RAID for Video Surveillance
In video surveillance industry, the resolution advantage of the megapixel IP cameras yields greater file sizes. Hence, larger data storage capacity is a requirement for video recorders. Factors such as camera resolution, frame rate and number of recording hours per day have influenced on the storage capacity requirements. Moreover, ensuring data security against disk failure is a basic need for any surveillance system. On the other hand, due to large recorded file sizes, the system should be able to write data to disk with sufficient speed. Applying RAID technology to video surveillance, improves system performance. There are numerous ways to use RAID technology in Network Video Recorders (NVRs) such as RAID 0,1,5,6,10,50,60. Each of these RAID levels has their own unique characteristics which make them suitable for different applications. In video surveillance, RAID 5 and RAID 6 are the most popular choices which increase both the reliability and performance of the system, especially in a megapixel video system. For instance, RAID 5 offers redundancy due to usage of parity information in its configuration. If one drive fails, the parity code is used along with the remaining data on the other drives to retrieve the missing video. This achievement is very important in high security applications in which the data integrity is very crucial to the user. Furthermore, RAID 5 increases speed by stripping data across multiple disks.
This white paper describes the effectiveness of RAID in video surveillance. At first, the concept of RAID is explained briefly. Then, implementation and different standard levels of RAID are presented. The features of every RAID type are also explained to help the users choose the correct RAID type based on their specific video security requirements as well as the required archive capacity of their video surveillance system.
What is RAID?
RAID stands for Redundant Array of Independent (Inexpensive) Disks. The term was introduced in 1987 by a group of researchers at U.C. Berkeley. It is a technique to use multiple disks to build a faster and more reliable disk with same or less total capacity.
Using this technique leads to excellent system in terms of the high storage capacity. The total array capacity depends on the type of RAID array and the number and size of the disk drives. In other words, by setting several hard disks next to each other and implementing RAID technique, these disks are converted to a single unit and the system will recognize them as one disk. Actually, RAID provides a way of storing the data in different places on multiple hard disks. Apparently, adding hard disk to the system increases the storage space and also the speed of the system. Moreover, this technology improves the performance and data security.
RAID technology based on the magnetic disk technology was originally introduced to be an attractive alternative to SLED (Single Large Expensive Drive). Therefore, improvements in performance, reliability, power consumption, and scalability are all expected.
Implementation of RAID
RAID technology can be implemented in both hardware and software. In this section these two methods are described.
Hardware RAID is most traditionally implemented in businesses and organizations where disk fault tolerance and optimized performance are so noticeable. In this method, RAID runs via CPU and RAM inserted on the independent device of the host computer, which means the RAID has its own processor and memory to run. In fact, in this implementation, the RAID system is an independent small computer system. Since, using hardware RAID, processing operation is conducted through RAID controller, server is not forced by additional tasks. This is one of the advantages of the hardware RAID, although the implementation of this technique is complex and expensive.
There are some advantages and disadvantages with hardware-based RAID. It's more expensive, because configuring it requires an additional hardware component, a RAID controller which is a piece of hardware that controls the RAID array. RAID controllers can be internal, meaning they connect inside of a server to the motherboard or external (usually reserved for enterprise, high-level RAID solutions). Hardware- based RAID is also considered a better performing, more efficient way to implement RAID than software RAID. Hardware-based RAID is used most in corporate servers and business-class NAS drives.
The simplest way to identify whether a solution is software or hardware RAID is to read the technical specification or data sheet of the RAID solution. If the solution includes a microprocessor, the solution is a hardware RAID.
Some of advantages of this method are presented as follows:
- Protection in boot
- The performance of the processor will not be influenced by hardware RAID while running other applications
- Feasibility of hard replacement after failure without turning off the server
- Ability to manage disks simultaneously on different operating system
- Ability to install backup batteries
- Increasing the data security in hard disks
Software RAID is not as reliable as hardware RAID, but it's definitely more economical and can still deliver basic fault tolerance. In this way, similar to other software, the RAID runs on the CPU of the computer. Software RAID can be implemented on the host without additional hardware. This method uses the hard disk which OS has been installed on it, preventing the need for additional hardware. You can't configure RAID arrays as complex with software as you can with hardware, but if you just want to implement mirroring (which is copying data from one drive to another, to keep that data accessible in case a drive fails) then software RAID is a cheaper, less complicated to set up option. Instead of using a bunch of disks and a controller to make an array, some software RAID solutions can use logical partitions on a single disk. That's what makes it both cheaper and less reliable, if that single disk fails completely, your data is gone.
