Application Profiling
Optimise application response with application profiling
services from Data Integration
Application development projects are usually, large, costly and
complicated, involving many cross-functional teams. One of the
objectives that needs to be met is to ensure that application
response times meet an acceptable standard - but this can be
difficult to predict and control because of the many different
components of the delivery mechanism that affect response
times.
Why use Data Integration Application Profiling Services?
Our Application Profiling services characterise, analyse and
predict the response-time performance of an n-tiered, TCP-IP
application (client-server or web-based). We provide you with rapid
insight into how your application, network and individual
client-server components interact as an end-to-end system to affect
end-user performance. We then provide recommendations to enable you
to optimise user response-times.
This service helps you to deploy and support applications more
efficiently and effectively.
The best results are achieved if you include application
profiling at the early stages of a project, but we can assist you
at any stage of the application life-cycle and are confident we
will deliver tangible benefits.
Pre-deployment Stage
- Helping planning teams understand how different application
deployment options (e.g. client-server vs. web-based and/or
distributed vs. centralised servers) affect application response
times
- Assisting development teams understand how the application will
perform and scale across a corporate wide-area network or the
Internet and identifying where performance bottlenecks exist within
the application code or other components of the delivery
mechanism
- Capacity-planning WAN bandwidth for optimised performance
- Comparing alternative client-server or web architecture design
options
- Characterising how sensitive application performance is to
changes in network bandwidth, load and latency
Deployment Stage
- Assisting your business to base line or benchmark the day-1
response-time characteristics of the application
- Characterising the sensitivity of application performance to
changes in network bandwidth, load and latency
- Setting Service Level Agreement objectives based on transaction
response-times
Production Stage
- Troubleshooting performance issues
- Understanding the effect of changes to the network or
client-server architecture
- Understanding the effect of changes to WAN bandwidth
- Comparing new software version releases
Benefits
- Objective information enables you to make informed decisions
based on user performance
- Fast and efficient - delivers application and network
performance optimisation and/or troubleshooting information in
shortest possible timescales with easy-to-understand, concise
reports
- Pinpoints the root cause of performance issues accurately -
avoids costly time-wasting and "blame culture"
- Eliminates the need for guesswork in understanding how to
resolve end-to-end performance issues
- Totally passive methodology no business risk
Service Options
On-site engagement
On-site engagement for a single 2/3-tiered application
project:
- Typically10-15 user transactions profiled
- Timescale - typically 7-10 days (to delivery of report)
Off-site engagement
This option requires you to provide Data Integration with
network traces of the application transactions to be profiled. The
traces are typically taken by a customer with network analyser
(e.g. Sniffer or equivalent).
Based on number of trace units:
- 1 trace of a single transaction across a single tier
client-server/server-server conversation = 1 unit
- Minimum project scope of 10-units (based on 5 application
transactions and 3-tier client-server architecture)
- Timescale typically 3-5 days (to delivery of report)
