A traditional mid-size police department with just a couple hundred officers typically will have about 100 terabytes of digital evidence, enough for a couple of robust servers to handle. But that’s changing quickly. Add in higher quality data from body-worn cameras and in-car cameras, crime scene footage, surveillance cameras, highway patrol helicopter cameras, video captured on smartphones and, soon, drones armed with cameras, and the organization now can have as much as three petabytes of video data that needs to be managed, stored, transmitted, secured, archived, analyzed and searched. All of a sudden, that mid-size police department has the largest IT platform of any IT organization in the city.
The police department, like many agencies in the public sector as well as enterprises, is caught in what I call the “video vortex,” a proliferation of new devices that are generating massive amounts of higher quality digital data that is overwhelming the traditional ways that these organizations have stored and managed their data. The challenge is acute for many in the commercial and public sector arenas, which must find new, cost-effective ways to address the oncoming wave of video data. For public safety agencies, they also must manage retention policies for this evidence data, which in many cases can require that such evidence is preserved for 10 years or more.
Traditional ways of managing surveillance video and evidence data won’t work in this increasingly digital age. Simply moving from closed circuit TVs to IP video has put a strain on the closed systems or appliances that typically have stored surveillance video. If an agency transitioned its 100 cameras from standard definition to 4K, the bandwidth, storage and compute requirements would multiply by 12X. Layer in the video from myriad new sources and the demand for more processing power and network and storage capacity grows significantly.
The momentum behind the video vortex will only accelerate. The surveillance market is growing almost three times faster than the broader IT market, and according to a recent survey by MeriTalk, 99 percent of federal IT and security decision makers say video surveillance technology will improve their agencies’ ability to fight crime, theft, and terrorism over the next five years. These agencies are using video data to drive greater insights into such things as suspicious behavior, object recognition, traffic monitoring, incident reporting and face matching. By 2020, video surveillance is predicted to grow to about 3.3 trillion captured video hours worldwide.
At the same time, the requirements around evidence data also are evolving, and cities now must consider the best technologies for supporting the retention policies for this data. The type of crime dictates how long the evidence needs to be stored and accessible. In some federal crimes, police departments are required to keep the evidence forever, while a domestic violence case may require a retention period of five years and a misdemeanor, 30 days.
Value of an Open Architecture
Simply putting the data into black boxes just doesn’t work anymore. What used to take one black box before 4K video now will take 12, and managing all those boxes – at a larger enterprise takes hundreds of boxes. This just isn’t cost-effective, scalable, reliable or secure enough. It’s also difficult to analyze the data when it’s locked away in a box. While almost all federal IT leaders in the MeriTalk survey said surveillance video technology is key to their future efforts, 54 percent of the video data now goes unanalyzed. The industry has largely embraced Internet protocol (IP) surveillance – about 76 percent of federal surveillance cameras are now IP – making it the standard where once it had been the exception. However, for many agencies, the deployment often still has the look of a closed system, with video management applications tied to the appliances they’re packaged with, which forces customers to upgrade the hardware and software in tandem and locks the value of the data in a single box.
That needs to change. The number of devices will only grow, and the amount of data generated by those devices will increase rapidly. A key theme we’re seeing in the surveillance space is that surveillance data is becoming an enterprise application, and organizations need to start treating it and deploying it like you would any data center-class operation. A law enforcement agency in a major U.S. city is facing this challenge right now. If we were to design the system the way they would have done it in the past, the department would need 700 servers with 150TB of capacity each, a management nightmare. Another option is five servers with 50PB of capacity behind them, a much more manageable number.
Public sector agencies and enterprises are looking for an architecture that will enable them to unlock the value of all this surveillance data they’re accumulating. They’re moving away from appliances that contain the necessary compute, storage and software and moving toward a more open, data center-class IT infrastructure. So what does that look like? Such an infrastructure needs application virtualization technology, enterprise network infrastructure for the campus and software-defined networking (SDN) for the core.
These organizations also will need enterprise-level servers with enterprise storage platforms behind them. This infrastructure should include security software, high-end desktops, and 4K monitors to display the video and an open platform for greater flexibility and cost efficiencies. The platform must be able to not only address demands in the core data center – where most of the data is stored, archived and analyzed – but also reach to the network edge, where there are some video management and storage done closer to the cameras themselves.
