Large Video Surveillance Systems Guide

By Brian Rhodes, Published Oct 29, 2015, 12:00am EDT

This 14 page guide explains the key uses, design factors, and players in the large system surveillance market.

A global group of 80 integrators responded, each offering insights in selling, implementing, and maintaining video surveillance systems in the 100+ camera per system market.

Note: this is a companion to our Favorite Large System Cameras and Favorite Large Scale VMSes reports. For more market survey results, see our School Video Surveillance GuideHospital Video Surveillance Guide, and City Video Surveillance Guides for similar vertical reports.

Questions Answered

We share these core insights on large-scale surveillance systems:

(1) Who Are Typical "Large System" Customers?

(2) What Is The Typical Resolution of Cameras On These Jobs?

(3) How Often Do These Projects Replace Cameras?

(4) Which "Large System" Features Are Most Popular?

Join IPVM Newsletter?

IPVM is the #1 authority in video surveillance news, in-depth tests, and training courses. Get emails, once a day, Monday to Friday.

(5) Are these Systems Always Awarded To Low Bidders?

(6) Do Large Systems Use Converged or Dedicated Networks, and Why?

(7) How Can Large Systems Be Improved?

The list below summarizes the key finding and patterns found:

  • Multi-Building: Whether higher education, government, or even large private enterprises, most 'large' systems are multi-building even when single-site or same campus.
  • 720P to 3MP Resolutions: In the market, camera models offering 1MP - 3MP were most commonly cited, with 1080P cameras most commonly used.
  • Network Performance Important: Of all performance factors, video network reliability and efficiency was most frequently cited. While large systems place a premium on low bandwidth, failover and redundancy were also commonly given as having elevated importance compared to smaller systems.
  • Low Price Purchasing Most Common: In general, large surveillance projects are bidded using RFPs/RFQs and awarded to the lowest price, despite many suggesting negotiated sales result in better systems.
  • Access Integrations Popular: The majority of large systems that are integrated with other security systems are done with access control. 
  • Scheduled Camera Replacement Rare: It is so uncommon, many suggest cameras are never replaced before they break, with most lasting 5 - 7 years or longer before becoming an issue.
  • Dedicated Networks Most Popular: In terms of frequency and preference, having a physical separated network for video surveillance is common. Integrators emphasized that if the cost can be afforded, the benefit is less trouble and better performance.
  • Better Planning Is Needed: If Large System users want better systems, they should improve planning them, not only during the initial purchase but also during subsequent maintenance and growth.

Our detailed survey questions are examined in depth below:

Typical Business Types

Question: What are the most common types of businesses or organizations that you install 100 cameras or more on a single site?

Summary: The biggest projects seldom are only one building, but generally multiple buildings on a single campus sharing the same network. Specific answers cited universities, hospitals, and government buildings as the largest customers of 'big' systems.

Representative samples of answers include:

  • "Transportation- Airports, metropolitan transit systems Chemical storage or refineries Schools - K-12 and colleges/ university Large hospital systems Government- City, state and federal."
  • "Government / Municipal / Education. I have worked on deployments of 100+ camera systems in two airports, one city, one police service, one college and one courthouse. The airports, college & courthouse were essentially one "site" each, the city and police systems spanned many remote sites with 1 - ~50 cameras in each."
  • "High security enterprises or organizations: Such as banks, manufacturers of pharmaceutical, medical grow-ups (large scale). Hospitals, large condominium complex with parkade. Certain Fortune 500 engineering firms. Oil Plants"
  • "Airports, casino's and hi-rise buildings. Typically places that have a security force in place. We have also done large schools - which typically do not have a security force but are more interested in deterrence and forensic review."
  • "In today's world it includes City's, Schools, Airports, Hospitals, and corporate organizations. K-12 school districts have been adding cameras at a very high rate."
  • "State government for us. We have not done any private or public companies with over 100+ cameras as a fresh roll-out."
  • "Primarily large hospitals, primarily for the Department of Veterans Affairs. Many VA sites have 200-1000 cameras per location, depending on budget, coverage requirements, and of course geography."

