USO0RE41440E
(19) United States (12) Reissued Patent
(10) Patent Number: US RE41,440 E (45) Date of Reissued Patent: Jul. 13, 2010
Briscoe et a]. (54)
GATHERING ENRICHED WEB SERVER
5,848,396 A
* 12/1998
ACTIVITY DATA OF CACHED WEB CONTENT
5,892,917 A
*
4/1999
5,913,041 A
*
6/1999 Ramanathan et al. .
709/233
8/1999 Logue et a1.
709/219
5,935,207 A *
(75) Inventors: Paul Roger Briscoe, Southampton (GB); Cameron Donald Ferstat, St. Leonards
(AU); Matthew Robert Ganis, Carmel, NY (US); Stephen Carl Hammer, Marietta, GA (US); Gary Bob Kip Hansen, Saugerties, NY (US); Sean Alan Harp, Atlanta, GA (US); Michael Shannon Nichols, Georgetown, TX (US); Herbert Daniel Pearthree, Cary, NC (US); Paul Reed, Brookline, MA (US); Brian James Snitzer, Lancaster, PA (US) (73) Assignee: International Business Machines
Corporation, Armonk, NY (US)
Gerace ...................... .. 705/10 Myerson ............ .. 709/224
5,991,735 A
* 11/1999
Gerace ........ ..
6,018,619 A
*
Allard et a1. .... ..
1/2000
6,018,763 A *
1/2000 Hughes et al.
709/213
6,023,726 A
*
2/2000
709/219
6,041,355 A
*
3/2000
Toga .............. ..
6,085,229 A
*
7/2000
Newman et al.
..
..... .. 709/203
6,094,662 A
*
7/2000
Hawes ............ ..
707/104.1
6,363,418 B1 *
3/2002
Conboy et a1. ............ .. 709/218
Saksena ....... ..
OTHER PUBLICATIONS
Computer Knowledge NewsletteriNov. 1999 Issue.* PCT/EP01/09308, PCT Preliminary Examination Report, Jul. 3, 2003, European Patent Of?ce.* Primary ExamineriLashonda T Jacobs
(22) Filed:
(57)
ABSTRACT
A method and system for gathering enriched web server activity data in a global communications network in which requested information ?les are cached at a plurality of net work devices. With the prevalence of web caching on the Internet, the origin web servers do not serve the majority of
Related US. Patent Documents
Reissue of:
(64) Patent No.: Issued:
709/227
(Continued)
(21) App1.No.: 12/437,5s1 May 8, 2009
705/10
709/224
7,216,149 May 8, 2007
Appl. No.:
09/641,495
Filed:
Aug. 18, 2000
requests for web site content. A single pixel clear Graphics Image Format (GIF) request is added to the HyperText
(2006.01)
Markup Language (HTML) source ?le for a web page. Appended to the GIF request is a Common Gateway Inter
(52)
US. Cl. ...................... .. 709/217; 709/203; 709/219;
face (CGI) string of data that contains enhanced web activity data information, including the number of images (“hits”)
(58)
Field of Classi?cation Search ................ .. 709/203,
that have to be retrieved by a client browser to build the web page, and the referring identi?er that resulted in access to the
709/217, 218, 219, 223, 224, 225, 226, 227, 709/228, 229; 711/138
web page. The single pixel clear GIF request is not cache able and results in the request being transmitted to the origin
(51)
Int. Cl. G06F 15/16
711/138
See application ?le for complete search history. (56)
References Cited
server to accumulate an accurate number of hits on the web
U.S. PATENT DOCUMENTS 5,796,952 A
*
8/1998
web server when the client browser interprets the HTML ?le. The enriched data is stored in log ?les at the origin web
Davis et a1. ............... .. 709/224
cams ENGINE
page.
