Today's enterprises rely heavily on networked applications to conduct business both internally and with
"Through 2005, more than 80 their customers. Yet, complaints about poorly performing applications are common, and the
percent of application perform- time to resolve such problems can extend for months, years, or even indefinitely. New
ance and availability failures application deployments are also frequently hindered by performance problems.
will be blamed on network
problems, but the network will Performance of networked applications depends on complex interactions
represent less than 20 percent among applications, servers, and networks. IT organizations need a
of the root cause (0.7 probability)." detailed, quantitative understanding of these interactions to effi-
ciently and cost-effectively troubleshoot and deploy applications.
– Gartner Symposium
OPNET's Application Characterization Environment (ACE) direct-
ly addresses these challenges.
“OPNET’s products paid for ACE embeds expert knowledge about how applications,
themselves in just a few networks, and servers operate. ACE enables you to cap-
months. OPNET’s advanced Expedite the resolution
ture, filter, and synchronize application traces from mul-
troubleshooting technology of application
tiple network segments. Automatically merge any
enabled us to avoid unneces-
sary costs and to focus our
number of application traces and interpret them in a performance problems.
limited resources more combined view. Visualize end-to-end application
effectively.” transaction dynamics graphically, from both network-
layer and application-layer perspectives. Diagnose
Nick Dauphinais problems automatically with AppDoctor and the Solve problems cost-effectively
Manager
Network Performance industry's most comprehensive application and pro- by understanding the
and Capacity Planning
tocol decode engine. Validate your approach to
FleetBoston impact of changes.
achieving required performance levels with OPNET's
best-in-class prediction technology.
Return on Investment
Increase Effectiveness of IT Operations
Quickly identify the root-cause of poor performance with a
unique view of the interactions between applications and
their deployment environments.
Visualize
Transactions
Visualize application transaction
flows both at the application level
and the network packet level. Under-
stand the interactions and dependen-
cies among clients, the network,
application servers, and database
servers.
Diagnose
Performance Problems
Identify and diagnose performance
bottlenecks with AppDoctor's auto-
matic diagnostic reports. Decode
captured application information,
for example, to see the exact SQL
statements that cause unaccept-
able processing delays.
Validate
Solutions
Quickly evaluate the impact of changing bandwidth,
protocol settings, application behavior, server speed, and
network congestion on end-to-end response times.
Deploy applications in OPNET's virtual environment to
quantify the impact on network and server utilization.
Capture Application Traces
ACE analyzes application transactions from your testbed or production network by directly capturing packet
“OPNET IT Guru and ACE have trace files or by importing traces from popular network analyzers, including from Sniffer and NetScout prod-
identified bottlenecks attributed ucts, and from utilities such as tcpdump.
to application packet sizing and
capacity thresholds with server The ACE packet capture utility greatly simplifies the process of acquiring application transactions from your
processors as well as conflicting network. ACE's Capture Agents are easy to install from a CD, from the web, or from an email attachment sent
applications. These accurate to a remote location. ACE allows you to capture traces of application transactions on a wide variety of plat-
diagnostics were instrumental in forms including: Windows (95/98/ME/NT/2000), Solaris, Linux, HPUX, and AIX.
preventing costly network
upgrades and extensive server MANAGE CAPTURE AGENTS
replacements. In each of these From the central ACE Capture Manager console, with one click, you can simultaneously trigger collection by mul-
instances, prior to OPNET, the tiple agents. End-to-end transactions can then be analyzed from different independent network segments.
initial indicators pointed to the
network as the root cause while INTEGRATE WITH APPLICATION MONITORING
all of the metrics and measure- The OPNET Distributed Agent Controller (DAC) Module, priced separately from ACE, enables capturing multi-tier
ments of the network indicated application traces using the free OPNET Capture Agents without the ACE Capture Manager. DAC is used in con-
sufficient available bandwidth.” junction with synthetic transaction generators that monitor response times in a production network. DAC can
Russ Rodrigues automatically capture and archive traces of scripted application transactions in an unattended mode. These
VP, Business Technology Planning
archived packet traces can now be used to analyze, using OPNET's ACE Module, performance problems that
Banknorth Group, Inc.
are intermittent or that cannot be easily reproduced. Typical commercial synthetic transaction generators that
DAC integrates with include: Mercury Topaz and Loadrunner, Segue SilkVision, and IBM Rational Robot.
ACE transforms the trace file into an
SYNCHRONIZE MULTIPLE TRACES
Application Message Chart and a Net-
work Packet Chart for a full picture of ACE is unique in its ability to automatically syn-
your application's performance. chronize multiple application traces of the same
transaction across different network segments.
