Planning a World-Class Video Surveillance System

Focus on Fundamentals

Before any system planning can begin, there are several basic questions that must be considered. The answers will establish the scope of the system needed and its level of video detail, the level of automation and staffing, and the operating environmental conditions. These questions are so important, and so basic, that some people may be tempted to assume the answers are known to all, and agreed by all – but this is almost never the case. Take time at the start of the project to address these questions clearly and gain agreement, and this investment will pay generous dividends later in the project.

  1. “What are the operational goals of the surveillance system?”

Designing any system involves many decisions and tradeoffs, and the first goal must be to understand what the system is intended to do. This is not as obvious as might be expected by many lay people, and the answers may be interrelated in unexpected ways. For example, is the system intended to deter rule-breaking, or is it intended to support prosecution after rules are broken? If the answer is the former, then the cameras may be large, visible and conspicuous, but low resolution, and the monitored areas would be well-lit, even at night. If the answer is the latter, every one of those directions would be different – cameras would be smaller, less visible, high resolution, and would operate using invisible infrared light or even using thermal imaging. Also note that these decisions affect every other element of the system too – from the required data storage to the nature of the monitoring staff. It should be clearer now why answering these questions at the start is so critical; users are often confused and unhappy, for example, that a system they installed as a deterrent may be useless to support a prosecution. Here are some more supporting points to clarify as part of this first question:

How much resolution is needed? Most users won’t know the pixels per foot they need, but the point can be discussed in terms such as, “Is it sufficient to be able to know that a tall slender male is wearing a blue hat and black shirt, or will it be important to know finer details such as facial marks, jewelry, clothing logos and other details.”

How long will records need to be retained? This element may be guided by regulation, or state laws, in addition to thinking about the timing of detected events and the level of monitoring.

What areas need to be covered? Answers here range from only key internal doorways, to large internal areas, to outdoor views of entrances, to wide outdoor views of parking areas, to complete campus areas. As was mentioned above, this direction affects data storage, operator views, and many other system elements. It may also be helpful to clarify the range of activities that will be monitored – for example, are there areas that are only crowded at particular times, such as popular evening and nighttime events that would affect parking and gathering areas. Planning needs to account for expected peaks as well as the normal average level of activity.

  1. How will the system be staffed?

It is important to understand at the start of the project whether the surveillance system will be monitored at all times, at some times, or not at all.

If the system will be monitored, then the staff may need the capability of having overall views, and to be able to pan and zoom into areas of interest. However, it may be undesirable to allow users to zoom into one area for extended periods of time if the system will miss events in other areas. Some newer digital systems allow operators to zoom digitally to review areas of interest while continuing to record the entire scene.

System engineers can also design unattended systems that will record surveillance video that can be reviewed if needed.

Automation and analytics may also play a role in the solution. Analytics are now available that can support a wide range of recognizable events, including License Plate Recognition (LPR), heat mapping, simple motion vectors, as well as some very advanced demographic recognition functions such as gender, ethnicity, approximate age, and clothing colors. Analytics can be used to automate some monitoring functions if the situation allows, helping monitoring staff cover more camera views efficiently. In unstaffed situations, analytics can even alert staff that events may need to be reviewed.

  1. What are the environmental conditions?

Conditions such as extreme heat, cold, humidity, corrosion, and high dust levels will play a part in determining what equipment can be considered for the installation.

Other environmental factors also can play an important role, such as ambient light levels, the availability and reliability of existing power and network infrastructure, and more. Night-time illumination is an important factor, for example; if an area needs additional lighting, it may require permission from neighbors, potentially limiting the options to infrared illumination or even thermal cameras. If any remote areas are involved, how will they be powered and connected to the rest of the system?

Based on the answers to these questions and discussion points, system designers will determine the range of possible equipment choices and start to flesh out a basic plan with some tentative early decisions. These are a few of the decisions they will make at this stage:

  1. How many cameras will be required, and of what type? This will include considerations for camera resolution, low-light capability, dynamic range capability, and the positioning of the cameras in cooperation with lighting and analytics needs, all in the context of the purpose of the system.
  2. What are the data storage requirements? At this point, a rough plan is made based on particulars such as the expected camera resolution, frame rate, and bandwidth and retention timeframe. This early rough plan is important to guide the selection of the correct VMS back-end with the right amount of throughput and storage. Normally, the system data storage is the 2nd most expensive part of a system (after cameras), so it is important to set expectations carefully at this stage. As the situation evolves, most storage platforms can be expanded later if needed.
  3. What is the existing infrastructure, and can it be used? – This can go hand-in-hand with the previous question, but it’s important for two reasons. One, it might save the user a lot of money if they have an existing data center they can use to run the surveillance system. But there’s a support tradeoff that needs to be seriously considered. Specifically, while most enterprise systems manufacturers give support for using their software on 3rd party hardware, problems can occur that may lead to finger pointing. This can be prevented in part by looking for surveillance systems manufacturers that already have solid partnerships with mainline hardware manufacturers.
  4. How is the system going to be used? – This was touched upon previously, but specific questions regarding system staffing will help the system integrator and the manufacturer recommend the optimal components for the system. For example, if the system is unmanned most of the time, the user may want to avoid PTZ cameras because of the high risk of the camera being out of position during a critical alarm event. Training is also a factor, because unmanned systems require very different training than systems that are manned constantly. Good manufacturers and their partners will provide appropriate levels of training for each of these applications, as well as support structures to provide the necessary technical and troubleshooting support for a wide variety of applications.