Low cost is the primary advantage of this type, although some of drawbacks of this method are presented as follows:
- The lack of disk protection at boot
- Additional tasks on server (because of running on operating system)
- Limitation in running on all OS
- Vulnerability to viruses
- Feasibility of data loss because of hardware or software problems of OS
- The lack of cash capability
Standard RAID levels
The standard RAID levels are defined based on striping, mirroring and parity techniques. In the following, the common levels of RAID used in video recording, are explained briefly.
RAID 0 consists of striping without mirroring or parity. This level stores data on a disk without error controlling. As we can see in following picture, transfer traffic is divided among several separate channels. Read or write can occur on every drive simultaneously. Also, traffic control is performed by different controllers, thus, the speed is increased.
Due to the lack of parity information, designing this level is very simple, although this is a disadvantage of RAID 0. RAID 0 provides no redundancy or data protection from the disk failure, so if a disk fails all information is lost. Hence, RAID 0 is inefficient in applications in which the data protection is very vital to the user such as video surveillance. Actually, RAID 0 is ideal for non-critical storages of data that have to be read/written at a high speed for instance in retouching or video editing.
- Biggest possible usable storage space and faster speed of reading and writing.
- Easy implementation technology.
- No redundancy
- RAID 0 is not fault tolerant. If one drive fails, all data will be missed and the whole RAID system will stop working. It should not be used for mission-critical systems.
RAID 1 mirrors all data to two or more peer drives without parity or striping. We need at least two drives for a RAID 1 array. This level is fault tolerant since all data is mirrored fully; however it requires twice as many hard drives to provide the needed capacity. So RAID 1 is rarely used to store large amounts of surveillance video.
Data is written to both disks simultaneously. If a drive fails, the controller uses either the data drive or the mirror drive for data recovery and continues operation. In this RAID level, reading operations are fast, since data can be read from both disks at the same time, while writing operations are slower because every writing operation is done twice.
RAID 1 is ideal for applications that need high performance and availability such as transactional applications, email, and operating systems.
- If a drive fails, the data don't have to be rebuilt, they just have to be copied on the replacement drive. Therefore, the second drive keeps running with no interruption of data availability.
- RAID 1 is a very simple technology.
- Reading speed of RAID 1 is twice faster.
- Cost. This type requires twice as much disk space for mirroring.
- Since all data has been written twice, effective storage capacity is only half of the total drive capacity.
- Software RAID 1 solutions do not always allow a hot swap of a failed drive. That means the failed drive can only be replaced after powering down the computer which RAID is attached to it. For servers that are used simultaneously by many people, this is not acceptable. Such systems use hardware controllers that support hot swapping.
RAID 5 is the most common secure RAID level. It provides both disk recovery options and entirely efficient use of storage space. Data blocks are striped across the drives. The parity data are not written to a fixed drive, they are spread across all drives.
RAID 5 array can tolerate a single drive failure without losing data. It delivers the best balance of data protection and usable drive capacity, efficiently. RAID 5 has been commonly used in video surveillance applications, the best DVR and IP storage systems over the past several years.
- Read data transactions are very fast while write data transactions are somewhat slower (only a bit slower than RAID 0).
- If a drive fails, you still have access to all data, even while the failed drive is being replaced.
- This is a complex technology. Sometimes if one of the disks in an array fails and is replaced, restoring the data may take a day or longer, depending on the load on the array and the speed of the controller.
- Available storage space is one disk less than the maximum.
RAID 6 is similar to RAID 5, except the parity data are written to two drives. It requires at least four drives and can tolerate two drives failure simultaneously. If a drive in a RAID 5 systems fails and is replaced by a new drive, it takes hours to rebuild the swapped drive. If another drive fails during that time, the data will be missed entirely, while RAID 6 array will even survive that second failure. It is preferable over RAID 5 in application servers that use many large drives for data storage. RAID 6 has become popular in video surveillance applications due to its increased fault-tolerance for mission critical security applications.
- Similar to RAID 5, read data transactions are very fast.
- If two drives fail, you still have access to all data, even while the failed drives are being replaced. So RAID 6 is more secure than RAID 5.
- This is complex technology. Rebuilding an array in which one drive failed can take a long time.
