When given a choice, is it better to use network-attached storage (NAS) or direct-attached storage (DAS) for your...
environment? While there is not an answer that holds true in every situation -- since each organization has unique requirements -- here are five reasons why an organization might choose NAS vs. DAS. For the purposes of this article, I discuss hardware NAS and hardware DAS.
Reason 1: DAS has limited capacity
One of the primary reasons why some organizations select NAS over DAS is that DAS has limited capacity. As its name implies, DAS is directly linked to the server it uses. The disks may reside inside the server or in an external cabinet. In both situations, disks are connected directly to the server's storage controller. The problem is that a server can directly support only a limited number of disks.
NAS appliances also have a maximum number of disks they can physically accommodate. But NAS generally scales better than DAS. While small NAS appliances may only support a few disks, a large 4U NAS appliance might support dozens.
Reason 2: Flexible architecture (RAID, tiers)
NAS appliances tend to provide a greater number of storage architecture options. DAS can be configured as a RAID array, but the supported architectures are limited by the number of disks attached to the server and by the server's operating system (OS). As long as a sufficient number of disks are present, you can configure DAS to use some of the more common RAID levels and, in some instances, configure tiered storage. In most cases, NAS provides more architectural options. For example, NAS appliances may support some of the more exotic RAID levels that are not commonly available to DAS. Furthermore, DAS is limited by the number of drives that can be installed. Some NAS vendors allow multiple NAS appliances to be linked to achieve a level of scalability that is simply not possible using DAS.
Reason 3: Shared storage (file system only, impacted by virus or downtime)
One of the best reasons to implement NAS vs. DAS is that NAS is accessible to multiple servers. DAS is directly connected to a server and is considered to be a local resource to that server. DAS storage can be shared with others, but only at the OS level. An administrator might enable a file share that allows other machines to connect to the server and access a portion of its storage.
NAS is also accessible without the aid of a remote server's OS. While a NAS appliance has its own OS, it is usually a small, lightweight, purpose-built Linux kernel that is designed to be stable and reliable.
Imagine that a storage device needs to meet a rigid service-level agreement. In theory, DAS could be used. However, remote access to the storage and its data is dependent on the attached server's OS. If the OS is compromised by a virus, security breach or by the failure of its system volume, the attached storage may become inaccessible. Conversely, a NAS appliance is not dependent upon a server OS and is not prone to any of these problems.
Because DAS is generally shared at the file level, it is generally not suitable for use as clustered storage. Higher end NAS appliances may support iSCSI connectivity and block-level storage, which may make such an appliance suitable for use as clustered storage or for use in other situations in which shared storage is required.
Reason 4: Easier to configure
Although somewhat counterintuitive, it tends to be easier to configure NAS vs. DAS. NAS vendors often build wizards into their products that walk an administrator through the initial configuration process in just a few steps. Because DAS is directly attached to a server, an administrator must know how to connect the storage at the hardware level and be familiar with how the server OS provisions storage.
Reason 5: Redundancy
Because NAS appliances are purpose built, they tend to support a high degree of redundancy (especially on higher end models). It is relatively common for NAS appliances to support the use of extra disks and network adapters that can be used as hot spares. Some NAS appliances even have a redundant power supply. Many NAS devices also have a built-in replication mechanism that allows the appliance's contents to be replicated to a secondary NAS appliance to protect against an appliance-level failure.
It is theoretically possible to achieve similar levels of redundancy using DAS, but it usually requires an advanced knowledge of the server OS and the underlying storage hardware.
About the author: Brien M. Posey, MCSE, has received Microsoft's MVP award for Exchange Server, Windows Server and Internet Information Server. Brien has served as CIO for a nationwide chain of hospitals and been responsible for the Department of Information Management at Fort Knox. Visit Brien's personal website.
Which is better for your virtual environment: NAS, DAS or SAN?
When it makes sense to select DAS over NAS
Pros and cons of server-attached storage products