59 Claims, 15 Drawing Sheets
US RE41,440 E Page 2
US. PATENT DOCUMENTS 6,385,642 6,393,479 6,606,581 6,742,040
B1 B1 B1 B1
7,003,565 B2
5/2002
Chlan et a1. ............... .. 709/203
5/2002
Glommen et a1. ......... .. 709/224
8/2003 Nickerson et a1. . 5/2004
2002/0004733 A1 2002/0147772 A1 2008/0052392 A1
702/186
Toga ........................ .. 709/229
* cited by examiner
2/2006 Hind et a1. ................ .. 709/224 1/2002 Addante ...................... .. 705/7 10/2002 Glommen 2/2008 Webster et a1. ............ .. 709/224
US. Patent
Jul. 13, 2010
Sheet 1 or 15
US RE41,440 E
US. Patent
Jul. 13, 2010
CLIENT
Sheet 3 or 15
US RE41,440 E
302
BROWSER REQUESTS HTML WEB PAGE
/ 310 304 PAGE
Y
CACHED AT
DELIVER I-ITML FILE TO BROWSER TO
INTERPRET AND w
WENT?
auILD PAGE WITH souRcE OR CACHED IMAGES
/
316 326
RETRIEVE
cAcI-IED IMAGES AND coMPLErE BUILD OF PAGE
PROCESSING
312 320
306
DELIVER HTML
313
FILE TO CLIENT
PAGE
Y
CACHED AT ISP'?
/ TRANSMIT
l?g'ggg? :30
HTML FILE
REQUEST FOR
BU'LD PAGE
CONTAIN C.GIF CALL
uc.GII= TO ORIGIN WEB SERVER
wITI-I souRcE OR cAcI-IED IMAGES
314
‘
aoa
TRANSMIT REQUEST TO
oRIGIN WEB SERVER
DEL'VER HTML
ENRCISQJEETNEB
FILE TO CLIENT BROWSER TO
SERVER ACTIVITY DATA &
INTERPRET AND __
BUILD PAGE WITH souRcE 0R cAcI-IED IMAGES
STORE ‘N LOG
F'LES
“1 END REQUEST PROCESSING
FIG. 3
322
324
US. Patent
Jul. 13, 2010
Sheet 4 or 15
US RE41,440 E
Site Level Analysis Week Ending
1,979 461 1
Week Ending 446,484,591 23 10 1 29
54.3
Week Ending
Week Ending
513 1,1 29 795, 20.760 4
30 1,781 2,364 1.357
86.
70.83
.
274 102.1
775 103.3
7.21
411.
11.1
7.
922.1
525
Download comma delimited format; Sitg Level Analygis
0 OverallU 0Custom MonthDate Granularity Range[ 6) 04/27/2000 Week Granularity MFMIZT/ZOOO 0 Day Granularity
F
l I Save Repor1
Run ReponOf?ine
FIG. 4
I
US. Patent
Jul. 13, 2010
US RE41,440 E
Sheet 5 0f 15
Referral Categories Referral Catego _
All "
All
Week Ending Week Ending Week Ending Week Ending Week Ending 05/27/2000 Traf?c Information 1 473 72 5 131
2,195,
6
95,341 Total Page
1,979
803,753
1 ,234,73
23,110,!
29,
2,364
Download comma delimited format: gage Vigws
m
Area of Ana_iysis Measurements O Tra?ic Type 0 Visitor Entry Page Tmmc VB" 0 Traffic Category 0 Visitor Exit Page E Page View‘ U visits 0 Visitor SubDomain 0 Visitor Browser D Seconds/Page view D Page VIEWS/W5"! 0 Visitor Domain
0 Visitor Piatfonn
@ Visitor Referral
0 Usage Group 0 Ad Analysis
D snows/Vii"
D Display Percentage of Total
[1 custom Date Range Louzmooo lv?'ommooo lvl 0 Overall 0 Month Granularity @ Week Granularity 0 Day Granularity @ Show data by categories 0 Show data without categories
E FIG. 5
US. Patent
Jul. 13, 2010
Sheet 6 or 15
US RE41,440 E
Referral Category: Search Engines and Directories - Display: Referral SubCategories
i
Re
‘i Search Engines and Direetorles
Citjcgory Week Ending Week Ending Week Ending Week Ending Week Ending All "'
Traffic Information 1 l
Engines Directories otal
Views
FIG. 6
US. Patent
Jul. 13, 2010
Sheet 7 or 15
US RE41,440 E
Referral Category: AltaVista - Display: Referrals ''_1
{
Referral Category
iSearch Engines and Directories AllaVista
Referral
mg
m
05/27/2000 06/03/2000 ;
I
Traf?c Information
E1 www.altavista.com/cgi-bin/query [:l vewwelttavista.fr/cgl-bm/qucry D
698E] 857
65,93 6|} 6,437‘?
