|
|
Summary
eValid brings a fundamentally different
technology to the work of monitoring WebSites.
This page outlines how eValid technology compares
with alternative monitoring methods by identifying the
kinds of data that is not made available by the weaker methods.
(See the Technology
Level Comparison diagram.)
This is a technical perspective. See also the Medical Analogy, which describes this same information from an informal perspective.
| Type of Action | Object | Action Description | Percent of Reality | Commercial Aspects |
| PING | Domain name | Only confirms at that target hardware running. No data transferred. | 0.001% | Free or near-free.
Basic warning of health only. |
| HTTP | URL | Page gets with HTTP are memoryless. Data transfer is with that URL only. No parsing and no subsequent analysis of component URLs. | 0.1% | Quite inexpensive.
What many monitoring services offer. |
| Augmented HTTP | URL sequence | Sequence of independent pages to simulate transaction | 1%-5% | More costly than simple HTTP monitoring.
This is what most monitoring services offer and they often call it "transaction test" even though that nomenclature is misleading. Error range can be 100:1 depending on the selection of URLs made. |
| Augmented HTTP with Secondary URL Retrieval | URL sequence plus dynamic analysis. | Each URL is parsed for download of extracted component URLs. | 5%-10% | Very few (if any) monitoring services
do this because of the extra work involved
and thus the extra expense.
Parallelism of page retrieval not recorded. |
| HTTP from recorded session. | URLs actually downloaded. | The recorder is a passive tap on the HTTP/S protocol, so all downloaded URLs are included. | 10%-20% | Even though all URLs are downloaded, they are not downloaded as the browser does it -- with multiple parallel execution threads. Instead, the download is serialized, and this can result in significant underloading of the server (i.e. potential false positive data). |
| eValid Rich Internet Application Monitoring | Fully realistic simulated user session. | Navigation to a page involves base page plus
all component pages.
Download of component pages is via parallel threads spawned internally by the browser. | 100% | Because the session playback is with a stateful brower, there is NO question about realism: eValid is totally realistic. The server cannot tell that the pages delivered to the client are requested by eValid rather than a human user. |
