As with all aspects of the security industry, video storage is progressing at breakneck speed. While there is always the temptation to stick with what works and most resembles what an integrator is accustomed to, the industry is increasingly becoming one in which those who don’t keep up will surely be left behind. 

Consequently, integrators must have a clear understanding of all aspects of video storage so they can offer their customers not just something that will work, but something reliable, cost effective and ideally tailored to each customer’s needs.

Integrators can better meet the needs of their customers by asking the right questions.


What storage architecture is most appropriate for our clients’ needs?

When considering distributed versus centralized storage options, the right choice will depend on the client and the network environment. “Distributed is often best in a multi-site user,” says Bob Dolan, director of technology, Anixter, Glenview, Ill. “However, if the user owns the data link between the multi-sites, then centralized becomes more cost effective.” Dolan says integrators should take into account the number of devices to be recorded. “With more devices at a site, local storage becomes cost effective due to the size of network pipe required to transmit the video to the recorder,” he says.

The robustness of the network is a key factor in making the right decision for storage architecture. “Centralized storage puts a larger load on the network,” says Tom Larson, director of sales and engineering, BCDVideo, Northbrook, Ill. He gives the example of a college campus, explaining, “If you were to centralize the campus to a single data center, that network has to be more robust because all the cameras have to stream out of that building across the campus to the data center.”

Larson breaks down the choices saying, “You have what I call distributed versus centralized. Now if it’s a small job — 20 cameras — it’s going to be a single server out there. But when it gets a little bigger and you have multiple servers, it’s distributed, meaning I distribute my cameras across multiple servers with storage in it, or we centralize it.”

Distributed, he says, can be broken into two categories. “We are doing work for a chain of coffee houses. That’s just a little server; there’s no RAID protection; there’s no redundant power. It’s just a box with 6 TB of storage. That would be considered distributed. At the same time, you could have a job that’s more enterprise, with a higher number of cameras, and now you want a server with the storage protected; that would be RAID protection of the storage and redundant power, and that would be in the distributed category still.”

Susan Brady, marketing manager, WavestoreUSA, West Palm Beach, Fla., adds, “Many users tapping into the server remotely also puts a strain on the system and a more substantial storage solution is required beyond a hosted service.”

As the importance of video access changes, so too does the appropriate storage solution. If a customer needs immediate access to video at all times, then they need to get a storage solution that is truly enterprise class.

“If the customer wanted a higher level of protection or redundancy, that’s when you’d do a failover server,” Larson says, “and you would have primary servers and then a hot standby for a redundant server. If your primary server goes down, you don’t have the ability to access that video.” Larson explains that businesses such as casinos and theme parks must shut down tables or rides immediately when they lose video access. “There is a huge cost for not having video.”

In such cases, Larson advises centralized storage arrays because they don’t have a single point of failure. “You can have a primary server and a failover server. If the primary goes down, the secondary picks up; there’s no storage in either one because the storage is in the centralized array.”


How scalable is this setup?

As integrators tailor video storage solutions to customers’ needs, the proper architecture has to take into account the customers’ needs, including the ability to grow — or shrink — with the customer’s business and with technology.

“Ensure your design provides a scalable and future-proof solution,” Brady says. “Storage is all about knowing and determining your clients’ current requirements and forecasting future system expansion. Integrators should make sure the solution offers flexible options for storing video, audio and data. Regardless of the current size of the project, you can be assured that clients’ requirements will change or evolve. At some point, there will be a need to add more cameras and the video storage capacity requirement could grow dynamically. Plus, to gain maximum advantage from their investment in a video surveillance system, clients will want to integrate with other new systems and technologies, and this in turn may have storage implications.”

Companies are not static, says Dean Drako, president and CEO, Eagle Eye Networks, Austin, Texas. “They continually evolve, and their video storage needs can change as they do so. For example, they may want to add cameras or extend the retention period; they may even need to reduce their video storage.”

Drako points to the flexibility of the cloud as a possible solution for scalability. “In addition to cloud services offering a predictable monthly ‘op-ex’ subscription with a lower up-front cost, they also offer the ability to immediately change the storage retention period up or down,” he says. “This means not only avoiding the cost of purchasing new hardware, but also installation and configuration costs. With a cloud subscription model, they can also reduce their storage. They only pay for what they use, not for what they thought they would need.”


What is the right balance of capacity to performance?

When integrators have an idea of what they want and begin designing a storage solution for customers, they often do so with storage capacity in mind without factoring in data throughput, or performance.

