Log analyzers Comparisons
Comparison between AWStats and other famous statistics tools
* This number is not really the number of browsers detected. All browsers (known and unknown) can be detected by products that support user agent listing (AWStats,Analog,Webalizer,Sawmill). The 'browser detection feature' and number is the number of known browsers for which different versions/ids of same browser are grouped by default in one browser name.
** AWStats can detect robots visits: All robots among the most common are detected, list is in robotslist.txt (250Kb). Products that are not able to do this give you false information, above all if your site has few visitors. For example, if you're site was submitted to all famous search engines, robots can make 500 visits a month, to find updates or to see if your site is still online. So, if you have only 2000 visits a month, products with no robot detection capabilities will report 2500 visits (A 25% error !). AWStats will report 500 visits from robots and 2000 visits from human visitors.Sawmill Analytics uses a "currently active" list of robots based on the robotstxt.org database.
*** AWStats has url syntax rules for the most popular search engines (that's the 'number detected'). Those rules are updated with AWStats updates. But AWStats has also an algorithm to detect keywords of unknown search engines with unknown url syntax rules. Sawmill uses unique syntax to detect 67 search engines, and you can add any number of custom SE's.
**** Most log analyzers have poor (or not at all) robots, search engines, os or browsers detection capabilities and less features (no or poor visits count, no filter rules, etc...).
It is not possible to add all AWStats features to other log analyzers, so don't forget that benchmarks results are for 'different features'. For this benchmark, I did just complete Webalizer and Analog robots or search engines databases with part of AWStats database. So Webalizer config file was completed with this file, Analog config file was completed with this file. Note that without this very light add (using default conf file), Webalizer speed is 3 times faster, Analog is 15% faster).
Benchmark was made on a combined (XLF/CLF) log record on an Athlon 1GHz.
You must keep in mind that all this times are without reverse DNS lookup. DNS lookup speed depends on your system, network and Internet but not on the log analyzer you use. For this reason, DNS lookup is disabled in all log analyzer benchmarks. Don't forget that DNS lookup is 95% (even with a lookup cache) of the time used by a log analyzer, so if your host is not already resolved in log file and DNS lookup is enable, the total time of the process will be nearly the same whatever is the speed of the log analyzer.
***** Some visitors use a lot of proxy servers to surf (ie: AOL users), this means it's possible that several hosts (with several IP addresses) are used to reach your site for only one visitor (ie: one proxy server download the page and 2 other servers download all images). Because of this, if stats of unique visitors are made on "Hits", 3 users are reported but it's wrong. So AWStats, considers only HTML pages to count unique visitors. This decrease the error (not totally, because it's always possible that a proxy server download one HTML frame and another one download another frame). Sawmill Analytics allows you to choose what you define as a visitor - by default the client IP is used, but you can use a cookie (persistant or session) or any custom string, or combination of string from teh log data.
(a) Data were provided by Sawmill company (Graham Smith).
(b) With such log format, there is no user agent information in log file, so some reports are broken. For example, it's not possible to make reports on browser or os for (information is not stored in log file). To solve this, use another log format (like the combined format).
Article written by Laurent Destailleur.