What is NAS and SAN?

What is NAS and SAN? please give detail explain and how to choose and implement! Thanks!
Who is Participating?

[Webinar] Streamline your web hosting managementRegister Today

Lee W, MVPConnect With a Mentor Technology and Business Process AdvisorCommented:
Hi mandy9,



Which to choose depends on your purpose - if you are planning on using Exchange or SQL server or setting up a cluster, then the most appropriate is usually a SAN.  If you want serverless network storage (no need for a Windows or Linux Server) to place company user home folders and/or group shares (basically File Sharing), then you want NAS  What is your purpose?

NYtechGuyConnect With a Mentor Commented:

NAS is more "standard" and cheaper.  It is most common in an Appliance like a Dell Powervault or a SNAP! Server.  Basically disks locally attached to a controller of some sort - could be running Windows, or some proprietary OS based on linux, etc.  Users/clients/server are able to access the information over the ethernet/LAN network using shares setup in the OS of the device.

A SAN (Storage Area Network) is much more of an investment.  Essentially, there are a large number of disks attached to "San Controllers" via fiber connections.  There is a Fiber switch involved, which allows you to connect *MANY* servers "to the fiber".  You then, in the storage controller interface (usually web based, or using an application) setup what servers are "presented" which volumes.  Those volumes then appear as a locally connected disk in the server's disk management.

The SAN is much more versatile.

What are you trying to store (what type of files, and how much size total?)
Who is going to access it (PC,Mac, how many clients)
Lee W, MVPTechnology and Business Process AdvisorCommented:
Technically, you can get a NAS for a couple of thousand dollars or less - a little box that holds a single hard drive and plugs in to the network is TECHNICALLY a NAS.  HOWEVER, In my opinion, buying a NAS that costs less than HIGH 5 figures is pointless.  The EXPENSIVE NAS systems offer a variety of features including expandability and features like snapshots (akin to Windows 2003's Volume Shadow Copy).  Smaller units cost $$$ when you can just buy a RAID controller and a few drives for an existing server or even buy a new server to act as a file server that is cheaper/more expandable than a Snap server or Powervault storage unit.
dlmarioConnect With a Mentor Commented:
Hi Mandy9,

the main difference (technical) between Storage Area Network (SAN) and Network Attached Storage (NAS) is the level of access.

On a SAN you access blocks on a logical (of a RAID system) or physical (of a bunch of disks) volume. There you have features like "SCSI reservation", which is important for clustering for example. SAN uses the SCSI Protocol, so every SAN disk is like a SCSI disk for your OS. You need special Fibre Switches for SAN. Speed is 4, 2 or 1 GBit.

On a NAS you access the filesystem. You have no access to SCSI commands or Blocks. You cannot format a NAS Disk, because you only access the files in the filetree. The kind of filesystem is in a blackbox for the client of a NAS system. The protocol of a NAS is mostly CIFS/SMB (Windows Share), NFS (most for Unix/Linux) or FTP. NAS handles the filelocking for you - SAN does not lock files while concurrent access. NAS uses "normal" network components.

Another differenciator are the costs: SAN storage is more expensive than NAS, because SAN is most used in enterprise high availability environments (most for Clustersystems).

An third system is I-SCSI. There you have SCSI protocol over IP. This combines the blocklevel Disk access system of SAN whith the cheap network infastructure. But be aware, that Ethernet/IP/SCSI has more overhead than Fibre protocol - but you can realize clustersystems with little money.

Hope that helps...

DisorganiseConnect With a Mentor Commented:
NAS can be thought of as a file server replacement - you plug it in and voilla: a bunch of disk space available on the IP network.  NAS (generally) has the same kind of limitations as a traditional server, except you don;t need to 'manage' the OS - it's like a black box or utility.  In some respects its like plugging in a thumb drive to your PC to add more space, except you plug the NAS straight into the network.

SAN is something you plug servers 'into'.  SAN is method of separating hard drives from the servers: instead of having IDE or SCSI cables connecting your server to disk or array of disks, you have a fibre channel cable plugging into a switch, and from the switch to a rack full of disk.  SAN's rely on a 'fabric', which can be thought of as the SCSI equivalent of a traditional IP LAN.  By pooling the disks together and allowing them to be shared, greater utilisation may be achieved  (very simplified example: if you have a 36GB drive in a server and only use 10GB of it, the other 26GB is going to waste.  With a SAN, another server is able to use that 26GB - each 'partition' is usable by a different server.)
Since sharing drives between multiple systems inevitably introduces contention, SAN's usually employ large amounts of cache memory - the result is that most disk IO's write into cache and thus you get significant performance benefits.  With SAN's costing both arms and legs, you also get extra smarts like; snapshots, block-level replication, lots of sofware tools for monitoring performance etc
All Courses

From novice to tech pro — start learning today.