15 I
883]!
76 L
424pl
82!
189“
J
105]
- 462]:
J
69 37
48 509]
40]
482
a 38‘
223} 151]
10]
9%
IE] www.altavistalcgi~binlquery
0]
2941
I[:laltavista.de/cgi-bin/query [jdimltavista.com/Top/Sports
][
l 1
218 100
i E] www.altavista.se/cgi-bin/query
]
22]
5s
.lUwwwvaltavista.ca/cgi-bin/query '|l:lwww.altavista.com/
I
a 3]
16] E 38] g
i U d1r.altav1sta.com/Top/Spo?s/Tenms/Toumaments
IL ]
T
[J altavista.fr/cgi-bin/query
dir.altavista.com/Top/SportsfI‘ennis/Toumaments/Grand_Slamsl [J www.altavista.nl/cgi-bin/query D www.altavista.dc/cgi-bin/query i'_ [J [:1 www.altavista.co.uk/cgi-bin/query ragingsearch.altavistaeom/cgi-bin/query
[J altzwistaadvalvas.be/scripts/avsearchdll
[:l wwwaltavista.it/cgi-bin/query i [j dir.altavista.com/Top/Sports/Tennis ‘_ [I] altavista.advalvas.be/scripts/AVsearch.dll
_
2674‘
1,279]?
U www.altavista‘fr/cgi-bin/ms [:l babel?slLaltavismcom/translatedyn
1
O} 45]
[E] www.altavista.com/cgi—bin/telia {El www‘almvistacanada.c0m/egi-bin/query
_] 7I
J 23“
0! 37'
1W
2] 33]
{Ualtavista.nl/cgi-bin/query lgwwwnltavisIa.c0m//cgi-bin/qucry LOther AltaVista Page Views [AltnVista Total Page Views
290]
FIG. 7
US. Patent
Jul. 13, 2010
Sheet 9 0f 15
US RE41,440 E
Content Category: Home Page - Display: Resources Content Categog Home Page
Week Ending
Resource Trat‘?c Information
1,415
1,187,831 657,21
Ellen
1,012,851
[j/en/indexhtml. III /fr/index.htm1= 1 7 U l?'l index. html= 1 6 , D/fr/index.html=15
1
1 [:l/fr/index.htrnl=l8
1
U/fr/index.htm1m=17 ome Page Total Page iews
227,2
,
Area of Analysis
1,845
2,428,731
J
238,92
Measurements
0 Traffic Type
0 Visitor Entry Page
Tram‘
@ Trat‘?c Category
0 Usage Group
E Page VIEWS
‘
E] Seconds/Pace ‘new U Display Percentage of Total
1 0 Visitor Referral
Display Totals For:
.
.
.
