The Importance Of Application & Protocol
Behavior Testing To Accurate Network Assessment
Author: Barry Koplowitz
Awareness of performance problems in a large enterprise network
comes from different sources. Sometimes it is a monitoring tool
and sometimes it is complaints from business users.
Monitoring tool notifications stay within IT--quietly. With all
the alligators to shoot, such issues may be left for a quieter
time. Nevertheless, since those alligators reproduce so
quickly—that quiet time may never come. Besides, such tools
usually focus on "Server Down, Switch Down" alerts and cannot
perceive the far more subtle signs of a problem.
Business user complaints are visible -- frequently high profile
-- involve Business User engagement. Such problems are discussed
in management conferences and must be addressed
quickly—and—successfully. Even if there are no alerts from you
monitoring system—they must be addressed. After all, those
automatic alerts (if they triggered at all) are not really
focused on the "User Experience". When a human being is having
trouble performing their job, which feeds their family, for any
reason—they do not want to hear statistics.
A standard Network Assessment is usually a review of Network
Architecture -- Bandwidth Utilization / Errors. All important
information—but third hand information at best.
Business Users see & feel how the Application ITSELF utilizes
the Network—how it actually rides your LAN and WAN. This is
first hand information.
Unfortunately, this perspective gets minimal attention in a
standard Network Assessment process.
This assessment methodology provides a two-pronged approach.
Simultaneously assess:
- Network Architecture
- Bandwidth Utilization
- Errors
AND...
* How the critical Applications utilize the network themselves
-- from the Packet Level.
* Use tools such as Sniffer / Ethereal / WireShark as well as
many other proprietary and open-source tools.
* Identify All delays regardless of source (Network,
Application, Servers, Database, OS, etc.)
* Quantify this data to the byte and the microsecond.
* Compare these results to the data from our Network
Architecture Assessment.
* The Interpath of every application is reviewed -- from Switch
Port to Switch Port -- Router to Router.
* Network Protocol behavior is evaluated regarding Network
Transport times, TCP Tuning -- discovered issues.
This two-pronged approach -- capturing the actual way
Applications behave on the Network -- from multiple locations --
and comparing it to the Network Architecture -- provides Clear
Understanding of the causes of the problems and provides a path
for resolution.
- It is a "Departmentally Neutral" approach.
- It does not assume the problem is network based -- although
it analyzes the network.
- It does not assume the problem is the application -- although
the application's network behavior is measured.
All results are quantified and objective. They allow all
departments come together onto the same point of view --
resolving the technical problem -- as well as the political one.
In many cases, the problem is one of balance. Neither the
Network -- Application -- or any other single source is the
cause. If you are only looking from the perspective of the
Application Support Department, or Network Support—you will not
discover issues that span multiple areas of responsibility.
This Assessment Methodology will uncover the truth regardless
of the cause of the problems. Then diplomatically provide
detailed documentation and guidance (if requested) on a path
towards Resolution.
About The Author: Barry Koplowitz founded Interpath
Technologies Corporation http://www.interpathtech.com in 1999.
Barry spent 3 years with Network General and NAI traveling the
United States teaching for Sniffer University ®. Since then, he
has worked consulting to large enterprise environments in the
area of in the area of Enterprise Network & Application Analysis
and Troubleshooting. He is the writer and host of The Sniffer
Guy http://www..itunes.com/podcast