APPLICATION MESSAGE CHART
shows the application requests and This essential feature makes ACE best-in-class at
responses between the server and visualizing and diagnosing the root-cause of end-
the client. In this example, the server to-end performance problems for multi-tier appli-
is sending a large application
cations. This feature is especially useful in
message to the client.
distributed production environments that prevent
NETWORK PACKET CHART full capture of transactions from a single loca-
shows packets at the network level.
tion.
In this example, a large application
message is composed of many
network packets.
ACE enables you to visualize application behavior by parsing an application trace file consisting of transac-
tions captured from a test-bed or a production network. ACE then graphically reconstructs application behav- “If you consider the time that
ior on any network segment into intuitive, powerful diagrams. ACE displays transactions at both the application we’re saving end-users, the
message level and network packet level to provide a clear view of the important dynamics. amount of CPU utilization we
cut, and the speed with which
Transaction visualization facilitates the quick resolution of performance problems in complex applications, we got the problem solved, you
before or after deployment. can see how a solution like ACE
can pay for itself pretty quickly."
ISOLATE APPLICATION MESSAGE FLOWS
Decompose the application transaction to collect Jack Mundale
Communications Network Specialist
information such as application data, protocol US Coast Guard
overhead, error messages, retransmissions, and Operations Systems Center
inter-arrival times.
In addition to powerful visualization capabilities, ACE enables several techniques for diagnosing end-to-end
“As it turns out, expensive performance problems. ACE applies expert knowledge to captured application data for automated trou-
upgrades were not required. bleshooting. Sources of delay are summarized in convenient diagrams. Thresholds for key application sta-
By optimizing server configura- tistics are used to generate informative repor ts that characterize problems. ACE also provides
tion and TCP settings, network recommendations that support fine-tuning of specific application and network parameters for achieving
resource efficiency was doubled, improved performance levels.
and average response time
reduced by more than 10%.” SUMMARIZE COMPONENTS OF DELAY
ACE generates summary reports that clearly show the factors that contribute to end-to-end response time,
"OPNET's software enables us
including transmission, propagation, network congestion, protocol overhead, and processing on each appli-
to avoid unnecessary costs and
cation tier.
improve application perform-
ance." AUTOMATE DIAGNOSIS OF PERFORMANCE PROBLEMS
ACE automatically analyzes each application tier to generate a summary of potential bottlenecks. Find out,
Jun-Min Liu
District Manager for example, if "chattiness" or TCP windowing between a specific client and server are causing bottlenecks.
Business Solutions
AT&T ANALYZE APPLICATION STATISTICS
ACE’s reports include a broad range of important numerical statistics for each application, including pro-
cessing delays, network delays, response times, number of application turns, number of application mes-
sages, and more.
Used with OPNET's IT Guru or SP Guru, ACE provides visibility into the performance of applications in different
deployment environments. IT Guru and SP Guru offer advanced analytical modeling for quick results, as well “Organizations benefit significant-
as advanced hybrid simulation technology to facilitate more sophisticated analysis of response times in var- ly by simulating and then optimiz-
ious “what-if” scenarios. ing performance of applications
prior to implementation, cutting
RECODE POORLY WRITTEN the high costs of system down-
APPLICATIONS "VIRTUALLY" time and lagging access of poorly
Evaluate the impact of changes to application design optimized applications when the
performance using QuickRecode. system actually goes live.”
Using OPNET IT Guru and SP Guru, analyze end-to-end performance, and take into consideration the complex OPNET IT Guru and SP Guru predict SLA
compliance with increasing application
relationships among protocols, network devices, servers, and applications. Evaluate end-to-end response
usage over time.
times; optimize server placements; analyze the impact of increasing application usage; and verify conformance
to Service Level Agreements (SLAs). Quickly make changes to your virtual network and/or application models,
and then deploy your application in the production environment with confidence.
Awards
Network Computing Winner of the
Company to Watch 2004 SUPERQuest Award
for Best E-Business Solution
at SUPERCOMM
FORBES
Best 2003
OPNET Recognized
for Financial Performance
Winner of NetworkWorld’s
OPNET Technologies, Inc. Best of the Tests Award
7255 Woodmont Avenue
Bethesda, Maryland 20814 USA
phone: +1 (240) 497-3000
fax: +1 (240) 497-3001
Winner at Networld+Interop
and COMDEX 2002
email: info@opnet.com
Best of Show
NASDAQ: OPNT
www.opnet.com
The OPNET application packet capture utility includes software developed by the University of California, Lawrence Berkeley Laboratory, © 2004 OPNET Technologies, Inc. OPNET, IT Guru, SP Guru, and OPNET Technologies are trademarks or service marks of OPNET.
and its contributors, as well as software developed by the Politecnico di Torino, and its contributors. Other trademarks or service marks appearing in this document are the property of their respective holders.