Features Business & Marketing
Ask the Expert: Strategies to ensure high network uptime

For mission critical applications, you may seek so-called “five nines”
availability, which represents an uptime of 99.999%. This level
reflects a five-minute downtime over the course of a year, and there are
associated costs for this high availability. For other scenarios,
99.99% — or 52 minutes of annual downtime — may suffice. In any case,
open-based digital systems offer far higher reliability than DVRs,
which may be offline eight hours or more per year. Whatever your network
situation, consider the four strategies that follow:


Outsourcing
High-end customers that can’t afford downtime often choose an
outsourcing strategy to ensure servers and networks remain up. There
are two approaches to ensuring the highest levels of availability.  You
can outsource the network structure to a co-location provider or build
it and maintain it yourself. Using co-location providers, which have
been around for more than 10 years, is generally accepted to be a more
cost-effective approach. Co-location providers operate physically
secure buildings with redundant capabilities (Internet connections,
power, etc).

These hardened facilities monitor the network 24×7 using
advanced software tools and highly trained staff. Providers implement
Border Gate Protocol 4 (BGP4) protocols to make routing decisions based
on path, network policies and/or rule sets. The bottom line is that by
outsourcing network infrastructure to the pros, you can obtain the
highest level of availability without having to deal with networking
hassles.


Isolating staff from equipment
If you house your own networking equipment, physically separating gear
from employees is critical in ensuring network uptime. I recall one
customer, who told me his firm suffered outages every night. Using a
network video camera trained on its equipment, he discovered that
cleaning crews were unplugging the router so they could use the outlet
for a vacuum cleaner. Another healthcare customer discovered that his
staff was using the network closet as a break room. Restricting access
to these areas will eliminate unintended incidents. Beyond physical
security, ensure that you implement the proper computer security.
Implement firewalls to protect the integrity of the system. For
wireless systems, ensure you set up WEP or WPA encryption for the
routers. In addition, use Secure Sockets Layer (SSL) encryption
techniques to protect data to protect against those trying to sniff
your network. Finally, use uninterruptable power supplies (UPS) for all
networking gear. The large majority of power outages are of short
duration, so these can protect against almost all power incidents.


Redundant networks
When I worked at SunGard Availability Services, a
business continuity provider, we had to pay attention to common
problems that interfered with network up-time. Interestingly, it was
often low-tech problems, like a backhoe cutting a network line while
digging a trench, that cropped up. Such examples point to the need for
redundant paths for the data to travel. In a sense, you want to create
a circle, so that if one path is cut, the network can repair itself and
go in the other direction.

Further, it’s safest not to send all data to
a central repository immediately. Some of the largest commercial
companies and government agencies have the network set up to move data
from distributed cameras to a regional center for storage and then
trickle up to a central location. Thus, if the network drops, they have
data in several spots (something you can’t do in the analog world.)
With network video systems, you can have redundant servers, storage and
routers (Layer 3) using Hot Standby Router Protocol (HSRP). For outdoor
wireless environments, you may set up a wireless mesh network that
provides for self-repairing nodes for reliability and redundancy.


Store data at the edge
The final strategy includes storing video data at the edge – in this
case the network camera. In this way, if the network fails, the local
camera can continue to store images locally. During that critical time,
network cameras will still operate, taking pictures and storing them.
One caveat —this isn’t a fully developed strategy yet. Today, if you
want to access the data, either you must take out the SD card
physically and read the data on a computer, or you must log in directly
to the network camera.  Video management systems will typically not see
this edge data as part of their security database. Still, that is
better than not recording and storing while an incident occurs. In the
future, video management software companies will provide an automated
system so that when connectivity is re-established, the software will
pull stored images into the network so you have a full library.

Whether you require a “five nines” strategy or something less
stringent, using these strategies will ensure that customers not only
have an outstanding setup, but can successfully record, store and
access important video data in case of human, equipment or
environmental failures.