Electrical computers and digital processing systems: multicomput – Computer-to-computer session/connection establishing – Session/connection parameter setting
Reexamination Certificate
2000-06-27
2004-01-27
Geckil, Mehmet B. (Department: 2142)
Electrical computers and digital processing systems: multicomput
Computer-to-computer session/connection establishing
Session/connection parameter setting
C709S224000
Reexamination Certificate
active
06684252
ABSTRACT:
BACKGROUND
1. Field of the Invention
The present invention relates to network communications. More particularly, the present invention relates to predicting the performance of computer servers.
2. Description of the Related Art
The Internet is a well-known, global network of cooperatively interconnected computer networks. The World Wide Web (“Web”) portion of the Internet is a collection of computer servers that store Web pages, which are accessible by the public. The Web pages consist of text, graphic, audio/visual, and the like (i.e., multimedia).
FIG. 1
shows a simplified diagram of a network. Client computers
115
connect to an Internet Service Provider (“ISP”)
110
, which provides Internet access to users of client computers
115
. ISP
110
connects client computers
115
via a network
105
(i.e., the Internet) to servers
100
. A browser, which runs on each of client computers
115
, requests Web pages from servers
100
. The browser allows the users of client computers
115
to navigate (or “browse”) between Web pages.
Web requests may represent a search for information, such as an e-commerce search for a product. As shown in
FIG. 2
, an e-commerce search interface presents a user with a plurality of options to select as search parameters, such as price range and/or keyword(s). The user selects the search parameters and clicks “search” to effect a search query. Server
100
returns a Web page with a list of matches to the query.
A concern of operators of computer servers
100
is allocating the limited resources of servers
100
. The goal of the operators is to allocate the most resources to the most requested Web pages in order to increase access to the most pertinent information. The success of a Web page depends, then, on the performance of servers
100
, which store the Web pages, because quick access to information is expected on the Web. Thus, it is crucial for the operators to predict with high certainty the performance of servers
100
in order to allocate their limited resources.
It is well-known in the art that a plurality of concurrent requests impact the performance of a server
100
. Thus, the present performance prediction of server
100
is based on the response times of server
100
to concurrent requests by users of client computers
115
.
Further, the types of request impact the performance of server
100
. For example, for one shopping service, it was determined that the type of request impacts the performance of server
100
by at least a magnitude of three. Because concurrent requests and the types of the requests impact the performance of server
100
, the determination of a model of the performance of server
100
must take these two factors into consideration. Otherwise, as determined from tests, the model of the performance of server
100
may miscalculate the actual performance of server
100
by 200%, which does not allow for the proper allocation of limited resources.
In addition, the use of the actual performance of server
100
fails as a basis for allocating limited resources. To begin, it is a measurement of past performance of server
100
, and not of future performance, which may only marginally correspond to past performance. More importantly, measuring the performance of server
100
fails to isolate whether the performance fluctuations of server
100
result from an increase in requests or from the types of requests. Accordingly, there is a need in the art for improved performance prediction techniques for computer servers.
REFERENCES:
patent: 5446680 (1995-08-01), Sekiya et al.
patent: 5812780 (1998-09-01), Chen et al.
patent: 5892937 (1999-04-01), Caccavale
patent: 6292827 (2001-09-01), Raz
patent: 6438592 (2002-08-01), Killian
patent: 6457143 (2002-09-01), Yue
patent: 6480749 (2002-11-01), Lee et al.
Daniel A. Menasce, et al., “Capacity Planning for Web Performance: Metrics, Models, and Methods”, 1998, pp. 1-15.
Subbash Sharma, “Applied Multivariate Techniques”, University of South Carolina 1996, pp. 1-16.
Sam Kachigan, “Multivariable Statistical Analysis: A Conceptual Introduction”, 1991 2nd edition, pp. 236-260.
George E.P. Box, et al. “Statistics for Experimenters: An Introduction to Design, Data Analysis, and Model Building”, 1937, pp. 453-483.
“SYSTAT®8.0—Statistics”, p. 235.
Geckil Mehmet B.
Intel Corporation
Pillsbury & Winthrop LLP
LandOfFree
Method and system for predicting the performance of computer... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and system for predicting the performance of computer..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for predicting the performance of computer... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3221452