Disaster Recovery
Surveillance plays critical role in natural and man-made disasters
- By Lee Caswell, Greg Pellegrino
- Feb 01, 2013
Hurricane Sandy disrupted power, Internet, phone and various
other technical services for millions of people and businesses
along the East Coast in 2012. In the aftermath of the storm, organizations
are reviewing the effectiveness—or lack thereof—of
disaster recovery plans across all of their systems.
Because video recording is increasingly used to monitor and manage the effect
of natural disasters, surveillance planners are especially taking care to review
internal plans for recovering from disasters, both man-made and natural.
The adoption of IP technologies across surveillance cameras, IT networks, edge
devices and hardware, such as storage platforms, opens up interesting new options
for disaster recovery since video can be stored, copied and moved across
local and wide area connections in ways that were simply not available in the
days of analog systems.
In the Time of Tapes
When dinosaurs—analog systems and VHS recording—ruled the world, surveillance
directors relied on sub-optimal methods for protecting stored video content
in the case of a local disaster. Video recording happened in one format—a VHS
recorder—and playback was limited to a VHS player running either the primary
tape or a subsequently created copy.
With this approach, disaster recovery preparedness often required physically
labeling, indexing and transporting bulky, unreliable tapes to a secure offsite location.
Creating copies of tapes was a time-consuming task and video was captured
and stored at a lower resolution or generational loss made the video suspect as
evidentiary material. Even more challenging was the fact that the entire disaster
recovery system relied on the delicate process of transporting and locating a single
physical tape.
Search capabilities in the analog world were
equally limited. Video content on a VHS tape represented
a fixed amount of recording time from
a specific camera. Content could only be reviewed
serially with a limited set of fast-forward and skip
options. The process of reviewing disaster footage
was cumbersome and required a massive mobilization
of staff at the physical tape location, who would
methodically plod through the recorded video from
start to finish, hoping to find relevant data. For example,
following the 2005 London Bombings, more
than 400 policemen were marshaled in to review
more than 2,500 videotapes.
The Role of Digital Copies in DR
IP video introduced, for the first time, the ability to
create identical digital copies of a video clip without
any loss in video content. Where tapes introduced
generation loss with each new copy, digital systems
eliminated this resolution loss by allowing copies to
be easily created and stored offsite for recovery purposes.
Today, many surveillance departments create
copies of locally recorded video on disks, DVDs or
USB drives, and then move the physical copy to an
offsite facility so that the video is not lost if the primary
recording site is compromised.
The simple creation of digital copies can result in
some unintended consequences, however. Creating
hundreds of DVDs with uncontrolled copy privileges
can be like mixing Mentos and Diet Coke, where the
results can quickly spray out of control. Multiple copies
can quickly increase unnecessary storage costs and
be nearly as difficult to track as their VHS counterparts.
Uncontrolled copy procedures also present the
unwelcome prospect of releasing sensitive video content
to the outside world. Ideally, the best practice for
DR includes a responsibility to track and control how
video copies are created, distributed and accessed.
How WAN Bandwidth
Throttles DR Options
The limited availability of wide area network bandwidth
forces a conscious examination of why video
is being retained, so that the appropriate selection of
cameras, image resolution and video retention can be
selected. In effect, an optimization puzzle has to be
solved for each use case.
For example, a retail store that is sensitive to liability
cases may choose to save offsite DR video at lower
resolution video over an extended period of time. DR
video used to protect critical infrastructure surveillance
may require exactly the opposite strategy so that
high resolution is kept offsite for a shorter retention
time to provide forensic evidence in the case that a
threat is carried out at the primary facility.
For at least the next five years, the quality of the
WAN available to a customer will dictate the video
surveillance DR solutions that are possible. Here are
three potential solutions.
For several cameras with limited WAN, copy events
remotely. IT systems frequently provide DR by copying
locally stored data across a WAN to a second and,
sometimes, third site. A small network pipe is often sufficient because most IT data changes in small incremental
bits each day. For dynamically changing
environments, replication over a WAN occurs at night
to catch up the remote site when system activity is low.