What Resolution is Most Common?

Question: For new cameras, what is the average resolution of cameras deployed? What drives that?

Summary:  About 50% of the answers cited cameras with resolutions between 2MP (1080P) to 5MP. However, 720P or 1080P resolutions were almost as common, with over 80% of all answers giving resolutions between 720P and 5MP.  The leading factor driving the final decision is cost, with many noting they specify resolution as a function of budget, not performance.

First, some color comments from the two most common groups:

2MP - 5MP Resolutions

  • "2 to 3 Megapixel, mainly cost. Everybody wants 5 or 10 Megapixel cameras these days, but few are prepared to pay for the higher associated device costs, and even fewer realize the storage and network implications and associated costs."
  • "Average resolution is 2 megapixel . Customers prefer to have higher resolution almost in all cases, still budget which is an important parameter makes them to go for 2 megapixel or lower"
  • "2mp for interior and 5mp for exterior. Driven by cost of cameras decrerasing."
  • "I typically recommend 2MP cameras, unless the customer specifies something else. I find that 2MP provides the best cost and image quality while still allowing the customer equipment that will perform for 3-5 years without requiring immediate upgrades after purchase."
  • "2-5 MP are common. It depends on the budgets and oddly enough on the personality of the active decider in the projects (customer side)."
  • "Assuming we're counting each head of a multi-head camera individually (since they're not always in a "panoramic" layout), the average would be 2 or 3 MP. Best bang for your buck (best price/value balance). Anything less restricts the size of the areas we can cover with good quality too much, and anything more tends to be overkill except for some specific scenarios."

720P - 1080P Resolutions

  • "720p-1080p. It comes down to cost-benefit. Those devices are pretty inexpensive now, and since most of the cameras will be mounted in 10 or 12 foot ceilings, you can get great evidence for a low cost. An additional factor is that these lower resolution devices can be set to a higher frame rate with less of an impact on storage and bandwidth, so they're good for settings where you're looking for quick, small moves."
  • "720p-1080p. Sales & the ability to use digital zoom on recorded video and still get a decent image."
  • "Most have been either 720P or 1080P; more than anything it's the storage requirements."
  • "1 MP /720p seems to be the sweet spot based on proformance to cost. 1080p is becoming more prevalent but storage concerns quickly come into play."

Jobs Vary

Some responses mentioned that resolution was driven by job specifics, not general practice. Rather than buying a specific resolution because of popularity or upselling effort, it is calculated based on scene requirements.

  • "Varies from 1MP to 10MP depending upon the FOV and the detail needed. Example: If the FOV is looking at the waterside of a marine port and the goal is to identify if an analytics alarm is being caused by an anomaly or rather a real intrusion object but detailed identification is not necessary from that camera then a lower resolution may be sufficient. But if the same coverage is required with identification then 10MP would be required."
  • "it is like Miles per the Gallon (MPG) - "How much camera do you get for your dollar?"
  • "I'm not sure there is an average resolution. Resolution is situational. Since every shot has it's one characteristics it hard to specify just what resolution should be used. I would say we are not deploying anything under 1280x720."

Most Popular Features

Question: What types of features or options are important to your 100 camera per site customers that your smaller customers do not require or care about?

Summary:  The three most common factors cited were network performance, redundacy/failover, and system scalability. Apart from smaller systems where they factors are not big attributes in the operational/buying decision they almost always play a key role in large system performance.

