Roundtable: Surveillance experts weigh in on standards

Oct. 28, 2008
What standards can mean for end-users, integrators and product manufacturers

As the weather turns cold and we reach the fourth quarter of 2008, it's clear that one of the big themes in video surveillance for 2008 has been "standards". With adoption in IP video and a boom in the number of capture and recording device manufacturers and video management system vendors, the time was probably ripe for standards to start appearing.

And while we sometimes become very U.S.-centric, these standards weren’t just a push in North America. Sure, in the U.S., you saw the roll-out of the Physical Security Interoperability Alliance (PSIA) network video API. But in Germany this month, the Open Network Video Interface Forum (ONVIF) kicked off at Essen, opening its membership to the entire industry. And there’s reportedly strong standards activity in Australia as well as the UK in relation to video surveillance.

To help us sort out what this interest in and work towards standards all means, we reached out to three video surveillance experts.

To do so, we got in touch with Tom Galvin, principal of the firm NetVideo Consulting Inc. Galvin is the former vice president of engineering at GE’s security division and the former vice president of product development for Verint-Loronix. He is also a regular contributor to Security Dealer & Integrator and Security Technology & Design magazines (sister publications of SecurityInfoWatch.com).

Also on hand to give the low down on standards was John Honovich, founder of IPVideoMarket.info and author of the book, “Security Manager's Guide to Video Surveillance”. John is the former director of product management for 3VR Security and was general manager for Sensormatic Hawaii. He is also a regular contributor on topic of IP surveillance to IPSecurityWatch.com and SecurityInfoWatch.com.

Finally, to balance us out and give us the world perspective, we reached out to Vlado Damjanovski of CCTV Labs Inc. Vlado is an Australian-based trainer and lecturer on CCTV and author of the book CCTV (Elsevier Butterwoth-Heinemann publishing) which has often been called "The CCTV Bible".

The roundtable with these three CCTV thought leaders appears below:

Is the time now ripe for standards in network video, or should this push for network video standards have happened sooner?

Tom Galvin: The time is ripe for standards. I’m not sure it could have happened sooner. Network video had to reach to reach a level of maturity and a critical mass of user acceptance before manufacturers could be motivated to put effort into standards. Network video solutions are now widely deployed in many critical infrastructure projects. End-users are seeking interoperability from the broad number of choices of available products.

John Honovich: This is a good time as IP cameras are now moving from a niche into the mainstream. Standards will help ensure that costs are minimized so that IP cameras can successfully and quickly become mainstream.

Vlado Damjanovski: It is never too late for something that has never been done. In actual fact this is exactly what we are doing now in Australia, and as a current Standard Australia CCTV sub-committee chairman, I invite your readers to have a look at the EL51 forum site. Comments and contribution are welcomed.

What are some of the problems currently faced in network video adoption and implementation that could be solved by standards in this area?

Damjanovski: First, there are differences in various protocols, both PTZ controls, image formats and compression. But, there are also big differences in CCTV in the type of products offered, both in terms of pixel count, images per second and physical imaging chips sizes (from 1/4" up full 35mm size). It becomes more and more important to consider all components in the CCTV chain: lenses, chip sizes, pixel count, DSP, compression, latency, even display quality. Comparing live video streaming at standard definition to a 5 megapixel camera at 3 images per second (for example) is like comparing apples and oranges. This is exactly what we tried to address in Australian upcoming 4806.5 digital CCTV standards.

Galvin: The biggest problem is the lack of manufacturer interoperability between network cameras, NVRs and video management software. All network surveillance systems are currently proprietary because every network camera integration with an NVR or software product is unique. With interoperability, network video can achieve a level of “plug-and-play” that has been the norm in analog CCTV systems and similar to how we expect our printers, digital cameras, networking gear and other IT devices to work with our PC.

Honovich: I agree. The biggest problem is constraints on what cameras work with what video management systems. Compared to analog, where all fixed cameras essentially worked with all video management systems (NTSC/PAL are standards), IP video specification is significantly undermined by the lack of standards.

We have a number of groups trying to create some sort of standards for network surveillance device interoperability. Is that good or bad? Does the existence of competing groups help or hinder the standards-writing and adoption process?

Honovich: Competing groups hurts the process. On the other hand, the reality is that the standards creation process is a fight to shape the future of a multi-billion dollar industry, so such competition is to be expected.

Galvin: The fact that there are multiple groups trying to standardize reflects the strong recognition that standards are needed. It will likely accelerate the adoption of a standard.

Damjanovski: I personally don't think it is a good idea to have competing standards in preparation. Eventually, only one of the groups will prevail, and the time lost by the non-prevailing groups could be used more productively and more efficiently if everybody works together. This is the reason we (in Australia) openly invite all readers and industry experts to comment and contribute to AS4806.5. Although the AS refers to the Australian Standards, we think it is a very good starting point. In the AS4806.5 draft proposal, we have yet to define protocols and formats, but the more important thing for us at the moment is to make normal users aware of the already existing differences in products and formats and how to evaluate them. Once this passes the industry acceptance (with possible proposed modifications) we would propose (or use) some of the agreed data and image protocol(s), if they ever come out from the various groups you are referring to [such as PSIA and ONVIF].

Could more than one standard for network video be adopted? What could the impact of that be? Are two (or more) standards better than none?

Galvin: Many standards have already been widely adopted by network camera manufacturers. Many manufacturers have adopted standards for video transport (RTP) and streaming (RTSP) that have been defined by the Internet Engineering Task Force (IETF). Other international standards for video compression (MPEG-4, H.264), audio compression (G.711, AAC) and IT standards such Universal Plug and Play (UPnP) have been adopted. Both the PSIA and ONVIF are packaging a set of existing network and IT standards while defining a new application level standard to address video interoperability. The application level is where the biggest opportunity for interoperability exists. The application level defines interfaces for PTZ control, device configuration and event handling. This is currently where the PSIA and ONVIF can potentially compete or collaborate. I believe that a single standard will emerge at this level.