This “store and forward” DR method is completely
unsuited for surveillance environments where
a single high-resolution camera can fill a T1 line, in
which every single bit of data is new every day and
continuous video capture offers no off-duty window
to catch up.
Customers also need to remember that typical
broadband connections only offer guaranteed download
speeds. Upload speeds are carefully monitored,
hence the name ADSL or Asynchronous Digital Subscriber
Line.
In smaller installations such as retail, banking or
remote enterprise environments, there is simply not
enough bandwidth available to support any offsite
capture of real-time video. For these environments, it
makes more sense to simply copy events of interest
and send these offsite to be retained in the case anything
happens to the primary site.
For multiple cameras with a heavy-duty WAN, dual
stream the cameras. Many critical infrastructure projects
can justify a high-speed WAN that enables live
video to be streamed off-site, as well as locally. In this
scenario, incoming video is saved locally at high-resolution
quality over the high-speed LAN while a second
lower-resolution stream is sent across the WAN
to a remote disaster recovery location. This approach
matches video resolution to the available bandwidth
and optimizes storage costs since the lower resolution
off-site copy does not require the same usable capacity.
In some cases, the offsite copies are retained for
shorter retention times to further minimize the duplication
of storage, servers and licenses. DR site storage
costs can also be managed by evaluating which cameras
merit a disaster recovery approach since camera views
are often not equally critical for forensic purposes.
Multi-streamed camera solutions also are ideal for
modern corporate campuses or distributed metro installations
where high-speed networks connect distributed
buildings or depots together. Video is captured
locally onsite and simultaneously streamed to a nearby
facility for disaster recovery. This also may be effective
for companies that offer more sensitive distributed services,
such as daycare or outplacement services.
For few cameras and moderate WAN, direct to the
cloud. For sites with limited cameras and moderate
WAN bandwidth, the cloud or Video Surveillance as
a Service (VSaaS) holds tremendous appeal for customers
looking to reduce capital and operating expenses,
and for resellers looking to offer a recurring
monthly revenue service.
VSaaS can involve video sent to either a public
cloud, such as Amazon Web Services (AWS), or to
a private cloud. The public cloud offers a lower cost
of entry and support but carries the security risks of
sharing equipment and access with other customers.
Many customers remain concerned that unauthorized
access to the public cloud could allow sensitive video
content to be distributed or that a hacker could access
onsite cameras through the Internet. Private clouds
provide enhanced security by dedicating network,
storage and server resources to one customer.
For small retail environments, manufacturing
sites, restaurants and non-critical infrastructure sites,
VSaaS may offer the perfect mix of cost and offsite recovery.
For larger critical systems, the bandwidth limitations
of the WAN continue to dictate local solutions.
Infrastructure Considerations
for Disaster Recovery
Whatever the type of DR site, it is critical to rely on
high-availability infrastructure that continues to operate
during power, component or system level failures.
Pivot3, for example, eliminates any single point
of failure throughout the server, network and storage
infrastructure so that the system will continue to
operate during adverse conditions and will self-heal
during component interruptions, such as network or
server failures. Disaster recovery is a type of insurance,
and for those who anticipate the need to deploy
such a system, there is no shortcut for the type of
high-availability that Pivot3 offers across all of the
key system elements.
For DR systems based on copies of local recording,
dual-streaming video to an offsite location, or
using VSaaS to private or public clouds, the expectation
is that video data is always available. WAN bandwidth
may dictate the specific DR solution deployed,
but a common decision process on high-availability
infrastructure leads users to select systems that have
no single points of failure, which offer simple support
and seamless expansion without downtime.
Today, surveillance directors experience and benefit
from the advantages of IP-based surveillance
technology including high-resolution video, remote
and on-site video flexibility, remote management and
operational efficiency. At the same time, digital video
surveillance creates a steadily increasing stream of information,
which requires a new approach for both
storing video data efficiently and protecting it in case
of a disaster.
Whatever the size of your surveillance infrastructure,
it is critical to have a DR plan in place to safeguard
surveillance operations in the most challenging
environments because less than optimal solutions can
be costly and a source of failure. Therefore, it is critical
to understand DR options for surveillance and
which work best for your organization.
This article originally appeared in the February 2013 issue of Security Today.