HOW TO CHOOSE A SERVER FOR VIDEO SURVEILLANCE
SPECIFYING A SERVER
The central inquiry is what is the purpose of using servers in the first place? In the past, the foundation of many business IP surveillance projects was to use a large number of cameras with an advanced video management system (VMS) on a high-performance computer, which would replace multiple digital video recorders (DVR) or network video recorders (NVR) devices.
A server that is specifically designed and equipped with the proper software can effectively record hundreds of high-definition cameras with a sufficient frame rate for several months. This implies that choosing the right server in a well-constructed business system can provide substantial benefits in terms of system design and efficiency, reducing power consumption and maximizing rack space utilization.
Optimized servers specifically designed for HD surveillance are now available. These servers can handle over 4000Mbps of camera data traffic and offer over half a petabyte of storage within 4U of rack space. This level of performance is comparable to using 15 commercial off-the-shelf servers or the highest-spec NVRs available in the market. The cost savings in energy consumption alone can easily pay for the optimized server in a matter of just a few years.
The conversion from analog to IP surveillance has been slow due to the different skill sets required between security integrators and IT infrastructure specialists. The divide between software and hardware in the IT industry has led to most VMS suppliers opting to recommend or rebrand commercial off-the-shelf servers instead of a specialized optimized product. However, IT-based rebranded servers are not always suitable, and many VMS providers prefer to recommend an optimized solution when necessary.
When choosing a server, it’s crucial to consider three key factors: throughput, storage, and resilience. Calculating server requirements can be challenging as industry calculators for throughput tend to be inaccurate and provide limited information. To determine the server requirements for an H.264 device, you need to consider both average bandwidth and peak bandwidth.
Average bandwidth refers to the total of I and P frames in one second during normal operation, expressed in Mbps (Megabits per second). It is also used to calculate storage needs. However, it’s important to note that industry calculators often fail to consider peak bandwidth, which is the highest possible bandwidth output required if all the pixels change simultaneously.
The selected server must be capable of managing the sudden increases in bandwidth demands from all cameras at once, or there is a risk of server failure during a crucial moment. Most cameras and video management systems have a bandwidth limit for easier calculation. However, as a general guide, the peak bandwidth usage could be double that of a busy scene and four times that of a quiet scene.
For instance, in a hotel setting, there are usually a limited number of people moving around in the halls, at the bar, and in the reception area, resulting in low bandwidth usage for most cameras. However, in the event of a fire emergency, with alarms going off, emergency lights turning on, smoke, flames, sprinklers, and chaos, all cameras will suddenly experience high motion, leading to a significant surge in bandwidth demands.
For a mission-critical event such as this, the server must be able to handle the high data demands. The last thing you want is for your system to fail and shut down at the most important time. So, what are your options after completing the bandwidth calculations? Most servers that can perform multiple tasks, such as hosting websites, databases, data backup, email, and social media, are also designed to run a video management system.
However, it’s important to note that video is one of the most data-intensive tasks, even more so for multiple streams of HD video. Due to their general purpose, IT-focused servers may have limited throughput, which limits the number of HD video streams that can be recorded on a single server.
There is a common misunderstanding that upgrading the processor is the only way to increase server throughput. However, with specialized knowledge and components, there are more effective ways to optimize throughput. A few server manufacturers have emerged who understand the workings of IP video and have analyzed the different VMS options to create surveillance-optimized servers that can handle increasing camera resolutions without sacrificing system performance.
The VMS is used and other parallel video applications, such as a mobile gateway, license plate recognition, or video analytics, will all affect the overall performance of a server, often requiring multiple servers or virtualization to meet the end goal.
Industry calculators can provide an average throughput and storage specification and suggest the required processor, but they never provide enough information for a server manufacturer to build an optimized solution for the specific application. Working with an HD surveillance server specialist can save time, money, rack space, and power, and provide peace of mind that the hardware solution has been designed with a deep understanding of both IT and IP video.
When purchasing a server to replace multiple products, consider the impact if it fails. Servers designed for mission-critical applications typically have multiple levels of redundancy and use high-quality components with lower failure rates than regular components. Hard drives are the most common component to fail in a server, and enterprise drives are recommended for valuable data. Consider using RAID5 or higher for increased resilience, multiple RAID arrays for better performance, and dual redundant power supplies. The operating system and application drives should be separate and mirrored to preserve the video archive. The quality of the RAID card and battery backup is also important to ensure correct shutdown in case of a power failure. Choose a server with hot-swappable redundant components for easy maintenance. Investing in a well-designed server optimized for surveillance can save time and money while cutting corners could have long-term consequences.
When it comes to getting a server for surveillance, it’s best to let the server supplier handle the details and ensure the solution works as expected. Ask the supplier if they have a dedicated video team with knowledge of the selected VMS and its intricacies, as well as good relationships with the VMS provider. Ensure they have the ability to independently perform the necessary calculations and design a solution based on the system specifications. If the supplier is unable to guarantee the solution, it raises questions about their ability to provide support if issues arise.
It’s important to keep in mind that the IT industry is rapidly changing compared to the security industry, and older systems may have had to make compromises in the past. Upgrading to a specialist server optimized for HD surveillance can greatly improve performance and reduce power consumption, cooling needs, and rack space.
When it comes to recording HD video, it’s not just about the storage but also the ability to decode and display multiple HD live and playback streams, especially with the rise in popularity of panoramic cameras.