DAS, NAS or SAN?

Posted on June 2, 2013

1


DAS, NAS or SAN, the Differences are Greater than the Arrangement of Letters               – Part 1

The acronyms, NAS and SAN get thrown around constantly and there is often some degree of confusion between these two terms.  SAN stands for Storage Area Network and NAS stands for Network Attached Storage.  Though the main purpose of this two-part report is to offer a comparative analysis between these two storage technologies, we have to discuss DAS, Direct Attached Storage, in order to complete the conversation. 

DAS is the more traditional storage format we often think of when discussing storage technology.  The classic example of DAS is a server consisting of an internal RAID 1 which hosts the operating system and a RAID 5 that hosts the data and/or server applications.  In this scenario, each server in your datacenter has dedicated internal storage.  The primary advantage of DAS is its relative low cost.  DAS also provides an ample degree of security since the user must access the server that the DAS is attached to in order to access the data that resides on it.   The large downside is its lack, or should we say dearth, of scalability because we can only insert so many drives into a server.  Backups can be complex and expensive as each server requires dedicated backup or a software client to initiate remote backups.   With the exponential explosive growth of data today, scalability is essential and having a dependable and manageable way to back up your vast ocean of data is critical. 

So let’s focus on NAS and SAN.  There is a great advantage to these storage topologies and that is the fact that all of your data is centralized.  This means a single point to manage and a single source of backup.  Rather than logging on each server to manage the storage hardware, all of the drives reside in a single rack or unit.  Backup is easier since all of your data resides on a single device.  In addition, disaster recovery is much easier as you can use these centralized storage devices for off-site storage backup.

A NAS in some respects is a computer dedicated to the task of managing data files.  Inside the system is a specialized operating system and storage disks to hold the data. The system usually supports multiple networking protocols so that different computers attached to the network can use the NAS by mapping a drive.  In this manner, the NAS can share files with multiple computers or an entire LAN.  A NAS compared to a SAN is relatively inexpensive and they are easy to implement and require very little training to manage.  If you only need a place for your users to store their files, a NAS is probably just what you are looking for. 

There are some disadvantages however.  It is important to note that data is stored at the file level and that the users communicate directly with the NAS.  Because it operates at the file level, a NAS is not associated with high levels of performance.  It also means that any user has the potential to read any of the files residing on the NAS assuming they had the permissions to do so since they are communicating directly with the storage itself.  Like a DAS, they are not very scalable.  You can purchase additional NAS devices, but a new NAS is simply just another decentralized device.  Finally, like a DAS, if you lose two drives in your RAID 5, your data is gone. 

And then there is the SAN.  A SAN is quite expensive compared to a NAS although in the past several years, the price has dropped significantly with the introduction of models geared towards small and medium size organizations.  The question is, is a SAN worth the increased cost versus a NAS.  The answer can be a resounding yes if you are concerned with issues such as security, scalability and redundancy. 

There is a huge fundamental difference in how a SAN stores data at the block level.  You can read more about the differences between file and block storage here.   Block level storage usually means higher performance levels.  In addition, because your data isn’t saved at the file level, a person with access to the SAN itself can’t access the data in any sort of readable form. 

Here is another key difference, a SAN resides on a separate network.  Like a DAS, a user cannot access the storage directly, it can only do so through the server.  The servers are connected to the SAN in a dedicated backend subnet.  This means that the server communicates privately and securely with the storage itself, preventing anyone from tapping into this communicative process.

We’ve only just begun discussing SAN technology so we have to wait until next month to finish our discussion where we’ll also talk about a couple of popular SAN models available today.

Advertisements