For many people, choosing a video surveillance system is as simple as buying the ingredients for an exquisite dinner. by Germán Cortés *
But the truth is, just as you choose the best foods for a delicious meal, you should choose a complete and good security system for your company.
Recently the systems of Closed Circuit Television, have had a boom of massification due to the proximity or convergence with the world of IT (Information Technologies) that we have commonly called the world of computers.
Many times I have said that this integration between the world of security and that of computing is a double-edged sword, but the guild of electronic security needed the massification, dissemination and dynamism that the world of computers has. Nowadays these computers are essential in any office or home, they are already another appliance; however, this closeness to end users becomes a problem for the world of security, due to several factors.
On the one hand, by demystifying video surveillance systems, many criminals also know the systems and sooner or later they will end up applying strategies to breach the equipment. On the other hand, the cameras, video recorders and other necessary accessories, fell into the same game of a PC, selling it as a black box, which is sold without knowing exactly what you want and if the technical specifications are adequate or not. There are many computer marketing companies that now sell cameras as an accessory and even give away the recording software.
Nothing more serious for the world of security, than to ignore the calculations, plans, risk analysis, study of specifications, planning and estimation of the way our "security system" would operate; as we were accustomed to experts and connoisseurs of the subject of electronic security.
Nowadays a computer vendor supplies cameras of any specification (almost always the baddest and the most economical), recording on the hard drive of any computer (almost always with a homemade configuration), using software (often illegal). The sad thing about this true story is that both the seller and the buyer take pride in the "security system" that they have just marketed.
So where was the safety engineering? Where was the technical experience of many years and the "indigenous malice" that characterizes us, especially Latinos, when we think about security? We turn an entire industry into a low-level barter, where what matters is volume and not quality.
With this article I want to raise awareness so that we understand why it is important to put a little engineering to the subject of video recording, and avoid mistakes that can be very expensive.
Types of video recorders
We see some possibilities we have today:
1) Cameras with analog output, with analog video recorders and analog control equipment. This is the traditional system, existing in many places and that really is already obsolete, and therefore these equipment are almost no longer installed.
2) Cameras with analog output with digital video recorders with analog inputs (Digital Video Recorder -DVR) and digital control equipment. This is the most used alternative today, where we replace the analog video recorder (VHS cassette), with a digital one, which stores the information on a hard drive. In this case the market can offer several alternatives for the video recorder:
to. The Dedicated, Autonomous:
i. Compact unit with analog video inputs (BNC connector), which physically has no resemblance to a PC, but which in its electronic architecture is based on a traditional computing scheme (source, motherboard, hard drive, communications ports, video capture card, input and output devices). It uses a proprietary operating system or any commercial embedded version and software dedicated to recording video.
ii. Compact unit similar to option i, but much more robust, with accessories that allow to offer a continuous and safe operation. (Watchdog, 24-hour source, ports for disk expansion, better processing capacity, among others). Additionally the computing capacity and internal devices are designed to store and play all existing video channels. Traditional devices and accessories such as Joystick keyboards and networking are accepted.
b. Those based on a PC:
i. Video capture card installed on a home-type PC. It uses traditional operating system. The recording software almost always comes with the video capture card and is installed as one more PC application.
ii. Video capture card with independent processors and dedicated video memory, installed on a home-type PC, similar to that of option i.
iii. High-specification video capture card, installed on a server-type equipment, with software dedicated to video recording. Communication to the outside of good speed.
iv.High specification video capture card, installed on a mission-critical server-type equipment, with high specification devices, according to the capacity of the input channels, with embedded operating system, software dedicated to video recording and accessories that guarantee a continuous and safe operation. Ability to receive joystick keyboards and send information at high speed through high-speed communication channels.
3) Analog cameras with IP encoders, video recorders (Network Video Recorder - NVR) and digital control equipment. This is a recently used option that converts analog video signals into fully digital signals and employs the recognized information protocol used on the Internet (IP). In this case, the video signal, converted into traditional high-speed data, is recorded on digital devices that receive the information through an ethernet-type network card (NIC). That is, the video recorder has no video inputs, nor video capture cards, all data traffic (video) enters and leaves through the LAN/WAN type communications network. It is used a lot, when we must concentrate the video signals from various remote points and above all, there is already a data network that we can use, with enough bandwidth available for video data. In some cases it is suggested to build a data network totally separate from the traditional one, with the aim of not sacrificing performance.
