At What Cost?
Keeping overall costs down and driving value up when installing video analytics
It is widely accepted that video
analytics provide value to the end
user, especially in security applications
such as perimeter protection and
the identification of vehicles versus
humans in unauthorized areas. But at
what cost? The real answer lies beyond
dollars and cents and is found in the
realm of costs in terms of manpower,
engineering, training, deployment design
and implementation method.
Today, many analytics vendors market
and sell proprietary solutions that
consist of software that must be hosted
on a server or on servers the customer
must supply; edge devices of the provider's
own manufacturer, like IP cameras
and encoders; or on a combination
of the two.
The commonality is the closed-end
approach to providing intelligent functionality—
basically it's an additional
system on top of systems you already
have. And an additional system means
more engineering, another GUI to
learn, additional time to implementation
and, consequently, higher costs.
This perceived value of analytics
causes potential buyers to remove
them from their purchases because they
would rather not expend all that effort.
Real Savings
We can easily conclude that any customer
solution will ultimately cost less if
it does not require as much manpower,
engineering or training in order for users
to become productive. But let's talk
about implementation method for a moment,
because the only true path to real
savings in the analytics world is to not
buy an analytics system but rather a system
with analytics inside.
What does this mean?
Today, intelligent functionality
doesn't have to come from a proprietary
stand-alone system. Rather, it can be a
small but effectively applied ingredient
to the overall solution. All within the
same video surveillance system, analytics
can reside at the edge where network
bandwidth may be a concern. They can
exist in back-end servers or storage
devices so the analytics can effectively
be cross-assigned to different video
channels based on alert level. And intelligence
can even exist in the middle
of the network, in a router. To increase
quality of service on the network, intelligent
routing of alerts and analytics
functionality could be needed based on
time of day or a requirement to push
alerts to additional responders.
From an increasing number of software
providers, the analytics and/or
alerts are accessed using the GUI of the
video management platform or physical
security information management tool.
With this significantly shorter learning
curve, implementation is completed far
sooner and users move quickly to the
business of solving problems instead
of worrying about how long it will take
to learn a new system. Already having
chosen a management platform, video
analytics functions possess equal standing
with many of the other management
tasks your platform provides.
Back to Basics
What are the building blocks of
these kinds of open, analytics-enabled
solutions?
The analytics must be flexible. That
is, they must be available to the integrator
to spec into the solution in a wide
variety of form factors. To get the most
efficiency from the ingredient approach,
the solution provider should be able to
acquire analytics not only in cameras,
encoders and DVRs/NVRs, but also in
servers, enterprise storage devices, routers,
other network devices, PCs, snap-in
daughter cards, USB dongles and discrete
intelligent appliances.
With these device choices, along
with a requirement that the analytics
be available in capability packages designed
specifically for the customer's
vertical market needs, the user achieves
high value while the integrator can
price accordingly, since he or she can
sell the customer only the functionality
and performance they initially
need. Additional analytics capabilities
can be upgraded in the field via new
license keys, further enhancing the value
of the solution.
Practicality is key, or the customer
won't see any return on investment. There
are still too many analytics vendors out
there who tout features and functions
that would be truly innovative if they
actually worked reliably outside of a
controlled lab environment. But practicality
and cost dictate we face the facts
of realistic, not lowered, expectations
of analytics. What would be preferable:
features that only work in PowerPoint
with an unfamiliar GUI, or the reliable
and straightforward capabilities we're
already accustomed to working with,
presented within the management platform
you're already using?
Bring these building blocks together.
The interoperability of disparate parts,
through the execution of an industry
standard for analytics configuration,
rule definition and alerting, will allow
users to gain real value from their solution
sooner and at a lower cost. The
Physical Security Interoperability Alliance
has already taken the first step toward
realizing a world of pre-integrated
solution components by forming a specific committee to work on a video analytics
standard. The net result would
be much greater adoption of analytics,
which would, in turn, drive down costs.
Users must keep demanding practical,
easy-to-implement analytics-enabled
solutions from a trusted installer
or integrator. Ask them to take a look
at what's available today—the building
blocks are beginning to show up in good
numbers from manufacturers and software
providers. These manufacturers are
trusted names in the industry, and they
understand full well the importance of
delivering open-architected solutions.
Integrators should push OEM suppliers
and distributors to offer a greater
selection of intelligent devices, with the
feature packages and licensing flexibility
they know they can sell.
Video analytics as a high-value ingredient
in the video solution renders
the popular edge versus central debate
completely moot. To achieve the lowest
cost of ownership over time, analytics
should reside wherever they represent
the greatest value to
the end user.