1
_
: 0 Visitor SubDomatn I 0 Visitor Domain I
23,921
a
B Site
[:1 Cusmm Date Range Fora/2712000 H] 04/27/2000 [v]
ic> L 1
l
Overall O Month Granularity (9 Week Granularity 0 Day Granularity
Minimum Page View Threshold: 1 1 [Z1 Enable Single Object Selection
View Topzl 25 IFind 11:]
FIG. 9
l
I 1
1
US. Patent
Jul. 13, 2010
Sheet 10 0f 15
US RE41,440 E
Saved Reports
UserId: fropenOO $2 Level Reports
Hourly Reports Visi Dlstri
i n Re
rts
ljref?g Reports Qgptept Reports sgpDgmeln Reports Referral Repgrts Domgin Reports
E_nt[y Page Reports Exit Page Reports Brpwser Reports Pl_atfgrm Reports U§QQ.Q.__G..IQU [2. Reports Ad Reports
FIG. 10
Site Level Reports Date Ra Most Recent 5 Weeks
Granula
FIG. 11A
Shared es Standard
US. Patent
Jul. 13, 2010
Sheet 11 0115
US RE41,440 E
Visit Distribution Reports Dls'flflbuuon Minutes / Visit Views /
Date Range
ranu
Shared
Recent 5
eekly
g'andard
Recent 5
es
Standa
FIG. 11B Traffic Reports Usage Filte uste
Traffic Measu
V|s|t
shared
Measu
one Hits
eek‘y
Standard
FIG. 11C Content Reports Fi
Traffic Measu
Visit Measu
Date
ranula
shared es
(Standa
FIG. 11D
US. Patent
Jul. 13, 2010
Sheet 12 or 15
US RE41,440 E
SubDomain Reports Traf?c
Visit Measu
Shared es
Standard) as
Standard]
FIG. 11E Domain Reports Traffic
Visit Measu
FIG. 11F
Shared
US. Patent
Jul. 13, 2010
US RE41,440 E
Sheet 13 0f 15
Visit
Shared
0p 10
Op 10
es
5
Standarz
eeks
ES
Recent 5 eeks
Standarc
/27/2
OP
7
FIG. 11G Exit Page Reports If
usage
-
Traf?c
Visit
-
Date !
| '
Name Cluster Fliters Measures Measures V'ew Range Gran" anty
IQQ
Most
1Q
Exit
Shared
All
None
None
. .
Top
Recent
Visits
10
5
E'??é
Yes
Weekly
(standard)
Weeks
FIG. 11H Entry Page Reports Usage
.
Traf?c
Visit
.
Date
.
Name cluster Fllters Measures Measures View Range Granulanty
1011.)
Most
m
Ent _ All
Shared
None
None
. .
Top
Recent
V|s|ts
10
5
.P_age§
Weeks
FIG. 111
Yes
Weekly
(standard)
US. Patent
Jul. 13, 2010
US RE41,440 E
Sheet 14 0f 15
Browser Reports Visit Measu
ula
Shared es
Standard
es
Standard
FIG. 11]
Platform Reports Usage
.
Traffic
Visit
.
Date
.
Name cluster Filters Measures Measures View Range Granularlty
IQp_1Q Platforms b Pa g All
Shared
Mos?“ None
Pages
None
Top 10
Recent 5 Weekly
Yes (standard)
\Liew_ed
Weeks
T92 10
Top
Most Recent Weekly
Yes
1O
5
(Standard)
Platforms All . . bv VlSltS
None
None
Visits
Weeks
FIG. 11K
US. Patent
Jul. 13, 2010
Sheet 15 or 15
US RE41,440 E
Usage Group Reports Usage
.
Traf?c
Visit
.
Name cluster Filters Measures Measures
mg
Qgstg All
Visits “—
None
None
Visits
View
Date
.
Range Granularity M st Rgcent
Categorical 5
Weeks
Weekly
Shared
Yes 5
FIG. 11L
Ad Reports Traf?c
Visit
Measures M
FIG. 11M
‘
( tandard,
Shared
US RE41,44O E 1
2
GATHERING ENRICHED WEB SERVER ACTIVITY DATA OF CACHED WEB CONTENT
to deal with the ?le that it is about to open. One of the most common schemes to access web pages is HypterText Trans
fer Protocol (HTTP). The second part of the URL is the name of a server where the ?le is located followed by the path that leads to the ?le and the ?le name. Sometimes, a URL ends in a trailing forward slash with no ?le name given. In this case, the URL refers to the default ?le in the last
Matter enclosed in heavy brackets [ ] appears in the original patent but forms no part of this reissue speci?ca tion; matter printed in italics indicates the additions made by reissue. This application is a reissue application 0fU.S. Pat. No. 7,216,149, issued May 8, 2007 on US. Ser. No. 09/641,495
directory in the path (i.e., index.html), which generally cor responds to the home page. For example, consider the web address “census.rolandgarros.org/rc/images/ . . .”. The
domain name is “censusrolandgarros.org”. This is the spe ci?c host computer on which corresponding web pages
?ledAug. 18, 2000.
