|
CSC builds, operates, and maintains
large-scale communications networks for
commercial enterprises and government
agencies. These networks, which carry
mission-critical voice, data, and imagery
files, must perform optimally and with
minimal interruption. To determine the
correct network capacity and topology for
these networks - most of which are based on
router and intelligent multiplexer equipment
- we use a wide variety of modeling and
simulations tools.
These tools automate the time-consuming
engineering problems associated with the
design, optimization, and performance
analysis of communications networks. They
provide quantitative answers to specific
questions about such matters as backbone hub
optimization, optimal locations of
intermediate points, tail-end circuit
connectivity, circuit costs, optimal number
of sensors, and voice, data, and video
performance.
Our expertise in these tools and
procedures allows us to test many different
configurations until we find the optimal
design. They allow us to perform
cost/performance trade-off studies to help
you determine the best network architecture
and configuration for your needs. For
example, we can use your e-mail data - such
as the size, frequency, and image content of
the messages - to generate a workflow
analysis that projects peak usage hours. We
also can provide you with cost/performance
trade-off studies if you are expanding your
networks or you plan to add applications to
your existing networks.
For more information,
please contact
Kumait Jawdat,
Business Development.
What We
Offer || Success
Stories
Analytical Modeling
Analytical modeling uses mathematical
algorithms to calculate exact measures of
network performance. CSC's analysts
are careful to collect all appropriate data
and verify its correctness before using it.
We also ensure that the correct algorithms
are used to solve each problem. Once the
data and algorithms are ready, we use tools
like the following to develop analytical
results:
- Graphical Radio Network Engineering
Tool (GRANET) to design analog and digital
wireless systems; and
- CSC's Reliability and
Availability System Predictor (GRASP) to
analyze system reliability.
Heuristic Modeling
Heuristic modeling takes user-defined nodes
and traffic loading matrices then
automatically and interactively generates
feasible network topologies, routing traffic
over these topologies, sizing
interconnecting trunks, calculating link
costs from tariff databases, and estimating
network performance characteristics
iteratively. CSC's heuristic modelers
start by collecting and verifying all
appropriate inputs. Next, they review the
technologies they wish to study and identify
"what if" questions to be answered. Then,
they use one or more of the following tools
to perform heuristic modeling:
- Asynchronous Transfer Mode Network
(ATM) Engineering Tool (ATMNET) to design,
optimize, and size ATM networks;
- Mind-Data to design, optimize, price,
and predict performance of multi-drop and
X.25 packet-switched networks;
- NetMaker XA to design intelligent
multiplexor and router-based wide-area
networks;
- Polygrid and Non-Polygrid Voice
Modeling Tools to design and optimize
voice networks based on a user-specified
grade of service; and
- Network Resource Planning (NRP) tools
to determine the best quality of service
at minimum network cost.
Simulations
Simulations employ discrete-event simulation
techniques to predict the performance
characteristics of computer and
communications systems. They are often used
to validate designs generated by heuristic
tools. CSC's simulation experts start
by identifying all parameters to control the
simulation, selecting the number of devices
to be simulated, and determining the outputs
to be generated. Next, they specify the
characteristics of the servers, hosts, and
workstations. Then, they describe the
activities performed on each platform and
specify the characteristics of each
application. After that, they identify
network data transfer, send, and receive
rates as well as specify network
connectivity. Once all this information is
collected and verified, our experts obtain
their results using a variety of simulation
tools, including the following:
- Block-Oriented Network Simulator
(BONeS) to simulate discrete events;
- Scientific and Engineering Software
(SES)/Workbench to simulate performance of
hardware, software, and systems; and
- Multi-Switch Simulation (MSS) to
simulate ATM, circuit-switched, and
packet-switched networks routed over
ground-to-ground or air-to-ground media
that will be combined into a single
simulation.
On
the
New York Stock Exchange (NYSE) our
modelers used CSC's Reliability and
Availability System Predictor (GRASP) to
ensure system reliability and redundancy.
First, we performed "what if" scenarios on
numbers of electronic mail nodes, message
sizes/frequencies, and other information to
project peak usage. We then determined the
correct size, number, and locations of mail
servers so that communications lines could
be constantly open between the Exchange
floor and back offices. Next, we calculated
the number of hand-held devices needed for
brokers to transmit trade information
without interruption. This detailed work
allowed us to prove that the NYSE network
will survive under any scenario short of
nuclear war.
On the European Composite Health Care
System (EuroCHCS) program (1996-1998),
CSC has studied and made
recommendations regarding the traffic impact
of migrating three Military Health Services
System (MHSS) networks to existing wide area
networks (WANs) in the European region. We
started by using our Study Tool for the
Assessment of Technology (STAT!) to identify
the functional data flow requirements
generated by new Internet traffic as well as
the three migration application systems -
Defense Blood Standard System (DBSS),
Ambulatory Data System (ADS), and the
Off-Board Server (OBS). We then quantified
WAN information transfer requirements for
each Medical Treatment Facility (MTF) within
the European region during a high-traffic
scenario. Next, CSC used NetMaker XA
to identify connectivity requirements for a
high-traffic scenario which linked the MTFs
to the Unclassified (but Sensitive) Internet
Protocol Router Network (NIPRNET). We
related projected system traffic impacts to
patient workload, identifying on a
link-by-link basis the total bandwidth
required to support network access. We also
determined link utilization based on the
total traffic load and existing access
circuits. CSC's work showed our
customer that the existing WANs could handle
all new traffic demands with just a few
modifications, saving them the cost of an
unnecessary investment in a major new
system. |