Network Performance

How significantly video impacts the network in terms of bandwidth, security, or device manageability was the single most common factor:

  • "Access to their own domain or area. Multicast is vital as the feeds are monitored by multie operators, local security or reception and campus security."
  • "Network separation, i.e. separate network, Uptime of the system."
  • "Bandwidth control is also usually a requirement as most customers have multiple locations across low-speed WAN links."
  • "Multicast is a big one. Supported in the largest VMS's. Ability to tune stream profiles for bandwidth savings"
  • "Several simultaneous "client" connections with different levels of permissions Site configuration (all cameras from different server under the same tree)"
  • "Typically either trunk slammer one man band types or integrators that come at video from a different angle (e.g. phone system guys). Trunk slammers used to sell cheap analogue but are now pushing HD SDI/CVI etc. Still cheap, just higher resolution. Phone system guys typically sell IP and tey to bundle it worth the phone system or network gear. A lot of these types use Axis."

Redundancy and/or Failover

  • "Local electrician, they offer most Analog. Only bigger electrician companys install IP cameras. The network know how is the problem."
  • "Redundant storage and disaster recovery solutions"
  • "Redundancy, able to integrate more stores into same view client. Multiple site management."
  • "High resilience on systems: RAID5 storage, redundant power supplies, fail over between recorders, scalability to expand, easy to administer and supervise system health."


  • "Scalability is usually the primary consideration in larger sites, separating the plethora of systems capable of running 16 cameras well from the systems capable of running 1,600 cameras well. We have had a few large customers require single vendor for the entire system (cameras, recorders, workstations, etc. must all manufactured by the same company) to guarantee compatibility, interoperability and support. Many have been bitten by two disparate companies arguing over whose system is at fault and pointing fingers at each other while the customer's system is down. In the (graphic) words of one security manager, "I like having only one throat to strangle."
  • "Scalability & Flexibility of the solution. The ability to view anywhere. Flexibility of the storage solutions."
  • "Growing the system to multiple sites and greater camera counts over time is a big factor.  Systems work differently at 16 cameras than they do at 1000 cameras."

Most Common Purchasing Method

Question: How are large surveillance (100+ cameras, single site) jobs awarded? RFP/RFQ? Lowest Bid? Negotiated? Which approach is best?  Why?

Summary: Most new large projects use a competitive or lowest bid method, generally a response to RFP/RFQ to award work, while expansions to existing systems are frequently negotiated or are consultative sales by the integrator/installer, and customer:

Competitive Bid via RFP/RFQ

  • "Mostly it is lowest bid on tenders in response to RFP's, but you have to have a good argument to make a client pay more and demonstrate the benefits"
  • "The majority are thru RFPs, where the bidder has to provide the best solution that suits the client main requirement at the lowest price. the best approach is to have the evaluation done based on both technical & financial, with equal score for both."
  • "Our large surveillance jobs are awarded mostly over RFT (public tender process). I believe this approach is the best but only if price is not the only option. I believe that customer should focus at least 60% to technology offered and then 40% or less to the price offered."
  • "Always RFP/RFQ, although sometimes these are preceded by and RFI looking for industry input and budgetary pricing. As far as lowest bid, the usual terminology is best overall value to the customer. This will often mean the contractor that falls in the middle wins, although we have lost bids despite having the best solution due to cost. RFP wins are generally only successful when we are driving the solution with the customer."
  • "They had to have a min of 3 bids. They didn't know what they needed, so each integrator had to propose their own solution."
  • "Large surveillance jobs are awarded via competitive RFP/RFQ and with QCBS criteria. As this requires expertise for installing outside, special monitoring and care becomes important. Due to multiple technology involved, only a person with know all knowledge can implement."
  • "Typically RFP, but when we can get into the project early enough we will run it through other purchasing vehicles"
  • "I can't think of one that was not RFP'd. Contracts are not necessarily awarded to the lowest bidder, but often to the best offer: Reputation, references and resume all seem to play into the mix when these large contracts are awarded, and experience can sway a bid if the prices are all close. Lowest bid typically leads to disappointing results, as the VAR making the low bid usually has to cut corners to get the price down. We've been hired to fix a number of these installs and they are often a mess to clean up."

