What Is Data Storage? Data Storage Types Compared
Learn what data storage is and which types of data storage architectures are best suited to accommodate different operational requirements.
In This Article
The digitization across sectors and greater reliance on digital mediums in recent years have accelerated Big Data production like never before. But where can businesses “park” this newly-minted raw intel?
With the advancement of new storage protocols and media, customers are faced with a myriad of choices.
What is data storage?
Data is the term for all digital entries. This can be an input (a specific query provided to the computer) and an output (the result, computed by the machine). Data storage, in turn, is the technology, architecture and processes computers rely on to collect, process and retain all this information.
Whether you copy a PDF document to a solid-state drive or codify network access permissions for hundreds of connected IPs, this information needs to be saved immediately and stored for the long term.
For short-term needs, computer systems rely on random-access memory (RAM). RAM functions like humans’ short-term memory. It allows machines to “remember” all aspects of the command while in operation.
But to retain and save bigger records, computers require long-term memory storage. That’s what different types of data storage solutions are meant to provide.
Data storage concepts
The ubiquity of data formats, record types and volumes has served as the springboard for different approaches to storing it. From personal data storage devices to limitless data center and cloud-based data repositories, you now have many options to organize and store your corporate data.
Direct-attached storage (DAS)
Direct-attached storage stands for all types of physical data storage devices you can connect to a computer. Portable and affordable — yet only accessible by one computer at a time — DAS is a standard solution for keeping small-scale records data backups or for transferring data between devices.
Popular types of direct-attached storage include external hard drives or solid-state drives (SSD), flash drives (USB sticks) and, although now dwindling in popularity, CD/DVD disks and other older methods.
Network-attached storage (NAS)
Network-attached storage (NAS) is a special hardware unit, featuring file-level architecture that can be accessed by more than one device as long as all users are connected to the internal network. In essence, a NAS unit features several storage disks or hard drives, processors, RAM, and lightweight operating systems (OS) for managing access requests.
NAS architecture is relatively simple to establish, and this is why many organizations rely on it to set up local storage systems for several users. Offering high accessibility, NAS units can be used for both data storage and local file sharing. You can also configure it as a data backup site by adding support for replicated disks or a redundant array of independent disks. NAS storage is often used as a synonym for unstructured data protocols like network file system (NFS), server message block (SMB) and object storage.
Storage area networks (SAN)
Storage area networks (SANs) help assemble an even more complex on-premises data management architecture that features two components:
- A dedicated network for data exchanges with network switches for load balancing
- Data storage system, consisting of on-premises hardware
The purpose of SAN is to act as a separate “highway” for transmitting data between servers and storage devices across the organization, in a bypass of local area networks (LANs) and wide-area networks (WANs). Featuring a management layer, SANs can be configured to speed up and strengthen server-to-server, storage-to-server and storage-to-storage connections.
For instance, you can set up a dedicated low-latency data exchange lane between a server running big data analytics workloads and a storage system (i.e., data warehouse) hosting the processed data. Doing so helps prevent bottlenecks and delays for other users on the LANs/WANs. The type of data storage that makes use of a dedicated SANs is typically defined as “block storage.”
Software-defined storage (SDS)
A software-defined storage (SDS) system is a hardware-independent, software-based storage architecture that can be used on any computing hardware platform. While NAS and SAN storage systems require you to use vendor-supplied OS and supporting software, SDS lets you bring your own license to any type of x86 server.
The two major benefits of SDS include:
- Scalability: You can reconfigure available hardware devices to accommodate the required data types and formats better and configure access for different types of networks or applications, using APIs.
- Total cost of ownership (TCO): Instead of purchasing another NAS unit compatible with your network infrastructure, you can re-deploy an existing one, saving on the TCO of your investment and avoid vendor lock in on the hardware.
Hyperconverged storage (HCS)
Hyperconverged storage allows you to virtualize on-premises storage resources to create a shared storage pool. Each hardware unit (node) gets virtualized and then connected in a cluster, which you then manage as a unified system.
The big boon of HCS is hardware consolidation and associated capital expenses (CapEx). Typically, 80 percent of data storage spending is tied to hardware. By using hyperconverged storage architecture and virtualizing some of your legacy servers, you can significantly reduce maintenance costs.
Additionally, many HCS solutions come with advanced file-management systems for optimizing data formats and volumes. This means your files can be packaged in more compact blocks to be dispatched over long distances. In that sense, HCS can also offer some cost benefits for WANs.
Cloud data storage
Unlike other options, cloud data storage assumes you will (primarily) use offsite storage of data in public, private, hybrid or multicloud environments that are managed and maintained by cloud services providers such as Amazon, Microsoft and Google, among others.
Unlike NAS or SAN, public cloud data storage doesn’t require a separate internal network — all data is accessible via the internet. Also, there are virtually no limits on scalability since you are renting storage resources from a third party that effectively offers an endless supply of servers.
But while cloud data storage assumes only operational expenses (OpEx), these too can add up without proper monitoring and optimization.
As with all the data storage options available, this approach works perfectly in some applications and presents drawbacks in others.
Check out the links below to learn more about how World Wide Technology can help your organization get the most out of data storage.