The present invention relates generally to client-server computer systems and, more speci?cally, to information
reside. The next segment of the URL is the directory (“rc” and subdirectory “images”) on the host computer that con tains a speci?c web site. The last segment of the URL, repre sented by the ellipsis mark, is the ?lename of the speci?c
access requests to a web site server over a global communi
web page being requested.
cations network. All web pages are written with HyperText Markup Lan guage (HTML). Hypertext and universality are two essential
shows the entire path to the ?le, including the scheme, server name, the complete path, and the ?le name itself. A relative
BACKGROUND OF THE INVENTION
URLs can be either absolute or relative. An absolute URL 20
features of HTML. Hypertext means that a programmer can create a link on a web page that leads the visitor to any other
web page or to practically anything else on the Internet. Hypertext enables information on the web to be accessed from many different directions. Universality means that
25
because HTML documents are saved as ASCII or text only
?les, virtually any computer can read a web page. HTML
lets the web designer format text, add graphics, sound, and video, and save it all in a text or an American Standard Code
for Information Interchange (ASCII) ?le that any computer
30
can read. The key to HTML is in the tags, which are key
words enclosed between less than (<) and greater than (>) signs, that indicate the type of content coming up next. While practically any computer can display web pages, how those pages actually look depends on the type of computer, the monitor, the speed of the Internet connection, and the
URL describes the location of the desired ?le with reference to the location of the ?le that contains the URL itself. The relative URL for a ?le that is in the same directory as the current ?le is simply the ?le name and extension. To view a single page, the browser running on a client computer, may request and download numerous ?les from a web site server. The number of object access requests
(“hits”) stored in the web site server’s access log will typi cally exceed the number of distinct client sessions in which clients are accessing information on the web site, reducing the accuracy of the access log. Data networking is growing at a phenomenal rate. The number of web users is expected to increase by a factor of ?ve over the next few years. The resulting uncontrolled
35
growth of web access requirements is straining all attempts to meet the bandwidth demand. Additionally, although the
browser software used to view the page.
volume of web trai?c on the Internet is staggering, a large
Advanced web designers often use a scripting language called JavaScript and a system of naming parts of the web
percentage of that trai?c is redundant, i.e., multiple users at
page called the document object model (DOM), together
40
with HTML to create dynamic content on a page. These
any given site request much of the same content. This means that a signi?cant percentage of the wide area network (WAN) infrastructure carries the identical content and iden
effects are sometimes called dynamic HTML, or DHTML. HTML tags are commands written between angle brackets
tical requests for accessing it daily. Web caching performs a
(< >) that indicate how the browser should display the text. Examples of HTML tags are BASE, FORM, FRAME, IMG and SCRIPT. There are opening and closing versions for
requests more quickly, without sending the requests and the
local storage of web content to serve these redundant user 45
Caching is the technique of keeping frequently accessed
many tags and the affected text is contained within the two tags. The opening and closing tags use the same command
word; the closing tag carries an initial forward slash (/) sym bol. Many tags have special attributes that offer a variety of options for the contained text. The attribute is entered between the command word and the ?nal angle bracket. A series of attributes can be used in a single tag just by writing
50
information in a location close to the requester. A web cache stores web pages and content on a storage device that is physically or logically closer to the user. This access to stored web content is closer and faster than a web lookup. By reducing the amount of trai?c on wide area network links
and on already overburdened web servers, caching provides signi?cant bene?ts to Internet Service Providers (ISPs), enterprise networks, and end users. The two key bene?ts of
one after the other, in any order, with a space separating each one. The attributes in turn, often have values. In some cases, a selection of value is made from a small group of choices. Other attributes are more strict about the type of values they
resulting content over the wide area network.