Negotiated Sales

  • "Negotiated/lowest bid 50/50. Your pricing requires explaining in terms of feature sets of a system to the relevant decision makers, and actual users of the system. Customers seldom fall for the lowest bid when comparing systems, as most quotes are not identical in terms of feature sets or hardware. You need to explain the benefit if your system is more expensive than lowest bid, and if yours is the lowest bid you are often called upon to explain what it lacks that the more expensive bids contain. Primarily negotiation, with complete honesty with the client."
  • "Negotiated is best as we are maintaining our relationships with existing customers."
  • "Negotiations. Most of the time we won not / just because of price, but because of competence. End user realized that asking the most important questions to others leads to questionable answers and unqualified reactions. Ends sometimes even in "this is not so important"-Answers even if end user is demanding it. We learned that listening carefully to end users and preparing a good project plan with most demanded features explained results in very good chances to win."
  • "Negotiated design build is best because you get paid to do a comprehensive analysis and design. Then, when all is agreed with the end user, you put it in and get paid on a previously negotiated scope. Much cleaner."
  • "Negotiated. I think is better when you have contacts inside that company to have an agreetment on the price, but a contact is key."

Most Common Third-Party Integrations

Question: What type of third party integrations (access control, intrusion, emergency dispatch, etc.) are typical with large surveillance systems?

Summary: Of all the specific integrations mentioned, access control was listed first or primary more that 75% of the time. Other major integrated systems include intrusion alarms, fire/smoke alarms and intercom or mass notification systems.

However, not all large systems integrate other systems or have the money to do so; a trend that several responses indicated could be changing in years ahead :

Access Control, Mass Notification, and Intrusion

  • "Access control integration with video is the most common, but we are also seeing alarm panel integration in buildings that may "close" at a certain time of the day. This is not common in large campus environments, because there is almost always someone who is in the building at off hours. ACS integration is still the most common and is easily achieved when 90% of our video deployments are Genetec anyway."
  • "Most of the time an integrated system is Access Control, Video surveillance, intrusion and Intercom. Then there are usually other peripharal systems to monitor and control such as vehicle barriers, interfaces with fire alarms for emergency situations, building management systems for critical alarms, visitor management systems, Radio communications systems, NPR systems, pedestrian screening systems."
  • "Access control and intrusion control are equally important third party integrations. In few projects, we even control hardware If cameras detect some movement"
  • "Larger camera systems typically have some kind of integration, either intrusion w/duress/mass notification, access control and door status."
  • "Access Control, intrusion, emergency dispatch, PSIM, Mass Notification, building automation, occupancy sensing for security and bldg control, traffic control and on and on and on."
  • "Access Control. Intercoms are becoming more prevalent as a consideration, but have not seen a lot of traction yet."

Not Currently A Factor, Or Emerging

  • "It is not really a factor in our business."
  • "Not really much we have seen."
  • "No current third party integrations - looking at adding access control and guard tour (Genetec Security Centre)."
  • "Sometimes there is interfacing to access control systems, but most of the time the systems are not integrated to other systems."

However, several responses mentioned that while integrations are not currently used, interest and funding for future work is just now coming together:

  • "Just emerging in my world. Very limited access control and intrusion."
  • "We have a couple of access control integrations but are ramping up for more intrusion."

Camera Replacement Frequency and Causes

Question: How often are cameras replaced due to failure or obsolescence?  If a camera is replaced before it fails, why?

Summary: Cameras are not commonly replaced, and failure rates are low per annum, with many units lasting 5 - 7 years or longer before the issue is a factor. The vast majority of responses emphasized the rarity of the issue, with only a few giving a specific percentage of deployed units.  That percentage was always cited as minor, usually 2% or less.