“The biggest challenge we have in this conversation [with integrators] is data throughput,” says Ken Mills, video surveillance strategy lead for EMC Surveillance Solutions, Hopkinton, Mass. He says often traditional integrators might know the number of cameras they need and usually the resolutions, but they often don’t have a good understanding of how all that works together to create data throughput — “basically, how much data are you cramming into a storage box at any given time.

“A data storage box is just like a network switch: if you push too much data through it, it fails. A lot of partners under-design the system from a throughput perspective. They have plenty of capacity, but not enough performance.”

Mills says integrators need to be mindful of performance when they call storage manufacturers or the partner or the customer — whoever is involved in the chain.

Mills says this lack of understanding about capacity versus performance is why many customers are moving away from appliances to enterprise storage: “They’ll have a bunch of cameras jammed into a DVR, and they have enough capacity and the software license works for the number of cameras they have plugged in. All of a sudden they have a hard drive failure. So the storage platform within the box is trying to rebuild that hard drive, and it’s trying to store video, and it’s trying to serve up video for playback that people are requesting, and it can’t handle it all. And so it starts dropping video because they designed around capacity, not around capacity andperformance.”

The results, Mills says, could be dire. “If you have a simple failover, it turns into a catastrophic failure,” he says. “A drive failing by itself is not a big deal for most systems because you have redundancy even at the basic level of RAID 5. But if you don’t design for performance, that simple drive failure turns into a complete data loss because the whole RAID controller shuts down. Then no video is recorded, and you have a very high chance of losing video that was there.”


Could cloud work for my customers?

When talking about video storage options, cloud is increasingly part of the discussion as it quickly becomes mainstream. “It’s now being widely used in non-commercial and personal applications for data storage,” says Melvin Gray, senior manager of product marketing management, Honeywell, Louisville, Ky. While many commercial organizations are using it for their data storage needs, Gray says it could also work well for video storage. “It can lower both operational and capital costs,” he says.

Larson points out that several companies such as Eagle Eye are really servicing the zero to maybe 10-camera market well. “Where cloud would work well is in the retail environment,” he says. “The problem is most Dunkin Donut stores don’t have a robust Internet connection, and they don’t want to pay for that robust Internet connection. So it is far cheaper for them to do a traditional system.”

Dolan says cost to upload will be the limiting factor. Larson agrees, saying, “The problem is the network connection. You really want a broadband connection, but most retail stores and nail salons have about a DSL-type connection. You have to consider that in the cost of your installation.”

Larson points to a cloud hybrid as a possible answer for those who want to take advantage of the benefits but whose system doesn’t lend itself to a total cloud solution. In this type of situation, a business with 100 cameras may have only five that are very critical or that might require longer retention because of compliance issues. Larson says he sees many businesses using cloud for these purposes.

For cloud video storage though, Larson says in his experience, onsite is still coming in at 30 percent the cost of cloud, without even taking into account the pipe needed on a large campus. “You would need to be buying 100 MB uplink connection or more, which would cost hundreds, if not thousands of dollars a month. So it’s still not there. When you call an integrator for a 12-camera job, I think their first thought is, ‘FLIR has this new bundle with 12 cameras and a 16-channel NVR.’ It’s not, ‘Let’s go give them a cloud solution.’”

Tom Cook, vice president of sales for North America, Samsung Techwin America, Ridgefield Park, N.J., says the reason the security industry hasn’t adopted the cloud is people are so afraid of it. “Apple created the cloud term, and right away, what do you think ‘cloud’ is?” People hear about the security issues and are scared, Cook says.

Cook offers the analogy of buying a car versus leasing. “I buy my NVR now and I put it into my building,” Cook explains. “I maintain it; I support it; I have it fixed if the hard drive breaks, etc. Maybe I don’t want to be that buyer. I want to rent. I want to lease a car — I want to lease an NVR. I don’t care where it is; I just want to lease it because I don’t want to pay for it. I’ll pay a monthly fee for storage, just like I do for my leased car, and you, Mr. cloud company, you maintain that server, you maintain that NVR, wherever it is; you tell me my data is secure and I can look at it when I want, and I’ll pay you $30 a month instead of buying it for $1,000 up front. And that’s really what cloud is.”

Regarding the fear of hacking that seems to be the most prevalent fear swirling around the cloud concept, Cook says bad guys are hacking companies for very important data: credit card history, credit card numbers, and social security numbers. “What hacker is really going to hack into the jewelry store security system that’s on the cloud somewhere? Is that really what’s happening? All these cloud access systems have preventative encryptions and secure systems, and since some of the recent hacks, they’ve put in more preventative measures — Microsoft has, Apple has, EMC has — all these companies have stepped up their game even further.”