55
web caching are cost savings due to the reduction of WAN
bandwidth and improved productivity for end users resulting from quicker access. ISPs can place cache engines at strate
accept. Examples of attributes are HREF, SRC, ACCESS
gic points on their networks to improve response times and
KEY and VALUE. A web page is nothing more than a text document written with HTML tags. Like any other text document, web pages have a ?le name that identi?es the documents to the web site designer, the web site visitors, and a visitor’s web browser.
lower the bandwidth demand on their backbones. ISPs can 60
overburdened web server. In enterprise networks, the dra matic reduction in bandwidth usage due to web caching
Uniform Resource Locators (URLs) contain information about where a ?le is located and what a browser should do with it. Each ?le on the Internet has a unique URL. The ?rst
part of the URL is called the scheme. It tells the browser how
station cache engines at strategic WAN access points to serve web requests from local storage, rather than from a distant or
65
allows a lower bandwidth WAN link to service the user base. Alternatively, the organization can add users or add more
services that make use of the free bandwidth on the existing WAN link. For the end user, the response of the local web
US RE41,44O E 4
3
FIG. 1 illustrates an implementation of Web cache engines
cache is almost three times faster than the download time for
over a global communications netWork.
the same content over the Wide area network. Therefore,
users see dramatic improvements in response times, and the
FIG. 2 illustrates an exemplary implementation of the
implementation of Web caching is completely transparent to
uncacheable single pixel GIF With CGI query string param
them. Web caching offers other bene?ts including access
eters added to enrich information recorded in Web logs.
FIG. 3 illustrates the processing logic for handling client requests for Web pages utiliZing the single pixel transparent
control, monitoring and operational logging. The cache engine provides netWork administrators With a simple,
GIF in accordance With a preferred embodiment of the
secure method to enforce a siteWide access policy through
present invention.
Uniform Resource Locator (URL) ?ltering. Network admin
FIG. 4 illustrates a site level analysis display that can be
istrators can learn Which URLs receive hits, the number of
generated based on the implementation of the single pixel transparent GIF of the present invention.
hits per second the cache is serving, the percentage of URLs that are served from the cache, along With other related
FIG. 5 illustrates an exemplary display of referral catego ries that can be generated based on the implementation of the
operational statistics. Web caching starts by an end user accessing a Web page over the Internet. While the page is being transmitted to the end user, the caching system saves the page and all of its associated graphics on local storage. The page content is
single pixel transparent GIF of the present invention. FIG. 6 illustrates an exemplary display of referral cat egory for search engines and directories that can be gener
noW cached. Another user, or the original user can then
access the Web page at a later time, but instead of sending the request over the Internet to the Web server, the Web cache
20
system delivers the Web page from local storage. This pro cess speeds doWnload times for the user, and reduces the bandWidth demand on the WAN link. Updating of the cache data can occur in a number of Ways depending upon the
25
ated based on the implementation of the single pixel trans parent GIF of the present invention. FIG. 7 illustrates an exemplary display of the referral results for a speci?c search engine that can be generated based on the implementation of the single pixel transparent GIF of the present invention. FIG. 8 illustrates exemplary content categories for various
design of the Web cache system. Web caching can be a major problem for publishers of
tion of the single pixel transparent GIF of the present inven
Web content. For example, a publisher can gather an inaccu
tion.
Web pages that can be generated based on the implementa
rate number of hits if some of the visitors access Web content
already in a caching server. Furthermore, if a caching server doesn’t update content promptly, it can return expired or
30
tion of the single pixel transparent GIF of the present inven
stale content to users.
SUMMARY OF THE INVENTION
Cache engines are becoming pervasive on the World Wide
35
Web. As a result, the origin Web servers do not serve or see
40
been used to ensure that some record is recorded by the origin server for advertisements for some years. HoWever,
tion pertaining to each HTTP request, including date and 45
50
during each period.
55
Internet broWser applications alloW an individual user to cache Web pages on his local hard disk. A user can con?gure the amount of disk space devoted to caching. The ?rst time a user vieWs a Website, that content is saved as ?les in a subdi rectory on that computer’s hard disk. The next time the user
remaining untouched. The single pixel clear GIF has been
points to this Website, the broWser gets the content from the cache Without accessing the netWork. Certain elements of the page, including buttons, icons and images, appear much more quickly then they did the ?rst time the page Was
in the Web logs for the uncacheable single pixel clear GIF by appending additional information to it as Common GateWay
Interface (CGI) query string parameters. This enables the log record created by the request for the single pixel clear GIF to function as a “surrogate” for the complete set of log records Which Would have been created if the page content had not been cached.
