What's the difference between SAN and NAS?

[Garacesh]

Making revision notes for the A+.. Confused by SAN vs NAS (but don't want to start a thread about it, since I'm sure it'll only take one post to explain it)

 

Am I right in thinking that a NAS is a big box full o' hard drives, and a NAS is a bunch of boxes full o' hard drives working together? (kinda like RAID, but with servers rather than hard drives?)

Edited June 13, 2016 by Garacesh

[Boredguy]

NAS is network attached storage, SAN is Storage Area Network.

 

Both have drives, but SAN's are often better than NAS

 

 

Edit

NAS = File level access, SAN = Block level access

Edited June 13, 2016 by Boredguy

[pcstru]

Making revision notes for the A+.. Confused by SAN vs NAS (but don't want to start a thread about it, since I'm sure it'll only take one post to explain it)

 

Am I right in thinking that a NAS is a big box full o' hard drives, and a NAS is a bunch of boxes full o' hard drives working together? (kinda like RAID, but with servers rather than hard drives?)

 

NAS tends to offer end user file and file system management (either via LDAP or local user management). A SAN tends to be block based via iSCSI. These days a lot of devices straddle both functions. For a SAN, the storage devices should be on their own network, so communication with the device is between the server and then device on a network, while users access their files via the server on a separate network.

[korifugi]

NAS: Network Attached Storage - Networked 'box' of disks.

 

Operates at file level - allowing user level access.

 

 

SAN: Storage Area Network - One or number of 'boxes' of disks.

 

Operates with multiple servers as if it were local storage to all of them. No user level access.

[Garacesh]

NAS is network attached storage, SAN is Storage Area Network.

 

Both have drives, but SAN's are often better than NAS

 

I know what they stand for, it's the functional differences I'm interested in

We use a NAS here, which is where I get my understanding that it's just a big box o' hard drives combined into a large logical disk, using RAID with a few hot spares for redundancy. It's the NAS I'm confused about, functionally. Google and Wikipedia are failing me with explanations. A network is a bunch of devices connected together, right? So the network itself can't be used for storage, only the devices on the network. So is it a bunch of SAN's (in effect) networked together working as one logical server, or...?

 

Or am I getting too bogged down here by the etymology? I think it's the use of the word 'Network' that's causing me the confusion.

Edited June 13, 2016 by Garacesh

[korifugi]

It's the NAS I'm confused about, functionally

 

Think of a NAS as a separate external HDD - just one that's networked. It's seen as an external device by everything, but can be set up with redundancy like any other storage device.

 

A SAN can also be set up with RAID etc. - but can be set up to be seen as local storage by servers.

[sted]

Think of a NAS as a separate external HDD - just one that's networked. It's seen as an external device by everything, but can be set up with redundancy like any other storage device.

 

A SAN can also be set up with RAID etc. - but can be set up to be seen as local storage by servers.

with iscsi so can nas boxes

 

honestly the difference seems to be $$$ and to some degree throughput

[Oaktech]

As I understand it a NAS is storage attached to your normal network which can be accessed by any client devices with the correct credentials via the normal network. a SAN is server-network-client arrangement in which a storage device is accessed via a dedicated network by a limited number of clients normally servers.

 

In a NAS the access arrangements are usually controlled by the NAS device (although many have AD intergration) in a SAN the server OS controls access.

 

A NAS works well in a mixed OS environment, a SAN better in a single OS.

[strawberry]

Think of a SAN like a normal hard disk, but instead of a sata cable you use a network cable to access it ( probably fibre just to muddy waters)

Think of NAS like a network USB stick. You just plug it in and access files.

[Garacesh]

Think of a SAN like a normal hard disk, but instead of a sata cable you use a network cable to access it ( probably fibre just to muddy waters)

Think of NAS like a network USB stick. You just plug it in and access files.

 

Eep, maybe I should've started a thread...

That description makes them both sound identical to me. It's a storage device* that is accessed over a network connected (which in essence they both are if you're only concerned about end-results), but it's how they work I'm getting confused by (and the questions I'm working through are asking about it, so I assume I'm going to need to know about them when I sit the actual A+ exam)

 

* In this case, server rather than single hard disk, but the point still stands.