"Only When They Fail" 

  • "Most of the time people continue to use cameras until they fail or are obsolete. No one likes to spend money if they don't have to."
  • "Cameras are normally replaced only due to failure. Obsolecent cameras are left installed and working until they fail. Rarely do customers just replace cameras unless it is done in a specific area that woudl benefit from newer technology. An example would be in a lobby area where traditional dome cameras were used but with teh newer 360 degree cameras the customer can achieve more with less."
  • "Generally for brands such as Axis & Bosch cameras lasts for many years, Most of the times cameras are replaced only when fail and customers don't want to invest in new cameras."
  • "In our market the only time a camera is replaced is when it fails. I am not sure how often IP cameras fail but I have seen analog cameras fail between 5-7 years."
  • "Failure rates are less than 1%. Obsolescence is not a factor until after 7 years. If a camera is replaced before it fails, it is generally due to space renovation and upgrades are part of the budget."
  • "5 to 7 years is what we promise clients. Often we put in a 5 year warranty. If we replace a camera after the infant mortality passes, it's generally because of an environmental factor - lightening, water ingress into mounts/conduit, rodents."
  • "In our market the only time a camera is replaced is when it fails. I am not sure how often IP cameras fail but I have seen analog cameras fail between 5-7 years."

Specific Percentages Rare, Low When Cited

  • "Around 0.5 - 1%.  We don't replace it before failure."
  • "Over the last 5 years, we have replaced maybe 1% of cameras previously installed."
  • "Our customers do not expect to replace working cameras.  Maybe 1% if the resolution or fixed lensing is an issue."
  • "We replace about 1% a year."
  • "Tough to say, maybe 2% per year....? Usually lighting. Sony's have proven very reliable in operation."
  • "We might see 1%-3% failures per year but no clear reason other than a reflection of MFR quality. Sometimes we have gotten lightning burns or vandalism but not often."

Dedicated or Shared Networks

Question: Are large surveillance systems installed using a dedicated (surveillance only) or shared (multipurpose) networks? What are the biggest problems with this approach? 

Summary: In general, Dedicated or Separate networks are strongly prefered and common to large surveillance systems. As previously cited in our Converged vs. Dedicated Networks For Surveillance note, this preference almost always is borne of simplifying troubleshooting, avoid blame, and increasing available bandwidth even at the cost of new network utility.

For responses indicating Converged or Shared networks common, the main factor for the decision was cost, not performance, and several noted sharing a network only possible if segregation strategies like VLANs are used:

Dedicated Networks

  • "Surveillance network is separate from the operations network. This is avoid network issues in case their are bandwidth or network maintenance concerns. A separate system isolated the issues."
  • "Dedicated video networks are required in some regions for compliance with the authorities. The advantages of dedicated networks is control of the network design, selection of appropriate switches, quality of service, bandwidth throughput."
  • "We use only for surveillance application dedicated networks. This solution helps to avoid discussions with other applications and it is easier to manage and to calculate the UPS back-up (battery capacity), if it is necessary but for such bigger systems it is typical."
  • "Dedicated network, in large systems the last mile is quite weak and is important to have a system that understand this and uses it (something like a frame/relay arquitecture)"
  • "In most cases (75%), the client is asking to have a separate security network (for surveillance and associated systems like ACS, Intrusion, etc..). but this network should be separate from his internal computing network."
  • "All our bigger projects running on dedicated Video Networks. As there is no other traffic, it's much easier to dig in troubles and find issues. On the other side, it makes it harder to have remote access to it. Sometimes, end user just don't allow remote access at all, so you need always to go on site for troubleshooting."
  • "Mine are all on dedicated networks, the Gov't does not want these sytems sharing their "public" network."
  • "Despite what some integrators say, we always recommend a separate security network. This allows access and video to stay off of the corporate network, improving both performance and security. The largest difficulty with this approach is often the cost involved. Nearly all of the security networks we build use Cisco switching, which is costly and customers often overlook the need for network connectivity when writing RFPs."

