Drobo BeyondRAID Technology

The Drobo BeyondRAID technology was developed as a reliable RAID storage aimed at the user who lacks technical skills: such users can turn the device on, put some disks inside of it and do only minor service, such as replacement of failed disks. There is no need to know what kind of disks (models, capacity, speed etc.) are installed, what kind of data protection is used, which RAID level is applied, etc.

How does it work?

The Drobo firmware creates special markup on disk, transforming it to the "component of pool of disk space". This markup includes the following:

  • The heading area to identify the disk as a Drobo BeyondRAID component;
  • The main storage allocation table that describes which disks contain data and what algorithm is used;
  • Storage area translation table: to translate the file system address to storage allocation.

With the help of its algorithms Drobo NAS can handle disks of different capacities, speed etc. and automatically create "micro-RAID" regions with small chunks of data for allocating data with redundancy and in optimal way for specific storage configurations.

The third component (storage translation) virtualizes the file system volume and enables data allocation on demand. Usually Drobo creates a file system of "virtual size" (such as a fixed size 16TB volume, even if much less disk space is physically available). This allows to dynamically allocate storage space from the pool and replace (on demand) the disks with disks of higher capacity.


What could be the problem there?

The problem arises from high system dynamics. When some changes occur, Drobo starts applying these changes by moving and rebuilding the storage allocation (and file system virtualization) tables. These tables occupy much disk space so moving/rebuilding these tables may take a long time. Hardware issues or power loss during this process are the main reasons for BeyondRAID metadata damage. The system has automatic recovery capabilities, however, not all data damages can be handled properly, so rare cases of NAS failure may occur.


What can be done?

Using special software, it is possible to search for older fragments or copies of metadata tables and use them to translate data chunks from the storage again to "virtually linearize" the file system. The quality of this reconstruction will depend on the quality and age of found metadata and it does not always guarantee 100% data recovery.



Copyright © 2016-2017 LLC SysDev Laboratories. All rights reserved.