4) Digital cameras (IP), with video recorders (NVR) and digital control equipment. It is the same previous possibility, but starting from cameras that deliver a digital output (IP) directly, without the need for encoders or interfaces. It is the most recent option and the one that is fashionable. However, I must say that as in the world of fashion, not everyone looks what is fashionable, even many times we look better with old garments but that perfectly fit our needs. The same goes for a CCTV system, the best system is the one that suits your needs, your budget and your operation. And sometimes, it's not necessarily the one that's trendy. For the last two possibilities, the market also offers several alternatives in the video recorder (NVR):
a.Compact unit, made in factory specially equipped to record video from the network; internal architecture based on a server-type PC with accessories and high-specification devices. Embedded operating system and software dedicated exclusively to the function of recording and playing video.
b. NVS (Network Video Software) solution, which is just the software package, installed on a PC that could be:
i. A PC of home configuration, with lower performance and performance.
ii. A mission-critical server-type equipment, with high specification devices, according to the capacity of the input channels, with embedded operating system, software dedicated to video recording and accessories that guarantee a continuous and safe operation. Ability to receive Joystick keyboards and send information at high speed through various high-speed communication channels (NICs) (10 Gbit).
Computing platform
In any of the possibilities mentioned, from a compact DVR that comes from the factory for this purpose, to an NVR based on the architecture of a mission-critical server, there are certain conditions that the computer equipment must meet.
In all cases the electronic equipment must perform the same function, that is, record digitized video on a hard drive; therefore data processing capacity is a very important factor.
In this case, technical factors such as the processor, the speed of access to RAM, the speed of writing and reading video memory, the speed of writing and reading the hard disk, the speed of data transfer to network communications cards, among others; are decisive in determining the performance of the VCR.
The speeds of video recording, expressed in FPS (Frames per Second), as well as the video resolution, expressed in the size in pixels of each pictogram (CIF, VGA, D1, xx MegaPixels), depend not only on the compression algorithm used, but also on the large data processing capacity of the PC.
Efficient recording
For example, we have a system with 32 video channels. We want that in a critical situation, that is, when an accident or emergency scenario occurs, all cameras can be stored simultaneously in real time (30 FPS) in high resolution (D1:720x480).
Note: Obviously this will be necessary only in an emergency situation. In the rest of the time it is possible that each channel is only recorded based on the motion detection of each camera, at lower speeds and at lower resolutions, depending on the scene generated by each camera. This will cause the hard drive to store many days, even weeks or months.
However, I insist, it is important that the machine has the ability to record in the best resolution, at the best speed in a moment of emergency. This will obviously cause the hard drive to run out in a very short time (hours or even minutes), however it is well worth it to have a perfect record of what happened at a critical moment of security; after all, that's what you buy a video surveillance system for.
After the accident, I can take a copy of the hard drive in a massive medium such as a backup tape or transfer the information over the network asynchronously and free up space on the hard drive again to continue with the traditional operation.
The opportunity to have had recorded real video of excellent quality in a high-risk situation is invaluable and unrepeatable, and guarantees that future research can be done based on a video material, which if it works. This is simply efficiency.
It happens today, that many systems are content to record the cameras at low resolution (CIF), at less than 4 FPS each, simply to make the hard drive last longer. The result is a useless investigative material, where no faces or criminal acts can be identified and then we return again to that feeling of frustration caused by the analogous video. That is, investing a lot of money just to have vague evidence that he stole "a human..." from us, but without clearly knowing who he was, how he did it, and without being able to review in detail the criminal actions. So I ask, is it justified to have invested all that for that poor result?
Basic calculations
Let's continue with our example. To record 32 channels of video in real time (30 FPS), I need a system that is capable of recording 960 FPS in D1 resolution to the hard drive.
If we assume that we have an effective compression algorithm based on the MEPG2 or similar standard (i.e. high quality) for local recording and an MPEG4 or similar standard (i.e. medium or low quality) for sending remote video over the network, we can assume that on average, for local recording, in excellent quality every second is taken 900 KB, and for transmission 250KB is taken.
Then the computing platform must store (32x900Kx8) 230,400 Kbps or 225 Mbps. The exact questions are: a) can the external bus of the motherboard handle this speed?, b) Can the system RAM temporarily handle this data?, c) Can the processor handle this information simultaneously with the other activities it must do (visualization, motion detection, graphical interface, control, registration, attention to remote users, among others)? and above all, d) is the hard drive capable of writing at this speed? surely the answer will be NO to most. So what can be done?