 

Edit: So.. a NAS would be connected to a server (or multiple servers) as if it were a local drive (even though it's connected over the network), whereas a SAN is mounted as a network share (such as \\FileServer.domain.local) and the operating system knows it's a server elsewhere?

Edited June 13, 2016 by Garacesh

[Norphy]

 

Edit: So.. a NAS would be connected to a server (or multiple servers) as if it were a local drive (even though it's connected over the network), whereas a SAN is mounted as a network share (such as \\FileServer.domain.local) and the operating system knows it's a server elsewhere?

 

 

No, you've got that the wrong way around.

 

A NAS is effectively a server which does storage and nothing else. It'll most likely run a derivative of a bog standard operating system (Windows, Unix, Linux, BSD) and present itself using standard network file sharing protocols (SMB, CIFS, AppleTalk, NFS, etc). These are generally used for file shares and suchlike.

 

A SAN is a storage device which has a certain amount of storage. That storage can be carved up into chunks (LUN - Logical Unit Number). The SAN uses a networking technology of some description (Ethernet or Fibre Channel usually) as a transport medium rather than USB/SATA/SAS/whatever. A server will either have a piece of software to connect the SAN (An iSCSI initiator probably) or a hardware driver (More common with fibre cards). Once the server is connected to the SAN, the LUN is then presented to a server as local storage. Multiple devices can connect to the SAN at once and have their own LUNs allocated to them or, depending on the Server OS and application, multiple servers may use the same LUN. These are usually used for clustered applications such as databases, virtualisation stores and similar

Edited June 13, 2016 by Norphy

[elsiegee40]

Eep, maybe I should've started a thread...

Thread split!

[3s-gtech]

Some devices can also act as SAN and NAS at the same time - we have a Synology box which holds backups (file level), but chunks of it are provisioned off into LUNs which our servers access over iSCSI for block level stroage.

[Garacesh]

No, you've got that the wrong way around.

Derp, I meant to type them the other way, I just got my acronyms mixed up.

 

A NAS is effectively a server which does storage and nothing else. It'll run a derivative of a bog standard operating system (Windows, Unix) and present itself using standard network file sharing protocols (SMB, CIFS, AppleTalk, NFS, etc). These are generally used for file shares and suchlike.

 

A SAN is a storage device which has a certain amount of storage. That storage can be carved up into chunks (LUN - Logical Unit Number). The LUN is presented to a server as local storage. The SAN uses a networking technology of some description as a transport medium rather than USB/SATA/SAS/Whatever. Multiple devices can connect to the SAN at once and have their own LUNs allocated to them or, depending on the OS, use the same LUNs. These are usually used for clustered applications such as databases, virtualisation stores and similar.

 

That helps carve out the distinction between the two. So I'm guessing that SAN's have their own custom O/S (probably some weird and wonderful fork of Linux, knowing how things usually go..) whereas a NAS will match operating systems with whatever clients access it (if you had Windows clients you'd use Windows Server to host your NAS for example) and uses the basic permissions as access control? (NTFS, in the case of Windows). With a NAS you can browse to the folder over the network (or, in the case of Active Directory, have a portion of the NAS mounted at logon) whereas a SAN mounts a portion of its files (via the LUN) to a server as a DAS, even though it's over the network?

 

Some devices can also act as SAN and NAS at the same time - we have a Synology box which holds backups (file level), but chunks of it are provisioned off into LUNs which our servers access over iSCSI for block level stroage.

NOT HELPING lol

I did imagine that it would be possible for devices to act as both though, just to confuse me even more. It's a conspiracy!

Edited June 13, 2016 by Garacesh

[Willott]

Hopefully what everyone else has said has answered, but I'll add my bit

 

SAN - is presented to the device accessing it as a raw hard disk, so you can then partition/format etc as required

NAS - is presented as an already formatted file system so files can be placed on it

 

As others have said, the definition is now much more muddied - for example, you used to be able to say that SAN was used in back end storage for virtualisation, but now SMB3 and NFS are used for storage (which are NAS protocols - they are used for file transfer). And SAN wise, you used to have fibre networks, but then iSCSI was introduced which allowed use of LAN/ethernet, and now FCoE.

[JackCPickup]

Comparing SAN and NAS | SMB IT Journal

 

This guy's a bit of a storage expert.

[Michael]

A NAS is an independent server, accessible over the network (Ethernet).

