An Open-Source Approach To Web-Server Analysis
The Webalizer is an open-source external log file-analysis product that is available for Linux and Windows platforms. It is capable of reading a standard Web-server log file, such as those generated from Apache and IIS. The Webalizer generates several HTML pages that are accessible through a browser and are usually set up as a group of secured pages within a site. It is fully featured and provides a good overview of the log file data, as well as several ways to drill down into the specifics of the data.
The Webalizer is easy to set up and reaps immediate benefits that can be seen by you and your client. To see an overview of site usage by month, open the generated index.html in a browser. A graph is presented at the top of the page with a numerical representation of the data below. Hits, files, pages and visits are included in this summary, along with a daily average and a monthly total. By clicking on a month, you can view a summary of usage statistics, including daily statistical totals, hourly statistical totals, most common URLs accessed within the site, URLs by kilobytes, most common entry/exit pages and usage by site. Scheduling a job to regenerate the Webalizer HTML pages each day ensures statistics stay current.
It is also possible to tell your clients the page last viewed in a typical visit, average length of a typical user visit and which external users generate the most page views. The statistics that measure the number of kilobytes downloaded can be of particular value for providing an accurate snapshot of bandwidth usage to solution providers that host sites.
A common concern with Web log-analysis packages is the need to keep large log files on hand in order to have a decent historical perspective on usage. The Webalizer resolves that by keeping its own history file, enabling log files to be rotated or deleted as desired without affecting historical averages.
The Webalizer is configured through a file that is read upon start-up. Elements that can seriously throw off statistics can be omitted from the analysis. For example, a client may want to track external users on a portal application that is also accessible to internal users. Virtual hosts are also supported. Another nice feature is the Webalizer's ability to approximate a user visit. A configuration file setting tells it to treat all visits from a unique IP address within a certain time frame as a unique visit.
If you are a VAR or an ISV with a need for a product like this, I encourage you to look at open-source tools in general. They bring to the table low-cost, high-value solutions that can be downloaded, installed and evaluated with a minimum investment of time and talent. Granted, some of the commercial products, such as WebTrends and NetTracker, have more ways of slicing and dicing the data plus 24/7 support, but open-source products like the Webalizer still can satisfy the requirements for the majority of analysis applications.
A quick download of the software from www.webalizer.com will demonstrate the software's capabilities. There is also a documentation package available through a link on the Webalizer home page. In addition, take a look at AWStats from www.sourceforge.net, another solid product with a slightly different user interface for perusing log statistics.
Kevin Carlson ([email protected]) is CTO of Verteris, an Atlanta-based ISV.