3 min read

Data Protection: Raid 101

Featured Image

I get the pleasure of working with several refrigerator-sized storage subsystems capable of holding several hundred to thousands of spinning hard drives and solid state devices. These drives combine to store and service Petabytes of data.  One of my main responsibilities is ensuring your data is available even if a few of those drives go bad and fail. I often get asked about data protection, so I thought I’d review one of the tools we utilize within our arsenal – RAID (Redundant Array of Independent Disks)

There are numerous variations to RAID protection, but most enterprise systems use one or a combination of the RAID protection types discussed below:

Prevailing RAID Options: 

Prevailing RAID

RAID 0 – Striping

(Distributes incoming data blocks over all of the drives within a given set of drives)


Striping

RAID 0 – Major Benefit:

  • Very Fast as data is distributed across multiple drives

RAID 0 – Major Drawback:

  • No data protection
    (A single drive failure will result in data loss)

     Striping 

RAID 1 – Mirrored Drives

(Copies incoming data across two drives as shown in the following diagram)


Striping

RAID 1 – Major Benefits:

  • Can survive a single failed drive from any mirrored pair (Basically, you could lose half of the drives in an entire array!)

RAID 1 – Major Drawbacks:

  • Overall capacity utilization is cut in half (A 100TB system can only store 50TB’s if all data is mirrored)
  • Write Input-Output (IO) is effectively doubled since the data is written twice

 Striping

RAID 5 – Striped Parity

(Stripes data across available drives like RAID0, but adds a parity bit so data can be reconstructed in the event of a drive failure)


Striping

RAID 5 – Major Benefits:

  • Can survive a single drive failure per RAID 5 set
  • Decent read response times (can read data from parity blocks)
  • Better overall drive space utilization versus RAID 1

RAID 5 – Major Drawback:

  • Slower writes due to parity calculations which increases write IO by a factor of 4 


Striping

RAID 6 – Striping with Dual Parity

(Similar to RAID 5 but with two parity bits)


Striping

RAID 6 – Major Benefits:

  • Very high redundancy and availability (Can survive two drive failures at the same time per RAID 6 set)
  • Decent read response times (can read data from parity blocks)

RAID 6 – Major Drawback:

  • Slower writes due to parity calculations that increase write IO by a factor of 6.


Striping

Follow up posts will discuss where and why to deploy the different RAID technologies.  We will also discuss newer strategies and technologies to ensure your data is always protected.