Converged Networks

  • "Large systems typically share an enterprise network that is supported by the customer's network administration team. While this minimizes the cost of deployment, it introduces the possibility of "finger-pointing" when a network outage causes video outages. It is common to have longer repair times for service issues where network administrators are involved in helping troubleshoot camera outages."
  • "We usually share network, but with dedicated VLAN. Using dedicated network will increase the number of the required switches, patch panels, etc. Share network will provide more flexibility."
  • "Shared and Converged Networks.There has been no problems as we sit on a seperate VLAN within the main."
  • "Surveillance only is what we quote the client with enough budget to deploy separately. But in most of already established sites client prefer to have. Multipurpose networks for cost saving. Biggest problem in Multipurpose network is the dependency of ip support team for addresses. And also bandwidth required to handle all is very high."
  • "Usually we use shared network, but we insist on VLAN-s. We depend of another company that do a network configuration."
  • "Shared networks are used more often for large systems. Restricted bandwidth and network terminations not in ideal locations."
  • "When you say large, I think of a LAN that is 250+ routers and switches in size, with 100-150 different sites. In a network this large, it is difficult to physically isolate surveillance systems from the general network because then you have to have nearly double the equipment. And that is not ideal in any situation. A properly configured network combined with proper management, is not a problem in most cases. A good network administrator or network administration team should easily be able to keep the networks "isolated" at the programming level, thus keeping out any prying eyes."

Improving Systems and Lessons Learned

Question: In general, How would you improve the 100+ camera, single site surveillance systems you have worked on?

Summary: While answers varied, two main themes emerrge - better planning and better networks.  Other factors like better cameras and more integration was cited, but not with the regularity and conciseness of these two points:

Better Planning

  • "Planning, planning and more planning, followed by strict implementation protocol. i.e. we install all storage, servers and infrastructure components first. Followed by Networking to all camera points, then test complete infrastructure prior to installing any cameras. Only when infrastructure is stable do we start provisioning devices. On systems that we get called to consult on, the most basic lacking requirement is that the network connectivity , storage and VMS was inadequate, therefore these require the most improvement."
  • "For large deployment users should be engaged right from inception stage not at the end. Design should be coordinated with other systems. Redundancy and Disaster recovery should be considered."
  • "Ensure that all the components that make up the system can be fully integrated so that all the functionality can be achieved. Sometimes it is a compromise where full functionality cannot be achieved because of disparate systems interfaces do not provide everything they claim they can."
  • "More participation from system end users in planning states. Two key areas are security requirements and end-user features. IT/IS folks tend to focus too much on the server, network ports, camera counts. When you turn the system up, sometimes end-users start coming up with all kinds of special needs, or the security folks raise issues with locations that have security issues but no cameras. Typical, everybody just needs to communicate more."
  • "Better planning/forecasting and documentation, shop drawings, and assurance that the parts needed are available before promising delivery dates."
  • "None of these clients understands what they want in Year 3, much less Year 10. Better planning is needed."

Better Networks

  • "Use fibre wherever possible. Avoid WLAN unless there is no practical "alternative" Have redundant power supplies on all switches. "
  • "Better bandwith at our disposal (Management says more than 15 Mbps upload is too expensive for monthly fee) causes end users to complain of slow FPS in live view."
  • "On sites with third party managed networks, we now have to contact the customer to create a help ticket with the network management company, (as a contractor, we cannot create work tickets for other contractors) who then responds to that ticket on their schedule, while we are dead in the water until they do. This incurred delay affects our "time-to-resolution" metric. Many times the network personnel come back with a "no problem found, it must be your device" requiring us to further substantiate our complaint with specific tests performed, MAC addresses in play, IP addresses and expected outcome, etc. and then after a few sessions back and forth they find that some update pushed to the switches or the core either missed something or broke something. "
  • "Being as though I typically am designing them, I wouldn't change much other than provide a dual network with dedicated Lan infrastructure for the CCTV equipment, so we can better manage it with the security issues we face working in the Govt sector. But i don't see that changing. I would certainly start employing more converged access and video systems."
  • "Having better underlying networks would be very helpful."

1 report cite this report:

Largest New Video Surveillance Projects 2016 on Mar 21, 2016
100 Integrators checked in with our latest survey looking at large projects...
Comments (0) : Subscribers only. Login. or Join.
Loading Related Reports