Damjanovski: There has been a long battle between PAL and NTSC, many products made in two different versions, a lot of un-necessary production costs. The High Definition TV puts an end to this, and today the HD is an international standard. It would be wise and desired for the small CCTV industry (compared to other bigger industries) to have one global standard. But, again, this is only possible if everybody works together. I hope this roundtable may help promote this idea.

Honovich: Adoption of two specifications should not cause significantly more problems than adopting one. Certain low-level protocols demands only one standard or systems' efficiency is severely undermined. For IP cameras, two specifications would be somewhat of a waste but would be workable.

Who would benefit the most (end-users, integrators, recorder/server manufacturers, camera manufacturers, video management software developers, chip developers, etc.) from generally accepted and adopted standards for network video?

Honovich: End users would benefit the most as standards lower costs by eliminating the power of market leaders to control markets.

Damjanovski: I think there is no doubt – everybody benefits.

Galvin: Initially, end users will derive the most benefit. When end-users benefit, the entire industry will benefit: integrators, software developers, manufacturers. Interoperability will allow the overall market for IP-based solutions to grow and benefit the entire supply chain. Video management software developers will face the most profound impact. Much of the development and value in VMS software goes into integrating network cameras, video encoders and NVR’s. With the wide adoption of standards, camera integration will become significantly easier. This will be a double-edge sword for the software developers and manufacturers. On one hand, it will free up significant development resources that are currently tied up to repeatedly perform basic video device integrations. With a standard for interoperability, camera integration will no longer differentiate software products; developers will have an opportunity to divert more resources toward value-added features in order to compete.

Why would network video standards matter for corporate security directors/end-users?

Honovich: Because it would lower prices on cameras and provide greater options on what cameras to use.

Damjanovski: Everything in the end comes down to lower cost of implementing a solution. Security directors will have a better functioning system; end users will pay less.

Galvin: Again, interoperability will allow end-users to select the best of breed products for the overall system. Corporate IT directors will appreciate that both the PSIA and ONVIF proposals are based on accepted IT and networking standards.

Do standards for network video devices help or hinder the manufacturers?

Galvin: It depends on how specific manufacturers respond. There is a lot of historical evidence to suggest that standards for interoperability are generally good for an industry as a whole. Think of consumer electronics (DVD, VHS) telecommunications (SIP protocol for enterprise IP phones) and IT (IEEE standards for Wifi, IETF standards for the Internet). If the video surveillance industry is successful in widely adopting a standard, it will force vendors to provide value in different ways. Camera integrations with a specific NVR or software package will no longer be a differentiating feature. More vendor resources will be spent on building end-user value from new features, productivity enhancements, analytics and new service models. The overall market will grow as end-users derive more value from network video.

Honovich: Standards help manufacturers with low market share including those looking to enter the industry. Standards can undermine the power of market leaders.

Damjanovski: Standards in general can never hinder a manufacturer. The fact that a device is compliant to a recognized standard makes it more attractive for the wider market. Certainly, different manufacturers may have their own way of doing things, which could be better than the others in some way, but a wisely written standard will not exclude innovation in technology, but only put a framework for common language between various products. The same is with the video compression standards. For example, MPEG2 does not define exactly how you do your compression, but defines how you put together the data so that a decoder can read it. There are variety of MPEG2 compression implementations that may have some difference in quality.

Do you think the manufacturing community would actually adopt standards on IP video devices, especially considering this industry's proprietary history?

Honovich: Yes, manufacturers will adopt standards. The industry has always used standards for camera interoperability. In analog, it was NTSC/PAL. The industry's proprietary history reflects a general poor business case for interoperability for other systems. Interoperability between cameras and video management systems has very high value and I expect vendors to be very motivated to adopt such standards.

Galvin: Yes. I think IT companies like Cisco and network cameras companies like Axis, Bosch and Sony will lead the charge. These companies have a history of supporting network standards. They also have the market leadership to drive software and NVR providers to accommodate a standard.

Damjanovski: The chances are much higher [that manufacturers would adopt these standards] if the multiple groups get together and produce one standard. By the same token, chances will be very low if there are more proposals. Time will need to pass to see which one will prevail, and that will cost somebody money, time and most importantly loss of faith in the industry as a mature one

Is it realistic for our industry to try to develop standards for IP video devices and not consider standards for IP access control and other IP-connected physical security devices at the same time?

Galvin: Yes. Again the standards exist at different levels. IETF and other network level standards are already being adopted by video and access control vendors providing network-enabled products. However, system interoperability can only be achieved with a widely adopted application standard. At the application level, video applications are different enough from access control that the standards should be able to emerge independently. In fact, the broad adoption of a video standard could help motivate the development of an access control standard. Ideally, the access control application standard would be modeled after the video standard so that they are technically compatible.

Honovich: To start, the industry should focus just on IP video devices. IP video is growing four to five times faster than any other segment of the security industry. There is no reason to slow down or muddle the process by complicating the matter.

Damjanovski: Like all standards, these things evolve. The most important thing would be to put the foundation down by having the whole industry agree to one common standard. Once this is successful, it will be very easy and natural to add everything that connects to a properly designed CCTV. Existing other standards, like lift management, access control, cash-registers, etc. (if any) will not to be re-invented, but only adopted to suit the digital CCTV. If there are none, we will have to create them. It’s a tough job, and usually unpaid, but hey, somebody has to do it.