What you will learn from this tip: This tip will enable you to make a more informed decision on whether a NAS vs. SAN environment is best for the applications you need to deploy.
Storage technology decisions can sometimes be complex. Determining whether NAS or SAN is the best platform for your applications requires insight into the products that are currently available for data storage and host connectivity.
Let's start by getting the terms into perspective:
- NAS stands for network attached storage
- SAN stands for storage area network
The common element between the two acronyms is "network." Prior to the advent of storage networking, traditional computing environments used internal disks within a server for storing data. Clients connected to the computers over an Ethernet network using the TCP/IP protocol. Sun Microsystems Inc. developed the Network File System (NFS) protocol to allow the sharing of data files over an Ethernet network. NFS allowed a local computer user to access files on a remote computer over a TCP/IP network.
To the user, the file on the remote computer could be used just as if it were on the local hard disk. NFS even allowed for multiple users to share access to the same file across the network. Microsoft followed by creating the Server Message Block (SMB) protocol to allow the same type of file access for Windows computers. SMB was later improved and renamed to the Common Internet File System (CIFS) protocol. When implementing a NAS product, you simply connect a computer or appliance with internal disk storage to the TCP/IP network and enable users to access to those disks using the CIFS or NFS protocol over TCP/IP.
Thus, NAS is used for access to FILE storage over TCP/IP on an Ethernet network using either the CIFS (for Windows) or NFS (for Unix) protocol.
NAS works best for these types of applications:
- File serving
- File sharing
- Users' home directories
- Content archiving
- Metadata directories
- E-mail repositories, such as enterprise .PST files
- GRID computing (using 10 Gigabit Ethernet)
- Peer-to-peer data sharing
SAN was developed to allow companies to get better disk utilization and performance. Using traditional internal disk storage, there was no way to allow another computer to share the unutilized storage capacity inside each computer, or allow another computer to access those disks using the SCSI protocol. (SCSI stands for small computer system interface, and is the protocol used by applications to talk to block storage devices, which is what a disk really is.) By networking the storage, the disks that were trapped inside the individual computers could now be placed in a central location and shared between multiple computers. This allowed the sharing of disk space among multiple computers, which lead to better performance and enabled more fault tolerant computing through the use of server clusters.
In order to facilitate better performance for block storage to disk, the network used for SAN was developed for very low latency, high reliability and high bandwidth. High bandwidth optical cables were used with the Fibre Channel (FC) protocol to develop a very efficient mechanism for applications to share block-based disk storage, using the already-in-place SCSI protocol. Using SCSI over FC allowed for fast access to shared disk without having to rewrite existing applications. These low-latency networks allow large blocks of data to move very fast from computers to storage, with speeds up to 400 megabytes per second today. Faster speeds are down the road.
Thus, SAN is used for applications to access BLOCK storage over an optical FC network using the SCSI protocol.
SAN works best for these types of applications:
- Server clustering
- Messaging applications
- Data replication
- GRID computing
- Data warehousing
- Recovery archives
- Any application that requires low latency and high bandwidth for data movement.
Do you know…
About the author: Christopher Poelker is the co-author of SAN for Dummies.
This was first published in April 2006