60
the accompanying draWings, Wherein:
opened. To limit bandWidth demand caused by the uncontrolled
groWth of Internet use, softWare developers have developed applications that extend local caching to the netWork level.
DESCRIPTION OF THE DRAWINGS
The invention is better understood by reading the folloW ing detailed description of the invention in conjunction With
time, the originating Internet Protocol (IP) address, the object requested, and the completion status of the request. The logs are analyZed on a periodic basis to determine the traf?c through the server in terms of hits, the number of pages served, and the level of demand for pages of interest
The use of a transparent GIF is a Way to discretely control
used before, but the data has not been enriched such that it can be used as a surrogate for the complete set of log records. The present invention enriches the information recorded
single pixel transparent GIF of the present invention. DETAILED DESCRIPTION OF THE INVENTION
Format) is the most ?exible tool in a Web designer’s toolbox. the layout of text and graphics on the Web page. No matter Where the transparent GIF is placed on the page, it Will remain unseen With all background graphics and ?lls
tation of the single pixel transparent GIF of the present
Web server softWare typically collects and saves informa
this solution only logs information about the request for the single pixel GIF ?le itself.
The single-pixel transparent GIF (Graphic Interchange
tion. FIG. 10 illustrates an exemplary display of the available saved reports that can be generated based on the implemen invention. FIGS. llAillM illustrate various available saved reports that can be generated based on the implementation of the
the majority of the user requests for Web site content. Packet sniffers Will not see the requests either, as they are satis?ed
by cache engines elseWhere on the Internet. The technique of using a single pixel clear GIF (Which is not cacheable) has
FIG. 9 illustrates an exemplary content category for a home page that can be generated based on the implementa
The tWo current types of netWork level caching products are 65
proxy servers and netWork caches. Proxy servers are soft Ware applications that run on general-purpose hardWare and
operating systems. A proxy server is placed on hardWare that
US RE41,44O E 5
6
is physically between a web browser client application and a
tion to router 18, routers 26, 46 are also shown connected to ISP server 30. Routers 18, 26, 46 are frequently referred to as Points-of-Presence (POPs). A POP is the location of an access point to the Internet and has a unique Internet IP
web server. The proxy server acts as a gatekeeper that
receives all the packets destined for the web server and examines each packet to determine whether it can ful?ll the request itself. If the proxy cannot ful?ll the request itself, it
address. A POP usually includes routers, digital/analog call aggregators, servers and frequently frame relay or Asynchro
forwards the request to the web server. Proxy servers can be
used to ?lter requests, e.g., to prevent employees from
nous Transfer Mode (ATM) switches. Shown connected to router 46 is cache engine 48. Connected to router 26 is cache engine 28 and router 24. Router 24 is connected to a corpo rate intranet 22. Because the router redirects packets destined for web servers to the cache engine, the cache engine operates trans parently to clients. Clients do not need to con?gure their browsers to be in proxy server mode. In addition, the opera
accessing speci?c websites. The problem with using proxy servers is that they are not optimiZed for caching and can fail under a heavy network load. Tra?ic is slowed to allow the proxy servers to examine each packet, and the failure of the proxy software or hardware causes all users to lose network
access. Furthermore, proxy servers require con?guration of each end-user’s browser, which is an unacceptable option for ISPs and large enterprises. Because of these shortcomings of
become popular. These caching-focused software applica tions are designed to improve performance by enhancing the
tion of the cache engine is transparent to the network. The router operates entirely in its normal role for non-web traf?c. A web object can contain a Hypertext Transfer Protocol
caching software and eliminating the other slow aspects of
(HTTP) header to instruct a browser in a caching server how
proxy servers, applications that create network caches have
proxy server implementations. Because a proxy server is run 20 to cache the web object. For a static image, such as a com
under a general purpose operating system that involves very high per-process context overhead, they are not easily scale able to large numbers of simultaneous processes. Networking product vendors offer cache engines as a single purpose network appliance that stores and retrieves content using caching and retrieval algorithms. Such cache engines are dedicated solely to content management and delivery. Since only web requests are routed to the cache engine, no other user traf?c is affected by the caching pro cess. For non-web tra?ic, the router functions entirely in its
pany logo, the expiration header can be set to “no expira tion” so that caching servers can keep the image in the cache forever. In order to gather the exact number of hits on a
speci?c page, e.g., an advertisement, a small image object 25
caching server will retrieve the object from the original web server, and the web server can then count the exact number 30
traditional role. The communications between a cache
engine and a router is de?ned by a cache control protocol. Under this protocol, the router directs only web requests to the cache engine rather than to the intended server. With a cache engine, a client requests web content in the usual man
35
40
content, it sends the request to the Internet or Intranet in the usual fashion. The content is returned to and stored at the cache engine. The cache engine returns the content to the
de?ne a mechanism to pass data about the request from the server to the script. Each element on a web page form will have a name and
value associated with it. The name identi?es the data being
45
over a global communications network such as the Internet.