Considerations for a World-Class System

Now that the basic questions are out of the way, some more advanced considerations can be addressed.

  1. What other systems, if any, will be connected to the surveillance system? – The most common example is access control. While most current systems have an IP-based interface that can easily be integrated into some VMS systems, some older systems rely on Serial-to-IP connections using ASCII communication. Working with an integrator and a manufacturer that support 3rd party integrations from older technologies all the way to the latest ONVIF and manufacturer-API standards can make a significant difference in the success of the installation.
  1. What are the needs for redundancy in the system? – For the vast majority of systems, simple RAID5 or RAID6 redundancy in storage is sufficient. Planners should also consider budgeting for “Failover” recorders and other server hardware on your VMS back-end, as well as spare cameras. If the intended system use has demanding uptime requirements, then the budget should include spare cameras that can immediately replace failed cameras, hard drives to replace failed units, and possibly extra joysticks, mice and keyboards so operators don’t face downtime if a peripheral device fails. Remember, a failed piece of hardware that doesn’t have an immediate backup will result in at least 2 days of downtime. If the replacement hardware is not available, or not shipped overnight, then downtime will quickly grow to 4-7 days, or longer. Budgeting for and providing spares eliminates this potential.
  1. What are the security needs of the system? – Will the system be isolated from the Internet? This has tradeoffs, such as easy access to upgrades, or simpler remote troubleshooting by the integrator or manufacturer. Do client PCs need antivirus software installed and do the manufacturers being considered have solutions that work with the AV solution being proposed? Are tight security controls such as camera authentication required and do the cameras and VMS being selected support it? This is one area where working closely with the manufacturers being considered is vital to understand their answers to security questions.
  1. Are there CapEx and OpEx Considerations? – Today, most VMS systems have Software Upgrade Plans (SUP) or Service Level Agreements (SLA) to cover everything from higher tiers of support to future upgrades. The recommendation is generally to buy as much of this as is needed and as the end-user can afford in the initial project, or alternatively, build it into operating expense budgets for future years. Without such a continuity plan, out-of-date software can become a significant expense to bring the system back into compliance or to obtain the required support.
  1. Are solutions manufacturers being considered? – It is best to work with integrators and manufacturers who are extensively steeped in solutions experience. A given manufacturer might have excellent quality, or the cheapest cameras, but if cameras are their only product, then they will not be in a strong position to support complete solutions. Suppliers with a complete solution will be better able to solve any problems encountered during and after installation without finger pointing.
  1. What’s the support escalation path? – Beyond the capabilities of the end-user to support its installations, how is support escalated? Does the installer have personnel that are trained and able to support the planned installation? If support issues exceed the installer’s capabilities, does the manufacturer have resources close to the site?
  1. What’s the long-term support picture? – Successful surveillance systems are as much about continuity and relationships as they are about getting a good price. Most consumers only keep their PC hardware for 3-4 years. Most VMS sites keep their equipment for 5-7 years, some much longer. Choosing an integrator partner and a manufacturer partner who have been there, and who will be there for the long haul, guarantees the ongoing support needed by most end-users.

Caution – Pitfalls to Avoid

  • Site references – Be careful in placing too much value in requesting references for end-users that are in the same industry as the site being considered. While references are certainly important, understand that reference sites of a similar size and with similar needs are more important than sites in the same industry. Just because a manufacturer or integrator has had a few successes in the same industry does not mean that they’ll do a great job. Conversely, just because they don’t have references in your industry doesn’t mean they won’t be a great partner and provide a superior solution. Look for an integrator and manufacturer that take the time to understand the application and its assets and limitations, who communicate a thorough understanding, and have a history of working together to provide a solution that meets the end-user’s needs. Manufacturers
    who are willing to meet frequently with the integrator and the end-user from the system design phase to RFQ through post-installation, and who have sufficient resources to back up that willingness, are the ones that should be the most strongly considered.
  • Example video from other sites – Along with reference requests comes putting too much weight on “application video”or video provided by manufacturers from similar sites. Understand that most end-users don’t want to provide video from their sites for use outside their four walls. So it’s not that manufacturers don’t want that library of videos, it’s that it’s generally inaccessible to them. Additionally, there are many variables that determine what makes video “look good,” so just because it looks good in the cherry-picked examples provided by a given manufacturer, that in no way guarantees that the next project will have the same results. Asking for demonstrations at the site being considered is the best way to experience firsthand what type of video quality can be expected.
  • Unneeded tech and useless specs – Does the system actually need IPv6 or is it in the spec because an over-anxious architect wanted to cover the bases? Does every camera need to have cutting-edge low-light performance or are some of them mounted in areas that have 24/7 illumination? Does the spec call for 60 days of storage retention, but the customer only needs 30? Does the site really need high-resolution cameras at every spot or are some of them mounted in areas where lower resolution cameras will do the job? Specifying what’s needed is very important, but avoiding ‘over specifying’ is just as important.

Commitment, Balance, and a Perfect Surveillance System

Every customer wants a perfect system, and they should get it. The challenge is to understand what is needed, what the challenges are, and what problems the customer is trying to solve. Achieving this, and responding with a system that meets those needs, is what separates success from disappointment. There are commitments on both sides of this equation, including manufacturers willing to roll up their sleeves and be intimately involved in the design of the project, and integrators and end-user customers willing to take the time to thoroughly and openly discuss the project and their needs. Having these discussions at the start of the project, and considering the full range of factors, will help to ensure the final result is a successful project.