Networking NetworkingEverywhere we turn today, from television commercials to the sides of buses, we see web addresses. Anybody who is anybody has a website. It is no surprise that the growth of the Web has exceeded the growth of the Internet as a whole. Today the Web has a doubling period of just under six months.
NYU Web's growth has been similarly strong. New York University put its web server in place in August 1994, and launched NYU Web in September 1995. The number of files transmitted during October 1997 was more than 7 million -- over 500 times what it was for that first month two years prior. More than 200,000 people from over 140 different countries hit NYU's website in October 1997 alone.
With the increased number of people viewing the Web, there comes a greater demand to determine exactly who they are and what they are viewing. Since the inception of www.nyu.edu three years ago, monthly statistics about traffic surrounding the entire server have been available. But, because this includes the entire NYU site, the files have ballooned to over six megabytes and are difficult to load on any browser but Lynx. In addition, charting statistics for individual sites within NYU Web is cumbersome, requiring cutting and pasting into external applications.
The introduction of a new NYU Web service will make all that work unnecessary. An improved program is generating statistical reports about activities on NYU Web. Plus, each webmaster has access to individual statistics in the same format as that of the new system-wide statistics. In addition to information about how many times each file has been transferred, there is new information including the most popular time of the day, transfers by country, full-color graphs and more. All of this is based on the accesses to your site only.
When you go to your site's stats page, you will see a summary graph of statistics for the past year. Because we started beta-testing the software in September 1997, this page starts with September's statistics. As the months go by, it will remember the previous months' numbers and show the differences.
Under the graph are the raw numbers for each category as well as links to previous months' statistics. At the top of that table is a link to the current month's statistics. The program is set to update the full statistics every morning. This way, you do not have to wait until the end of the month to see how your site is doing.
To get more detailed statistics for a month, including the current month, follow one of the links on the y-axis of the table. If you click on the current month, information is shown up to the last complete day.
The main graph is actually showing you the progress of four very different elements: hits, files, sites and kilobytes.
The first table gives you overall results for the month including hits. There is a misconception about what a "hit" is. You cannot simply take the value for the number of hits your site has received and translate that to the number of people. Anytime a person goes to a page, the logs record at least one hit -- the request for the HTML page. If that HTML page also contains two graphics, the logs record two more hits. That one page now turns into three hits! So, think of hits as the number of objects (pages, graphics, files, etc.) requested.
The number of hits (or requests) is equal to the sum of all files sent, "304 responses", and other responses. For the purposes of this statistics program, "files" refers to the number of objects actually returned to the person. A "304 response" is when a server returns the status number 304 to a browser, which tells it that the object or image or file hasn't been modified since the person last requested it. Therefore, the browser doesn't waste time downloading the object again. Essentially, the file is cached on the computer running the browser (or on some other caching mechanism such as a proxy server -- see below). If your site uses a graphic, such as a navigational bar, repeatedly, you will see more occurrences of a 304. The "other responses" category shows how many times the object was not transmitted because of an error of some sort. Perhaps the person requested a page he or she didn't have permission to access.
The first table will also show you the number of unique URLs which have been sent from your site, and the number of unique sites that have visited your website. The "unique URLs" statistic shows how many of the HTML objects you have within your site, including graphics and HTML pages, were viewed at least once by somebody this month. If this number is low in comparison to the number of objects within your site, you can probably figure that either the number of people visiting is low, or those who are visiting are essentially asking for the same things.
The number of "unique sites" in a month shows how many different domain names visited your site at least once in the past month. If somebody visited once a day for the entire month, it would only count them once since they were only unique the first time they visited. However, this value is somewhat inexact for several reasons (see the explanation of proxy servers below).
As you scroll down the page, you will see some of the aforementioned concepts broken down by day. If, in the middle of the month, you changed the design of your site, you could use these breakdowns to help determine its success. You will also see a list of the 30 most commonly accessed URLs within your site. Remember that not only HTML pages have URLs. Every script, graphic and object has its own URL. While this table only shows the top 30, you can still view the hits for every page. At the bottom of the page is a link to "Total Transfers by File" which gives you this information in a predictable format.
Finally, you will also find a breakdown of the type of domains visiting your site. In America, domain names typically end in .com (US Commercial), .edu (US Educational), .org (Non-Profit Organization), or .net (Network). Most primary and secondary educational institutions use a geographical domain name such as schoolname.k12.ny.us (United States). Anything ending in .us is from the United States. Other countries have their own unique two-letter endings, which the chart and graph will reflect as well. For example, if we look at October 1997, we can see that a user from Burkina Faso accessed NYU Web. The exact host name of the person hitting your site is recorded, and the statistical program creates a page of this information. However, due to privacy concerns, we suppress that information.
Many organizations on the Internet, including NYU, use a network version of caching to some extent: a shared computer on the local network acts as an intermediary between end-user machines (running browsers like Netscape) and World Wide Web sites out on the Internet. Such an intermediary is called a proxy server. The first time a person retrieves a specific website via a proxy server, a copy of that page is kept (cached) on the proxy server. Subsequent attempts to access the page by anyone else whose browser is configured to use the proxy server will receive the cached version -- until such time as the proxy server determines that the real page may have changed and the cached copy should no longer be distributed. Many computers at NYU, including all the machines in ACF's public computer labs, are configured to use NYU's proxy server. Such use of a proxy server can result in a dramatic reduction in the time it takes to obtain a web page, since any Internet delays are eliminated.
Most large Internet service providers, such as America Online, use proxy servers. If AOL didn't use such a device, the Internet backbone -- the main "pipes" for transmitting information -- would be even more burdened with traffic, and the speed at which web pages are transmitted would be considerably slower. While this helps to save bandwidth, it prevents an exact statement about the number of visitors to a server.
If you would like to have these statistics generated for your site, send the URL of your site's home page to webmaster@nyu.edu. It does not make a difference if your site is located within another site. For instance, statistics may be generated for www.nyu.edu/gsas as well as www.nyu.edu/gsas/dept/french. The URL of where your statistics will be stored will be provided to you at that time. You will also have the option of password-protecting your statistics directory.
Please note that these statistics are currently only available for sites within the www.nyu.edu domain. They will not work on pages.nyu.edu or on any other servers at NYU. If you are running your own server, you may wish to install the same statistics program on it.
![[ C ]](../icons/CSmSp98.gif){kind=icon}
Posted January 20, 1998. Page last reviewed March 14, 2007.
 |
 
|
|
|