Geekly Articles each Day
📜 ⬆️ ⬇️

Selection of servers for video surveillance systems



Today it is difficult to find a company that does not have its own video surveillance system. First, safety is paramount, and secondly, which boss doesn’t want to control employees during working hours? As they say, trust, but verify. Obviously, the wider the network of cameras, the higher the requirements for data storage and processing. Let's look at one of the approaches to choosing the appropriate equipment.

Important criteria for industrial / corporate surveillance systems are:


The system architecture in general includes the following components:
')

In this article we will not consider the configuration of the client site, the choice of the cameras themselves, motion sensors and other conditionally secondary elements, since all these components must be selected for specific conditions and taking into account the requirements on the spot.

We will focus on choosing the “kernel” of the system (consisting of a server, file storage and control software) necessary to set up video surveillance in a medium-sized company with several dozen video cameras.

CCTV Server Hardware


When choosing servers and storages for a sufficiently large-scale IP video surveillance system, you should first evaluate:


Secondary parameters determine the possibility of further scaling without allocating additional servers:


In general, a video surveillance server is a kind of multi-tasking server, differing from it, on average, with less RAM needed, more emphasis on processor processing power, and intensive use of write operations (there is some similarity with file servers). If analog cameras are used, then the server will require installation of several video capture boards, i.e. 1U, 2U servers may not be suitable. If recording comes from IP cameras, then there should be a good bandwidth at the network interface. When choosing the necessary "iron" can be repelled by the following configuration:


According to our practical experience, the capacity of such a system is enough to serve 80 cameras with a video stream resolution of 1920x1080 (using the H.264 codec), or up to 128 video cameras with a resolution of 800x600 pixels.

Solid state drives are commonly used for operating system and control software. For known reasons, storing video on SSD is not a rational idea, and in general does not bring any benefits. Now there is a line of hard drives for video surveillance systems and streaming video. WD Purple for example.

Which server models offered by the largest suppliers - HP, Dell, IBM - can be used in the video surveillance system?

Hewlett packard


HP products are relevant to the market in the form of two generations of the well-known ProLiant brand - Gen8 and Gen9.

ProLiant DL380p Gen8



Server previous generation with good configuration. According to the manufacturer, the best-selling server in Russia.

Specifications :


ProLiant DL60 Gen9


Often this server is used as a basic file server, for organizing the work of not too large-scale applications or as part of cloud clusters.

Specifications:


Proliant DL80 Gen9


It belongs to a new generation, it differs from the DL60 in an advanced storage system, which allows in many cases to do without external storage.

Specifications:


Dell


Here we look at models from the popular PowerEdge line.

Dell PowerEdge R420


Pretty typical application server.

Specifications :


Dell PowerEdge R520


A disk subsystem with support for 2.5 ”and 3.5” SAS and SATA drives with a capacity of twice that of the previous Dell server series.

Specifications :


Ibm


IBM System x3550 M5


Compact and convenient server with decent performance.

Specifications :


IBM System x3630 M3


This model has a very large disk subsystem capacity.

Specifications:


In fact, the model range of servers applicable to the task is much wider and more diverse. For example, there are tower / blades for many of the above models.

Network and extended storage subsystem


network hardware


The network and equipment used for the video surveillance system must withstand a constant load of 150 Mbps - approximately this amount of traffic generates 50-60 working cameras. The basis of these calculations is information from the sites of equipment manufacturers, indicated for a single camera that transmits video in 1920x1080 resolution. In this case, the data stream is transmitted at a speed of about 2 Mbps.

Direct connection of cameras is carried out using a network switch with PoE support, which serves as an intermediate link between the server and active elements.

To reduce the load on the network and the server, use the mode with two data streams from the camera, which must be supported by the IP camera itself. The main stream comes with a base resolution, and the second stream is data at a lower resolution (for example, 320x240). In this case, the main stream is not sent to the server (therefore, it is not possible to analyze or record it) until it switches to it under the specified conditions: for example, as a result of the triggering of video analytics trigger, or operator actions at the client site.

Thanks to this approach, the load on the network can be significantly reduced, and hence the requirements for the server components will be significantly lower.

External storage systems


In many cases, the server itself has enough hard disks to store records, but if you need to store video for a longer period and in large volumes, and if you need to provide additional data backup, you can think about connecting additional storage.

Key storage requirements:


As a result, SAS disks assembled into hardware RAID-5 with hot-swappable support are most preferred. The number of discs is selected depending on the number of cameras and the time requirements for the preservation of video recordings. We can assume that 90 minutes of compressed video at a resolution of 1920x1080 require 20 GB of disk space to store.

Control software


Video cameras can be controlled by one of two types of tools:


Firmware guarantees maximum compatibility, but does not always satisfy in its capabilities, it can be inconvenient and poorly scalable. Software of independent developers is universal and often more functional, but it is desirable to pre-determine its compatibility with specific cameras (or choose based on the availability of support).

Below are some well-proven systems for working with large surveillance systems.

Camera Control Systems for Windows


XProtect Express



Software manufacturer Milestone. Designed to manage IP video in systems containing up to 48 network cameras under the control of a single server, with no more than 5 client stations. This software allows online viewing of video from connected cameras, motion analysis using its own detector, searching the archive, playing back recordings, as well as exporting individual fragments.

XProtect Express is just one product from the line. There are both simpler versions (XProtect Essential) and more advanced ones, in which the mentioned limitations are missing and / or additional functions are added (XProtect Professional, XProtect Enterprise, etc.).

Axxon next



Manufacturer - AxxonSoft . The software supports more than 1400 camera models, all cameras compatible with ONVIF are also supported.

A remarkable feature is a flexible licensing system - you can buy a license only for the required number of video channels, at least 1, at least 1000. The functionality does not depend on the number of licensed channels, the full set is always available, including video analytics. Among the additional features there is the ability to work through a web browser and mobile applications (iOS, Android).

Camera control systems for Linux OS


Of the projects for Linux, first of all, we can mention Zone Minder .



This complex supports many USB, IP and analog cameras from different manufacturers, provides basic video processing functionality (software motion detection, etc.). There are no artificial licensing restrictions, and, in general, the program’s capabilities are limited only by the characteristics of the video server.

The advantage is the availability of the source code, which allows you to better adapt the software for the required tasks with a special need. Like many open-source programs, the project exists thanks to voluntary donations. Otherwise, its use is free and unlimited.

Also from Linux projects, take a closer look at the line of tools of the Netavis developer company (applications for the organization of video surveillance and analytics), as well as the FelenaSoft Xeoma application.

Final word


In general, the corporate video surveillance market in Russia began to gain momentum relatively recently, but now the number of available solutions and options is very large. Especially it concerns the selection of servers : we have resulted only some of the suitable models. But remember that building a well-optimized video surveillance system in any case requires careful planning in the “field conditions”, taking into account both the specifics of the object and specific tasks.

Source: https://habr.com/ru/post/308320/

More articles:


All Articles