• October 11, 2024, 11:55:17 AM
  • Welcome, Guest
Please login or register.

Login with username, password and session length
Advanced search  

News:

This Forum Beta is ONLY for registered owners of D-Link products in the USA for which we have created boards at this time.

Author Topic: DNS ShareCenter - Data Backup Versus Redundancy  (Read 18370 times)

JavaLawyer

  • BETA Tester
  • Level 15 Member
  • *
  • Posts: 12190
  • D-Link Global Forum Moderator
    • FoundFootageCritic
DNS ShareCenter - Data Backup Versus Redundancy
« on: October 28, 2011, 01:54:41 PM »

RAID 1 and 5 are designed to provide data redundancy (in the event of a HDD failure) and are not designed as an alternative to maintaining a backup to a second physical storage device. Data redundancy offers a means of minimizing lost productivity by providing immediate access to data following a HDD failure. Conversely, data backups are designed to provide a means of restoring a completely failed primary storage device, but are never intended to be accessed by users or replace the primary storage device.

Why are RAID 1 and RAID 5 Not Considered Data Backup Strategies?

On its face, RAID 1 may appear to be a form of data backup since data is mirrored on two physical HDDs. Although RAID 1 mirrors data on two identical volumes, accidental deletion of a file or directory on one Volume will immediately (and permanently) delete the file/directory on the mirrored volume in real time. Newer DNS ShareCenter models contain a "recycle bin" option (which is disabled by default) that will move deleted items to a recycle bin directory. Only after the files are deleted from the recycle bin are the files permanently deleted from the NAS. If this feature is not enabled, accidentally deleted files are perminantly deleted from both mirrored volumes in real time.

RAID 5 provides additional levels of redundancy, but suffers from the same accidental deletion scenario as RAID 1. The failure of the RAID 5 array, which can happen from data corruption, failed re-syncing, and other issues can result in the catastrophic loss of all data.

RAID 1 and RAID 5 are both subject to malicious attacks, such as viruses. If a virus infects one Volume of a RAID 1 array, the mirror volume will become infected in real time. Without a backup, a RAID 1 and RAID 5 array present no recovery options from malicious attacks, other than trying to restore the damaged or compromised data.

What is the Value of Using RAID 0?

RAID 0 stripes data across two physical HDDs, offering a large, single-volume, storage capacity with exceptional performance. The performance gains come from the fact that data is distributed across two HDDs, so reads and writes are concurrently pulled/pushed from both HDDs, substantially increasing throughput.

The performance gain comes at the cost of no redundancy. If one HDD fails in a RAID 0 array, 100% of the data will be immediately lost on both HDDs. RAID 0 is a good solution so long as the RAID 0 array is backed up to a second physical storage device.

Are Standard Volumes that backup on a Schedule Considered a true Backup?

Yes. Standard Volumes are considered a backup (unlike RAID 1), because data modifications to one volume are not propagated to the second volume until a scheduled or ad hoc backup are performed.

What is the Best Backup Strategy?

Ideally, your data should be backed up to a second (and even third or fourth) physical storage device at another location. This strategy will ensure that environmental factors such as theft, fire, and water/storm damage do not compromise the source data and backup. Combining RAID 1 or RAID 5 in concert with separate physical storage offers redundancy with the source and backup data, providing additional layers of data security. Using this approach, RAID 1 and RAID 5 offer immediate access to data following a failed HDD, while ensuring the backup is never accessed by users.
« Last Edit: February 13, 2012, 11:49:40 AM by JavaLawyer »
Logged
Find answers here: D-Link ShareCenter FAQ I D-Link Network Camera FAQ
There's no such thing as too many backups FFC