the same Ethernet components that connected servers and users could be used to attach those assets to shared block storage. About five years later, another storage networking protocol – Fiber Channel over Ethernet (FCoE) – was introduced. FCoE has retained some of the performance aspects of FC but can operate over a standard Ethernet network.

What are the benefits of file storage?

File storage is familiar to anyone who has used a computer. It’s easy to use and easily accessible by a wide range of applications, including popular productivity apps like Microsoft Office.

As the basis of a shared storage resource – NAS systems – file storage offers relatively easy management with minimal administrative tasks required to make storage available to users and applications. And because the NAS can use existing Ethernet networking facilities to make storage available, it doesn’t require any special networking knowledge or specialized networking components.

What are the benefits of block storage?

Because it operates close to the native operations of the disks themselves, block storage typically performs at a higher level than file storage. This level of performance, combined with its block-based access to data, makes block storage particularly suitable for databases. It is also suitable for other applications that use large amounts of data that are increased and updated regularly.

In a disk drive block storage environment, even higher performance can be achieved by performing short strokes on the disk drives. This involves using only the outer edges of disks which move a clip faster than the inner parts so that data can be written and retrieved faster. All-flash SAN arrays and flash-assisted arrays are likely to produce even better performance for block storage systems.

What are the disadvantages of file storage?

The main disadvantage of file storage is that file systems are limited in the number of files they can handle. This number will vary greatly from system to system, but the commonality is that when the limit is reached, the system will not be able to store any more data.

File systems are hierarchical, containing levels of folders and subfolders that contain files that need to be tracked, managed, and updated with current metadata. By limiting the number of folders and files, the storage system can handle everything in a reasonable amount of time so that users and applications have access to data without major delays.

Its hierarchical nature also limits file storage performance. Although it is possible to run modest databases and other non-office productivity applications on file storage, the performance may be insufficient to handle more complex relational databases and related applications.

What are the disadvantages of block storage?

Block is generally more expensive than file storage. A network attached storage resource may require a special network to support Fiber Channel, unless the block storage is part of an iSCSI SAN using traditional Ethernet networking protocols.

Managing block storage in a SAN can be complicated and require more expertise than necessary for NAS administration. Although most SAN vendors selling block storage have simplified administrative procedures over the past decade, capacity allocation and overall management often remain complex.

Block storage also does not attach as much metadata to the data it stores compared to file storage, so other tools may be needed to identify and manage the stored data.

Unified storage: file storage coexists with block storage

Unified storage systems combine file and block storage into a single storage resource. Since iSCSI block storage can use Ethernet protocols just like file storage, it was possible to provide both types of storage in a single shared resource. Further developments made it possible to include FC and FCoE storage protocols in unified storage systems.

With unified systems, users can decide how to allocate storage capacity as a file or as a block. These systems are generally less expensive than block-only arrays, although block performance may not be on par with dedicated block storage appliances. But the flexibility to allocate resources as needed can outweigh the performance trade-off.

Conclusion: Use Cases Determine File or Block Storage

File storage and block storage have been around for a while. These are two mature technologies that meet the specific needs of data centers. This is true even as storage media have evolved from spinning disks to solid-state drives to flash memory. The choice of file versus block depends on intended use cases and performance expectations.


About Author

Comments are closed.