Having an open platform is a key part of this. Our customers overwhelmingly tell us they want open platforms so they can pick and choose the software that best fits their environments, and being able to choose the surveillance apps they need is a significant benefit of an IT-grade infrastructure. They are not locked into a vertically-integrated solution that requires hardware and software upgrades every time they want to make a change.
Such an infrastructure also makes the surveillance data portable. Once that happens, the data can bring much more value to the organization – for example, it can be moved between on-premises and off-premises boundaries and shared across applications. Moving away from the silos of the black boxes and embracing a more enterprise approach to the data enables users to create a “public safety data lake,” a single platform where all of the video data and evidence can be stored and managed. From here, the data can be more easily accessed, searched and analyzed.
Now, agencies can more quickly determine what’s happening where and to whom and make better decisions around everything from staffing to resource allocation. Search is more efficient because the data from the disparate sources are all in one place, and agencies can more easily apply such applications as facial recognition to the data. The combination of video data and big data can deliver real-time access, instant event search, the ability to connect with other agencies for real-time surveillance and a greater chance to capture criminals after an incident occurs.
Evaluating Vendors When There is a Cloudiness of Cloud Options
Organizations that want to implement an enterprise solution will want to be careful when choosing a vendor. They will need a vendor that has a broad and open portfolio, given the myriad components that are needed to deploy such a platform, from servers and storage to networking, security, and virtualization. They also will want to ensure that the vendor they choose actually tests and validates its products and provides implementation and sizing guidelines. Sizing data is key – with poorly designed solutions, you risk spending too much on technology you don’t need or developing a platform that can’t handle your data needs. Customers also want a vendor that has a proven commitment to the surveillance industry, and that includes having customer references.
Original Equipment Manufacturers (OEMs) often work with larger vendors to bring their unique solutions to the market. There are numerous OEMs offering specific appliances and solutions for the surveillance industry, but all are not created equal. Look for companies that use reliable, best-of-breed, industry-leading hardware as the foundation of their solutions. Vendors also should provide – or have partners who can provide – a range of services such as assessment, consultation, design, cost-benefit analysis, and integration, as well as ongoing support.
The cloud also is becoming an option for some customers, and the same rules should apply to the cloud as they do to any other infrastructure – is the cloud environment scalable, tested, validated and proven? What about data? Do you own the data if it’s in the cloud? Those kind of domain issues is a big deal in evidence data more so than with a lot of other data.
The key to the cloud is understanding the costs behind it, especially as it applies to surveillance and evidence data. In the surveillance arena, the cost lies in the bandwidth. Consider a larger organization that has 4K video on their hundreds of surveillance cameras. The bandwidth needed to send all the data to the cloud and then back from the cloud when you want to play the video back requires a significant investment in the networking infrastructure to make all this happen, so all the cost savings the organization may see on compute and storage by going to the cloud can be negated by the investment needed for networking. Private clouds are a more viable option for surveillance video because most of the data can be kept on-premises, with the most important archived clips being put in the public cloud for disaster recovery purposes.
Convenience and cost also are issues when considering public cloud and evidence data captured by body and in-car cameras or crime scene footage. The cloud is friendly because most of the time you’re talking about backing up files from onsite to off-premises and not streaming the video. But let’s say a city pays a cloud storage fee of $55 per month per body camera that officers wear, and there is an officer-involved shooting. The body-cam video becomes evidence that the department is required to keep for 15 years, so now the city has to pay $55 a month for 15 years for that one incident. So when you do the math over the long-term, public cloud costs can be daunting, so the pendulum is swinging back toward hybrid environments. Surveillance and evidence video data are one of the few instances where the cloud isn’t the obvious answer.
A final key to handling the influx of data and devices that make up the video vortex is collaboration between an organization’s physical security managers and IT leaders. According to the MeriTalk survey, 79 percent of respondents said better collaboration can ensure a successful video surveillance program, though the survey also showed gaps between physical security and IT managers in whether the responsibility is shared (only 33 percent of IT managers said so) and what the top technical challenges are.
But the benefits of collaboration are clear. Eighty-one percent of agencies that require collaboration say they are more prepared for the influx of video data are two times more likely to operate an edge-to-core architecture, are more likely to analyze more than 50 percent of their video data, and are more likely to implement mobile video surveillance on vehicles, drones, and people. Communication is critical.
About the Author: Ken Mills is the General Manager for Dell EMC Video Surveillance Solutions.