The answer to the first question is simply to make sure that the motherboard uses a high-speed communications bus, this is only achieved in models that handle multiprocessing, that is, where we not only use a microprocessor, but two, four or even eight processors operating simultaneously so that each one takes over a part of the process.
The second question is solved by expanding the RAM, the recommendation is to have between 2GB and 4GB for each existing processor. The third question is solved as in the first option, multiprocessing.
For the fourth question it is asked similar but at the level of hard drives, that is, using disk configurations in RAID 0 arrays, which allow segmenting the data and storing a part of them on each hard disk, providing a high storage speed, however these arrays tend to be damaged with relative ease, because all hard drives (mechanical parts) live in continuous operation. To avoid these failures, then we suggest RAID 10 or RAID 50 fixes, which in addition to increasing the speed double the information and in this way we reduce the risk of data loss.
So the solution became technically simple, multiprocessing, increasing RAM and using RAID 10. However, the cost of this configuration is very high, compared to the initial option (a processor, a hard drive). Additionally, these equipment require additional control cards, special cooling, higher power consumption, double or triple hot swap power supply, and many traditional accessories in computer systems of very high specifications.
So what can we think of a system that promises us (remember that paper holds everything) simultaneous recording of 64 cameras with a single processor and a traditional 120 GB disk? I leave the answer exclusively to you and your conscience.
Also, assuming that all the computer problems are solved in a computer, I do not recommend having a single unit to record all the cameras, for security. In case of damage to the unit, the images of all the cameras will be lost. It is preferable to have several units (three or four) that allow you to efficiently handle the computing needs and offer adequate security.
Hard Drives
Hard drives are the ones who carry the main role in video storage. A hard drive must have high performance (speed in reading and writing), a large capacity (measured in giga or tera bytes), excellent reliability (ability to operate as expected, without being damaged), must be easily managed by the operating system or the software that uses it, must be small in size (improvement in seek time), moderate power consumption and finally the data transfer interface must be fast and reliable (IEEE1394, USB, IDE/ATA, SATA, SCSI, e-SCSI, among others).
Let's look at some concepts:
Performance: expressed in access time, where the average access time is known as the sum of the minimum duration of various electromechanical processes: Command Overhead Time + Seek Time + Settle Time + Latency.
Seek Time: time that the head lasts to move between traces.
Settle time: time that lasts the head to stabilize, once it has been moved, to be able to start recording or reading data.
Command overhead time: time for the work that the DD does to begin.
Latency: Time it takes for the cylinder to pass through the head. The maximum is half a turn of the disc.
This time is a very important factor, because let's remember that the video recorder is permanently storing information on the hard drive. In addition, on many occasions while storing the data you may also be reading data on another part of the disc to play video that is requested by the end user or by a remote user. The lower the access time, the better the performance of the hard drive.
The cache of the hard drive itself is an important factor, 512KB is not the same as 4MB. Performance is very large when the cache is large.
Additionally, many hard drives encode and decode digital signals (1 and 0) by modulation to ensure that the information is magnetically fixed on the magnetic substrate of the hard disk. This takes a while and spends performance on transferring information.
Another important aspect is the geometry of the hard disk, number of platters, number of heads, number of cylinders, number of sectors. The sum of all this gives me a data transfer rat, which is the one that really interests me.
There are two types of data transfer rats, the external and the internal. The External, refers to the speed at which the internal circuits of the PC, but external to the hard drive, deliver data. It is almost always a large capacity and depends a lot on the memory buffer that the DD has. However, the internal one refers to the speed in the mechanical and electronic process of accessing the heads, cylinders, actuators and others, to record a series of modulated and encoded data on a magnetic substrate. This is really the lowest speed and it is the one that marks the limit of what the hard drive can actually record in a greater capacity. This rat is known as STR (Sustain transfer rate).
Another speed specification is MTR (media transfer rate), however it does not work for these cases where the files are long, because MTR is valid for data in the same sector or track and in the case of video files, they most likely occupy different sectors, therefore the most real comparison index is the STR.
Many of these deficiencies are solved with arrays (many disks that in a synchronized way behave as a single one of higher speed and greater capacity), in RAID configurations 0, 1, 2, 3, 4, 5, 6, 10 and all its variants ... however this topic will be developed in a next installment.
With these details I hope you have clarified some issues and realize how critical is the proper choice of the machine where our video recording system resides. The conclusions are taken by yourselves.
* Germán Alexis Cortés is an electronic engineer expert in building automation and General Manager of Insetrón Ltda. He is dedicated to project consulting and providing specialized training of a Latin American nature. If you want you can write to the email [email protected]
Leave your comment