Air RAID
Gaining performance of a video surveillance system
- By Tri Nguyen
- Nov 01, 2013
As IP camera usage increases in video surveillance systems, there
is a need for larger data storage capacities. The resolution advantage
that is inherent with megapixel IP cameras comes with greater
data rates and file sizes. Because a 1080p IP camera contains six
times the resolution of a VGA analog camera, the storage capacity
required for the same amount of recorded time is six times as much. Ensuring that
large amounts of data are safeguarded against disk failure should be a requirement
for any surveillance system.
However, another challenge posed by larger data rates and file sizes is the video
system’s ability to write this data on a disk with sufficient speed and reliability.
A RAID 5 storage array increases both the dependability and performance of a
video surveillance system, especially a megapixel IP video system.
What is RAID 5?
Redundant Array of Independent
Disks (RAID) is a
method for configuring multiple
disks or hard drives in
a single array. While there
are many levels of RAID,
focusing on RAID 5 for video
surveillance is key.
Because RAID 5 requires
a minimum of three
disks and uses a combination
of striping and parity
to write data across all the
disks rather than sequentially, the data is spread across multiple disks and faster
speeds are achieved. Parity is an error-checking technique that is used to add redundancy
to a RAID 5 configuration. If a drive in a RAID 5 array fails, the parity
information is used with the remaining data on the other drives to recover the
missing data. Like the data, the parity information is striped across all the disks.
The effective storage space of a RAID 5 array can be calculated by the formula
1 – 1/n, where n is the number of disks in the array. For example, if you have 8
drives of 1TB each, the effective space is 8TB total x (1 – 1/8) = 8TB x 0.875 =
7TB. In other words, one drive will be used for parity. Overall, RAID 5 increases
performance and offers redundancy with efficient space, making it an effective
configuration for megapixel IP video systems.
Increased Performance
Hard disk drives have read and write speed limitations. Due to their mechanical
nature and physical limitations, by configuring drives in RAID 5, the number of
locations that video information can be written increases.
In a megapixel video application, a large amount of random disk operations
occur. Video information is being stored along with the indexing of this video. If
motion detection and audio is used, the motion detection and audio index must
be stored as well. In a non-RAID 5 configuration, these non-linear disk operations
are typically handled by the CPU, which can cause the entire system to slow
or crash.
In a RAID 5 system, a dedicated hardware RAID 5 controller manages the
large number of disk operations. Depending on the quality of the RAID 5 controller,
the overall system throughput can be doubled or tripled. For example, a
16 camera, 1080p, megapixel IP video system at 15 fps requires approximately
118 MBps throughput. However, a Windows-based NVR typically has throughput
capability of 90 MBps, due to disk write limitations. When configured as RAID
5 and using a robust RAID 5 controller, this throughput increases to 250 MBps,
resulting in the ability to connect 33 1080p cameras at 15 fps, compared to only 12
cameras per server.
The Requirements
The main requirements for using RAID
5 are applications that need data redundancy
with throughput requirements
exceeding the non-RAID 5 capabilities.
Data redundancy is on video data. An
office environment, for example, may
not be as critical as a banking or public
safety environment. Often the need
for redundancy will be dictated by the
specifications of a particular project.
Calculations may be needed to determine
the required throughput and
the throughput capability of the surveillance
system. The camera bit rate
should be determined and multiplied by
the total number of cameras to figure
the minimum throughput requirement.
Overhead requirements, such as remote
client connections and backups, need to
be considered. This throughput value
should be compared to the manufacturers’
system throughput to determine if
RAID 5 is needed.
Advantages and Disadvantages
RAID 5 gives an end user the security
of data redundancy and the ability to
recover data in the event of a disk failure.
It increases the throughput performance
of a system because of its ability
to write on multiple drives at once.
Coupled together, these two factors
make RAID 5 a good option for high
camera count and high megapixel applications.
Because of the added hardware, and
in many cases, a higher grade of hard
disk drives, RAID 5 will be more costly
than a non-RAID 5 system. The effective
storage of a RAID 5 system will be
decreased as well, due to the extra parity
information. Other disadvantages
include longer build time of a RAID 5
system, the complexity of adding hard
drives in the field, and the knowledge
level needed to recover lost data in the
event of a failed drive.
Today’s IP megapixel applications
require more data storage and performance
capabilities of a video surveillance
system. This requires the system
to be robust enough to write large
amount of data onto disk without
compromising the performance and reliability
of the system.
RAID 5 adds data redundancy as
well as performance increases to handle
megapixel IP video applications. Critical
video should be recorded with a
fault tolerant system and high IP applications
need greater performance.
RAID 5 is an effective solution to both
of those needs.
This article originally appeared in the November 2013 issue of Security Today.