Drako agrees, saying, “Video storage for true cloud solutions is more cybersecure than traditional systems.” He cites U.S. CIO Tony Scott in an article on “I see the big cloud providers in the same way I see a bank. They have the incentive, they have skills and abilities, and they have the motivation to do a much better job of security than any one company or any one organization can probably do.”

Drako adds, “Almost every security breach we have had in 2015 was nota cloud provider. It was a team located inside of a corporation doing it themselves.”

Because of these fears, whether warranted or not, Cook believes it is incumbent upon integrators to show customers the return on their investment.

Absolutely it’s not for everyone, there’s no doubt,” Cook says. “That’s why we have all these multiple ways to store it, whether it’s a camera, an NVR, a NAS — or a cloud.”


How do you define your terms?

Part of the reason Cook believes people are still dubious of cloud is the failure as an industry to properly define it. People are afraid of cloud, but what “cloud” conjures up in their mind might very well not be what the integrator could offer. “In the industry, we make up our own terms,” Cook says. “We made up ‘2 MP’; we made up ‘5 MP’; we make up terms all the time. I’ve always told my sales guys here, stop using made-up terms, use what the consumers and the end users understand. They understand what 1080p is; they understand what 720p is — they don’t know what a 1.3 MP is at all. They have no clue.”

There is often confusion about terms such as enterprise, distributed and centralized. Often, such as in the case of the simple failover resulting in a catastrophic data loss, the results of such confusion can be detrimental. The industry must be on the same page when it comes to educating customers and presenting information.

Mills describes the difference between enterprise class and non-enterprise class storage as “the ability to have multiple levels of fault tolerance.”

For the most part, says Marc Cisneros, senior vice president sales and marketing, HDSTOR, Phoenix, manufacturers abide by accepted classes of storage. So integrators should be asking suppliers, “What’s the difference between drives? Why can’t I use solid-state drives?” etc. These kinds of questions will prevent companies from unknowingly putting non-enterprise drives in RAID arrays, raising fail levels. “Integrators must be on the same page with manufacturers when it comes to defining what enterprise is and what it isn’t.”

A line of communication between integrator and customers is crucial. Speaking about video analytics, Dolan says, “Teaching the clients about analytics and specific information that they can collect (people counting, LPR, etc.) provides value to the client. Most end users today want to talk about it. They may not install it today, but if the integrator doesn’t start the conversation today, their competitor will come in tomorrow and discuss it.”

And so it circles back to keeping up or being left behind. In the video storage world, when all is stored and saved, it really comes down to knowledge and education. When the security systems integrators know what questions to ask and the manufacturers are speaking the same language as the integrators, the entire industry benefits. 

ONVIF Works to Future-Proof & Standardize Video Storage

Systems integrators are frequently charged with creating an IP-based video surveillance installation for a client’s facility that brings together the best possible solution, regardless of the manufacturer. Where does this integrator begin? ONVIF is an open industry consortium committed to standardizing communication between IP-based physical security products for interoperability. Integrators can choose from ONVIF’s more than 5,000 profile-conformant products for the best possible combination of products for the video system.

Knowing that a product is Profile S conformant doesn’t require the integrator to know what is within that product, only that it will work with other Profile S devices for activities such as streaming video, streaming audio, and supporting pan-tilt-zoom (PTZ) functions.

Profile G is a specification designed to store, search, retrieve and play back media on devices or clients that support recording capabilities and onboard storage. Profile G encompasses devices ranging from cameras and encoders to NVRs and video management systems, as well as building management systems, physical security information management (PSIM) systems, and other physical security systems.

Here’s an example of how Profile G can work: Profile G can be deployed between a PSIM solution integrating video playback from a networked video recorder, including specific features such as starting and ending recording; searching video using various filters such as time, event or metadata; video retrieval and playback; and, on the receiver, side, creating a source of IP media.

While Profile S and Profile G are related, the two are also independent profiles that encompass different functionalities of a network video system. Some devices and most clients may implement both profiles, such as a camera with onboard storage or a DVR. But while a camera may implement Profile S for transmission of video, an NVR would encompass the functionalities from Profile G.

For now, our integrator who is building a future-proof video surveillance system for his client’s facility will still need to do some custom work, but the existence of Profile S is helping connect more head-end devices and easing the process, while Profile G broadens the potential even further by bringing recording and storage into the mix.  —  Contributed by Steven Dillingham, former chairman, ONVIF Profile G working group and principal software engineer, Oncam, Billerica, Mass.



For more on the basics of video storage, visit SDM’s website where you’ll find the following article by SDMcolumnist David Engebretson:

“Surveillance Video Storage”

For an in-depth look at storage device options and data explaining each of the classes, visit