There are some types of RAID systems available that they have been composed of standard RAID level for certain objectives. For instance RAID 10 (also called 1+0) combines RAID 1 and RAID 0 configurations.
In this configuration, RAID 1 and RAID 0 combined with no parity. Data is mirrored and then striped. If a drive in a stripe set is failed, all access to data must be from the other stripe set. This level provides the improved performance of striping while still providing the redundancy of mirroring. RAID 10 is secure because of duplicating all data. It is fast because the data is striped across two or more disks, meaning that data can be read and written to different disks simultaneously. However, Usable capacity of RAID 10 is 50% of available disk drives, due to mirroring, this is a disadvantage in high storage capacity requirement. The reading speed of RAID 10 is nearly equal to RAID 0 and RAID 1, but its writing speed is slower than RAID 0, because of mirroring property, although this speed is faster than RAID 1. In comparison to RAID 5, RAID 10 has faster speed, due to having no parity, but less storage space.
RAID 50 combines both RAID 5 and RAID 0 features. Data is striped across disks as in RAID 0, and it uses distributed parity as in RAID 5. Striping helps to increase capacity and performance without adding disks to each RAID 5 array.
Although overall read/write performance is highly dependent on a number of factors, RAID 50 provides better write performance than RAID 5 alone.
Usable capacity of RAID 50 is between 67% - 94%, depending on the number of data drives in the RAID set. Actually, RAID 50 is RAID 5 with better write performance and fault tolerance. It is the recommended solution for arrays that need a high fault tolerance without compromising space efficiency.
Consequently, RAID 50 provides data reliability, good overall performance and supports larger volume sizes.
RAID 60 is a type of composed RAID level that combines the stripping feature of RAID 0 with the dual parity of RAID 6. Dual parity allows the failure of two disks in each RAID 6 array. Striping helps to increase capacity and performance without adding disks to each RAID 6 array (which would decrease data availability and could impact performance in degraded mode). RAID 60 also provides very high reliability because data is still available even if multiple disk drives fail.
The Table 1 summarized the features of RAID levels.
Implementation of RAID in video surveillance
Implementing RAID at the software level relies on system resources which imposes delays and makes it unsuitable for video surveillance. Hence, a hardware RAID controller is a requirement for any surveillance video storage.
How to choose a RAID level for security systems
Security systems start with video recording while playing the recorded video is the ending point. The data storing is one of the main parts of these systems. The demand for storage capacity to record video surveillance images is growing at an unprecedented rate.
Due to high capacity of recorded video in NVR devices, using multiple hard drives is required. Furthermore, in order to make the whole system more reliable, recovery options should be available in case of physical damage of any of the hard disks. These two features are possible by using RAID technology. Redundancy, the main reason for using RAID in NVR, is actually the backup version of data when a disk fails, thus increases the security of recorded data on hard drives greatly. For instance, as mentioned earlier, RAID 1 relies on the profit of mirroring property in data protection.
To choose a RAID level, there are some important points that should be taken into consideration, depending on what level of performance is required. For instance, capacity and availability needs, cost, error detection, data keeping, system performance and also the speed of both data reading and writing can be mentioned to choose the right RAID level in different applications.
In video surveillance, capacity determines the amount of video being recorded. System performance determines how many total frames can be recorded at any time, which can be critical in situation related to data loss.
There are several types of RAID configurations available in NVR devices. By using some of these configurations, the video recordings on disks can be recovered even in case of physical failure of several hard disks. However, while using such secure configurations, the total available storage space will be smaller compared to less secure configurations. Selecting a RAID level depends on the level of security along with capacity, quality of resolution, data protection and budget. It is up to the user to decide having more available storage space or more security. In some NVR devices, whichever RAID type the user chooses, it can be configured with just one click in the setup of NVR.
Regard to the description of RAID types, here some applications of different RAID in video surveillance is presented:
RAID 0 is suitable for the projects with lower security, where the lost video records is acceptable while the usable storage space should be as big as possible in order to serve more cameras with one NVR server. Example: schools, hospitals, bus stops, etc.
RAID 1 is suitable for the projects with high security and small number of cameras due to the limited storage space. Example: Currency exchange booths, etc.
RAID 5 is suitable for nearly all types of projects that are about data recovery options and need to work with many cameras and therefore would like to use as much storage space as possible. Example: Retail, city surveillance, commercial buildings, etc.
RAID 10 is suitable for the projects with high security and with medium number of cameras. Example: Branches of banks, etc.