A client computer 12, 14, 16 can request web content via a router 18.
The router 18 intercepts TCP Port 80 web tra?ic and routes it to the local cache engine 20. The client 12, 14, 16 is
independent manner. CGI is simply a standardized way for sending information between the server and the script. The CGI script is a program that communicates with the server in a standard way. Currently, the supported information servers are HTTP servers. Each CGI server implementation must
transaction, and no changes to the client or browser are
client. Upon subsequent requests for the same content, the cache engine ful?lls the requests from local storage. FIG. 1 illustrates an implementation of web cache engines
of requests. The Common Gateway Interface (CGI) is a simple inter face (protocol) for running external programs, software or gateways under an information server in a platform
ner. A router running a cache control protocol intercepts Transmission Control Protocol (TCP) port 80 web tra?ic and routes it to the cache engine. The client is not involved in the
required. If the cache engine does not have the requested
can be added to the page with the object set to expire immediately, so the caching server won’t cache the object. Then, every time a user requests that page, the browser or
50
sent. The value is the data and can either come from the web page designer or from the visitor who types it in a ?eld. When a visitor clicks the submit button, the name-value pair of each form element is sent to the server. CGI scripts gener ally have two functions. The ?rst is to take all the name
value pairs and separate them out into individual intelligible pieces. The second is to actually do something with that data, such as printing it out, multiplying ?elds together, sending an email con?rmation, or storing it on a server. The
not involved in this transaction and no changes to the client computer or browser are required. If the cache engine 20
form has three important parts: the form tag, which includes the URL of the CGI script that will process the form; the
does not have the requested content, it sends the request via
form elements, such as ?elds and menus; and the submit
router 18 to the Internet to access an Internet content server 55 button which sends the data to the CGI script on the server.
Scripts are little programs that add interactivity to a web page. Simple scripts can be written to add an alert box or some text to the web page; more complicated scripts can be
40, 42, 44. The content is returned to, and stored at, the cache engine 20. The cache engine 20 then returns the requested content to the client computer 12, 14, 16 via the router 18. Several cache engines 32, 34, 36 can be placed in a
written that load particular pages according to the visitor’s
cache farm in a hierarchical fashion at an Internet Service 60 browser or that change a frame’s background color depend
Provider (ISP) site 30. Requests from clients 12, 14, 16
ing on the visitor’s mouse clicks. Most scripts are written in
directed through router 18 and ISP server 30, are diverted to
a scripting language called JavaScript that is supported by most browsers, including Netscape Communicator and Microsoft Internet Explorer.
the cache farm 32, 34, 36 to ful?ll the client request from its storage. If the cache engines 32, 34, 36 are unable to ful?ll the request from local storage, a normal web request is made via ISP server 30 over the Internet 50 to an appropriate
JavaScript is an object-oriented language, which means that it works by manipulating objects on a web page, such as
server 40,42,44 for the requested Internet content. In addi
windows, images and documents. JavaScript commands are
65