 

A SAN is an extension of your server's own storage, connected to the server itself externally. That's all there is to it.

 

So for example, your server may only have 2TB, but you could add a 12TB SAN as a separate volume without having to 'mess' considerably with your server.

 

NASes are accessed typically by mapping a network drive, so you drag/drop files as you would normally.

[Norphy]

That helps carve out the distinction between the two. So I'm guessing that SAN's have their own custom O/S (probably some weird and wonderful fork of Linux, knowing how things usually go..)

Most likely an entirely custom OS developed by the SAN manufacturer. These things have to be hard and I don't think that *NIX derivatives are suitable

 

whereas a NAS will match operating systems with whatever clients access it (if you had Windows clients you'd use Windows Server to host your NAS for example) and uses the basic permissions as access control? (NTFS, in the case of Windows).

 

No, not necessarily. As long as the NAS can speak the appropriate networking protocol, there's no need to match OS. Most *nix based NASes will talk SMB/CIFS, NFS and AppleTalk. The file system on the NAS is also irrelevant, as long as there's a common source (such as AD) where it and your clients can get user information from, you'd be able to set permissions.

 

With a NAS you can browse to the folder over the network (or, in the case of Active Directory, have a portion of the NAS mounted at logon)

 

Correct

 

whereas a SAN mounts a portion of its files (via the LUN) to a server as a DAS, even though it's over the network?

 

A server connects to a SAN over a network, not the network. Even with iSCSI, you should always have a separate network to your main one to avoid contention issues and having some physical separation improves security too. At the very least, it should be on a different VLAN.

 

The driver on the fibre card or the iSCSI initiator presents the LUN to the operating system as local storage, yes. Just as the driver for your SCSI card would present a SCSI hard drive as local storage.

Edited June 13, 2016 by Norphy

[3s-gtech]

NAS = files available over a network

SAN = storage available over a network (obviously not what it stands for!).

 

While a NAS can be a SAN, generally a SAN can't be a NAS as it doesn't run an OS with file level access. In a way, NASs are more sophisticated, but of course SANs tend to be bigger, more flexible, have more disks, and often have fibre channel connections.

[Rob_D]

whereas a NAS will match operating systems with whatever clients access it (if you had Windows clients you'd use Windows Server to host your NAS for example) and uses the basic permissions as access control? (NTFS, in the case of Windows). With a NAS you can browse to the folder over the network (or, in the case of Active Directory, have a portion of the NAS mounted at logon) whereas a SAN mounts a portion of its files (via the LUN) to a server as a DAS, even though it's over the network?

Tangent I know. But, once you've got them up and running, NTFS shares can be fairly OS agnostic. We map Macs to shares on a Windows server here, and my home NAS runs some-kind-of-linux and talks to my windows machines fine.

 

I'd be inclined to think of a SAN being more part of the hardware layer, kind of like a raid card in a different box.

While a NAS is a complete system on its own so any connections to it are on more of a software layer.

[Garacesh]

A server connects to a SAN over a network, not the network. Even with iSCSI, you should always have a separate network to your main one to avoid contention issues and having some physical separation improves security too. At the very least, it should be on a different VLAN.

 

So the SAN exists separately to the main network?

[Norphy]

So the SAN exists separately to the main network?

 

Ideally, yes. You can run an iSCSI network on the same network segment as your generic one but I wouldn't recommend it. Mostly for the contention and security concerns I listed above.

Edited June 13, 2016 by Norphy

[pcstru]

So the SAN exists separately to the main network?

Yes, in an ideal world. Our iSCSI devices are all on a physically distinct network from the client devices. Servers straddle a plug (or 8!) into both.

[Garacesh]

So is a SAN used more where it's a permanent fixture (as in, the same result could be created by slapping another disk in a server)?

 

Databases, for example, were mentioned earlier. I'm guessing a SAN wouldn't be used for, say, user home folders, but might be used to ensure a server could always access the database (and users would run whatever software needed the database from the server?) rather than being used for users to access files?

 

Or could it be used for user home folders by cutting the drive into smaller chunks and assigning each user their own individual LUN?

 

If the SAN exists separate from the main network, what bridges the gap?

[DJ-1701]

Use of a separate network also of course helps mittigate against direct data attacks on the SAN and freezing up multiple services.