It is worth mentioning that since keeping data, in addition to storage capacity, is very important in video surveillance, RAID 5 or RAID 6 are often used in video recorders. Therefore, in the following section, we focus on these two RAIDs in video surveillance.
The effect of RAID 5 and RAID 6 on video security
As mentioned earlier, in video surveillance, keeping the data and recovering them when a disk failures is a basic requirement. So, video recorders should use the RAID levels which are able to retrieving the recorded data after disk failure. The most common RAID levels for large amount of storage in video surveillance are RAID5 and RAID 6.
RAID 5 is composed of three or more disks and supports both stripping and parity features. In configuration of this level, stripping data across multiple disks instead of serializing them, leads to faster speed. Also, the number of locations that you can write video information is increased compared to single hard disk drives, thus performance is improved.
Parity information is also stripped across all the disks in RAID 5. If a drive in RAID 5 fails, the parity information is used to recovery the missing data. It is especially critical for IP video surveillance system, since it means the system is still running even if a drive is lost. After replacing the faulty drive, the data will be rebuilt by using parity information and the remaining data on the other drives. It should be noted that when a hard drive is replaced, the system may still be vulnerable to data loss, since it takes time to rebuild the entire data that was lost. Regard to this drawback, if a second drive fails during the first is still rebuilding, the data stored in RAID array will be unrecoverable. So, RAID 5 provides fault tolerance to just a single drive failure, confirming the urge for new RAID technology.
For extra redundancy, RAID 6 has been recommended for extreme priority of keeping data. It strips data across multiple disks like RAID 5, but it also adds a second set of parity. In fact, RAID 6 extends RAID 5 by parity information for each file to two drives. The second set of parity information is what allows RAID 6 to protect from two simultaneous disk drive failures without loss of recorded data. Hence, the system tolerates the failure of two drives in an array without losing data. On the other hand, the reading and writing performance is decreased slightly especially during rebuild step. Moreover, it also requires more disk space and costs an additional drive per array, thus is less efficient than RAID 5. However, RAID 6 ensures data is not vulnerable during the rebuild process, make it efficient for high security video surveillance.
Eventually, for large scale and critical video surveillance, the enhanced redundancy provided in RAID 6 over RAID 5 is more efficient to prevent video loss and ensure system performance. As RAID 5 protects data from a single drive failure, while RAID6 protects from two drive failures.
In this work, RAID technology is discussed and Application of RAID types in video surveillance have been studied. Since RAID 0 has no protection against disk failure, it is not efficient for critical video surveillance in spite of great reading and writing speed. On the other hand, RAID 5 will be no worse for reading data than RAID 0, but it is much slower at writing data due to parity information. Also, RAID 5 is robust against one drive failure.
RAID 6 solved the problem of RAID 5 in terms of taking long time for rebuilding, but RAID 5 has better writing performance. Because RAID 6 possesses two parity on each drive, therefore the efficient storage space is also decreased compared to RAID 5, while having high storage space to record video is important in video surveillance. RAID 5 only uses one disk for recovery purposes. The rest of the disks can fully be used to store videos. Therefore, the more disks in RAID 5 system, is used for storage. For example, if you have 6 disks and each of them with the size of 1TB, then the total available storage space will be 5TB. In other words, 83% of the storage space in given example will be used for storing videos and 17% will be used for recovery purposes. However, if one of the disks in an array of RAID 5 fails and is replaced, restoring the data may take a long time. Therefore, if another drive fails during the failed drive replacement, the entire data will be lost simply. On the other hand, due to extra redundancy in RAID 6, if two drives fail, you still have access to all data, even while the failed drives are being replaced. So, by using RAID 5 the risks of data loss are too great and RAID 6 is more secure than RAID 5 for critical video surveillance. Therefore, RAID 5 and RAID 6 may offer the best choices in video security.
NVRs which support Hardware RAID technology:
- RedLeaf Security - RLN1616U-MF
RedLeaf Security Product,RLN1616U-MF, Recording Solutions, , 16 hard disks and 16 channels , Hot swap support , Intel Core i7 CPU , 2 HDMI ports , Maple series
- RedLeaf Security - RLN8032U-MF
RedLeaf Security Product,RLN8032U-MF, Recording Solutions, , 8 hard disks and 32 channels , Hot swap support , Intel Core i5 CPU , 1 HDMI port , Maple series