SAN, NAS and DAS: The inside story

Why Storage at all?

First of all, what do we mean by storage? Are we talking about Rent It storage, EZ Mini storage, or the many other personal and business self-storage commercial places used to store your excess furniture and books and toys? No! We are talking about information storage. Keeping it safe, retrieving it when necessary, saving it to protect it and being able to access it whenever, wherever to manipulate it for your business needs.

What is Information Storage?

So what does the ability to store data provide to a user or organisation? Very simply put, information is a valuable resource. It needs to be protected, be easily accessible, always be available and be saved for later use. This resource may be saved, transferred, communicated, restored and manipulated in a multitude of ways. What you don’t want is to lose it, have trouble finding it, spoil or corrupt it or, worse still, destroy it forever.

Types of Storage Media Devices

There are multiple devices that are used to store data. In the earlier days of IT, the most popular media forms included tape cartridges, disc drives and automated tape libraries. Of late, other forms have become more economical--including optical, CD, DVD and NAS appliances.

In Server Storage

In the beginning, there were servers--and lots of them. Each of these servers had their own built-in, dedicated storage. You could save a file to disk, you could backup that file, you could even restore it when the time came. This type of storage, however, did have its limitations. It would monopolise your server processor power when saving, backing up or restoring, and it was not easily shared with other users or servers on the network.

DAS--Direct Attached Storage

Direct attached storage--more commonly known as DAS-- and also referred to as Server Attached Storage (SAS). DAS is storage that is one step more evolved than ‘in server storage’. It is the storage equivalent of the missing link in anthropology.

Source: Juldee

DAS allows you to share storage with other users and devices. It utilises a dedicated storage server and contains a file system to manage the storage. A dedicated storage server allows you to free-up your application server from storage management concerns. It also allows you to share the storing and back-up of important data. There are however, a few drawbacks. Firstly, DAS is likely to be a bottleneck--if many servers or hosts are accessing the same DAS storage there can be performance issues. Secondly, data is still being accessed indirectly through a server that is not optimised for storage management.

NAS--Network Attached Storage

NAS was developed to address some of these limitations. Using NAS, the application server is absolved of the arduous task of having to look after storage tasks as required for In-Server Storage. Instead, a totally dedicated storage server or device takes on this responsibility. But unlike DAS, where there is a dedicated, high-speed link between a server and the storage device, the storage device is easily accessible by any server or user over the existing network--be it LAN or WAN.

Source: Juldee

In DAS systems, the file management software resides in the server accessing the storage device. For the NAS device to be successful it has to control its own destiny--its own file management system accesses, restores and backs up data as required for many servers, users and other resources (known as ‘hosts’) operating on the same network.

NAS, however, tends to run out of steam in larger enterprises where there is a need to manage large amounts of data, and there are stricter service level objectives that need a more mission-critical, robust approach to data storage management.

SAN--Storage Area Networks

Enter SANs. To understand SANs, you need working knowledge of Local Area Networks (LANs). LANs enable servers, clients and other resources to be shared amongst users over a small geographic region-- such as an office building. A LAN server usually contains applications and data that are accessed over a shared communication network--usually Ethernet--enabling speeds of up to 10 Mbps.

Source: Juldee

A SAN is an independent, high-speed network (up to 1Gbps) that is dedicated to executing storage-related tasks and handling storage network traffic. Without a SAN, the LAN or WAN would be used not only for network communications but also for tasks such as back-up, restoration, data migration between storage devices, or archival and retrieval of information. This would put an extra stress on the LAN/WAN and cause slowdowns or blockages that may affect business applications and end-user response times. Additionally, a SAN is transferring data at speeds of up to 100Mbps whereas a LAN is operating 10 times slower at 10Mbps. So an enterprise-wide back-up would negatively impact a LAN--in terms of performance--and disrupt normal operations. By using a SAN, that back-up can occur totally on the SAN.

In an environment where there are front office or back office applications executing heavy database updates, the SAN’s ability to offload storage tasks will result in low latencies and predictable OLTP performance.

SANs use existing technologies like IBM’s ESCON (using optical fibre) or Fibre Channel (FC), which is about three times faster than SCSI. FC is very flexible and allows devices to be as far apart as 10km--if using optical fibre as the medium--or closer together if using co-axial cable or twisted pair. FC is a transport service independent of protocol. It is an ANSI standard that allows the delivery of sustained bandwidth of over 97 MB/second for large file transfers, and tens of thousands of I/Os per second for business-critical database applications on a Gigabit link. This makes it the connectivity standard for storage networks.