The Difference between NAS and SAN

The Difference between NAS and SAN

When you are looking for a local storage for business, there come two options: NAS (Network Attached Storage) and SAN (Storage Area Network). Today we will find out which one fits best to your needs.

NAS vs SAN – the Basic Difference

First, let’s figure out the essence of these two technologies. SAN storage is generally an array of disks, which are attached to the server via a special network. It is a block-level system with full server’s access to the disk volume using such fast protocols as Fibre Channel, Fibre Channel over Ethernet (FCoE), iSCSI, etc. That means that a client (for example, a server) detects SAN as a local drive, thus manages filesystem volume directly. Roughly speaking, Storage Area Network device is an external hard drive, which connects to the computer.

In traditional implementations, LUNs can be accessed only by a single server. It is possible, however, to allow Logical Units to be shared, but that requires additional input. This is why business often uses SAN alongside with special access managers, e.g. cluster services, which use clustered file systems and supervise read/write access rights at the every instant of time. The management of the IT-infrastructures with SAN requires a knowledge of low-level block protocols and their hardware and software medium, such as FC switches, optical cables, SCSI-powered protocols, etc.

Typical SAN infrastructure model

Typical SAN infrastructure model

SAN infrastructure implementation costs a can be high from a hardware and management perspective. Nevertheless, this storage type is faster in most cases, but the exact performance depends on the storage network and how the LUNs are defined within the SAN. The right SAN infrastructure consists of a dedicated network typically relying on a fibre-optics, enterprise-grade storage systems, and special connecting hardware. The wrong SAN setup leads to network overload and instability.

NAS also contains a bunch of disks, typically assembled into the RAID array. But this device has a core operating system aboard, which deals with all low-level processing and file system maintenance. This system shares disk contents over the attached network on object level using shared folders. Clients work with high-level protocols, such as NFS, SMB, AFP.

NAS systems are often easier to integrate into existing infrastructure because NAS can be configured like a network share. Besides, these systems are typically slower than SAN and have higher latency, because of additional query mediums and the architecture based on the regular LAN.

Since NAS doesn’t need any particular infrastructure, it is usually cheaper. Moreover, auto-backup, anti-virus protection, file versioning and a lot of other features may be integrated into the OS aboard. These features often come out of the box.

 

Typical NAS infrastructure model

Typical NAS infrastructure model

NAS is a ready-to-go software and hardware bundle, designed to use with multiple users. Both storage systems can be scaled up and down and work with numerous applications and machines types. 

Typical Use Cases

File Server

The basic use of onsite storage is a file server. Here NAS is the first idea you remember, because of built-in multi-user sharing features. Work files can be kept safe thanks to the backup and versioning support. 

SAN and NAS access difference

SAN and NAS access difference

Applications and Databases

There are some apps which need block-level access to the data so that they won’t work on NAS. The main example is Virtual Machine infrastructure, which usually requires a block level storage as the datastore, as they maintain their own file system and features. SAN suits here best, though some NAS can have a special hypervisor interfaces to satisfy its demands.

Database tables and logs are files, so they can be stored at NAS. But network errors and storage OS performance may affect database integrity - you may find detailed recommendations on Microsoft SQL Server Support Blog page. Therefore, SAN fits better for DB engines and the applications with strong latency demands.

Backup and Recovery

We have already mentioned that NAS is a good repository for files and documents. Its OS can also be configured as the backup server or storage endpoint. As far as SAN provide block-level access, the data can be accessed quickly without downloading to the recovery destination. That accelerates disaster recovery and helps to minimize a downtime. We have also described the common challenges and solutions in our Image-based vs. File-level backup article.

The Devil is in the Optimization

As you may see, SAN usually shows better performance, but it depends on the network configuration. Investments and administering skill here straightly determine the system capacity.

SAN switch with optical Fibre Channel connectors

SAN switch with optical Fibre Channel connectors

NAS can also be made faster by using the high-end devices, routing planning, using a dedicated network and overall optimizations. Both storage solutions are often used within one organization. You may have a file server for storing user files and a block storage for the disaster recovery at the same time.

How to Backup the NAS with CloudBerry Backup

CloudBerry Backup can access and backup files from any NAS. The user should simply configure NAS to be shared with Windows devices. After that, the storage device will be available for backup as a network share. To restore files to the NAS, simply choose the network share as a location for a restore.

Map NAS as network share to backup with CloudBerry

Conclusion

Now you are aware of fundamental differences between SAN and NAS devices and can find your bearing on the storage technologies ground. We have created a chart with the key features of both storage types so you could choose the right one.

SAN

NAS

Block-level access

File-level access

High performance due to the infrastructure nature, commonly faster

High performance can be achieved using the network and software optimization, commonly slower

May be configured in a very custom way

Easy to configure a basic data storage use case

Needs changes in the existing network

May be published in a network as it is

Needs separate servers for application or user access

Independent device with server functions

Suitable for any apps

Suitable for latency-tolerant apps

Grants read and write access for multiple users using external manager

Grant read and write access for multiple users out of the box

Costs more due to the infrastructure expenses

Cheaper due to the simplicity of deployment

Durable

Scalable

Effective for big data or performance-crucial business

Can be handy for a business of any size

 

Featured Product