Now that you know what Propel Accelerator does,
you probably want to evaluate it. And, you may be comparing Propel Accelerator
to other web accelerators. Following are the key issues that must be considered
when benchmarking:
Page
load times may vary for reasons that have nothing to do with the
accelerator:
- Network and server traffic variations
- Rotating and customized content on test pages
- Hardware and modem connection variations
|
- Take many sample timings for each URL in each
test configuration
- Use averages and percentiles to summarize
results (link to section in white paper that explains this)
- Do parallel testing on multiple matched machines
- Verify that line quality and modem speed match
across all tests
|
Acceleration can vary depending on page type and content type:
- Graphics heavy vs. text heavy pages
- Number of ads on page
- Dynamic vs. static pages
|
- Evaluate acceleration for different page types
separately
|
Acceleration
can vary depending on surfing pattern:
- First time visit to page
- Return to page later on the same day
- Return to page a day or two later
|
- Evaluate acceleration separately for first
time visits, repeat visits on same day, and next day visits
|
About
Propel
