Figure :1 Client-server paradigm
Client-server computation - is a distributed model used in the Web, which are two types of entities, implemented applications: customers (clients) and servers (servers). The term "server" should not be confused with a computer that is running the server process.
Protocol, on which clients and servers communicate - is a protocol such as "challenge-response" (request-reply), for which clients send requests to a server responds to them (Fig.). Clients and servers can operate on the basis of the TCP protocol or model, without a connection, for example, UDP.
Figure :2 Request-response in the client-server systems
If at the same time to serve only one request, the server resources are not fully utilized, its carrying capacity (number of requests serviced per unit of time) will be low and response time to requests from customers to grow with the increasing load on the server (see Fig. A) . Therefore, most of the servers provide multiple processes or flows, performed as one process, in order to serve the queue of incoming requests (see Fig. B).
Figure: 3 a) server with a single process b) server with multiple processes
Length of service in networks
Message from client to server must go through several levels of protocol and can be transmitted through one or more networks (see fig.).
Figure: 4 Connections between the client and server
Time communication services network - this is the time required to transfer the message through the network. That is equivalent to being against the number of bytes needed to send the message, including header and tail (service information), the bandwidth.
Conclusion
Modern computer systems are becoming more and more complex and dependent on network technologies that are accessible through the Internet. Accommodation applications that depend on Web-sites, intranets, and client-server technologies pose a problem to ensure their proper operation with acceptable performance. Creating a scalable Web-services leads to the appearance of certain difficulties because of the large gap between average and peak loads. A detail model of client-server and address the impact of architecture on performance. Web-request is used to carry out its various resources, including processors and disks client, server, networks and routers. Length of service includes time, which holds the request and receive services from a resource (for example, performing the operation I / O on the server database). Time in the queue - this time the request spends in waiting until the resource is free (for example, waiting for the release of the processor).
References
F. Doughs and M. F. Kaashoek, "Scalable Internet Services," IEEE Internet Computing,July/Aug., 2001.
A.Ginige and S.Murugesan,"Web Engineering: An Introduction," IEEE Multimedia,vol. 8, no. 1, Jan.-March, 2001.
J. М. Andrade, М. Т. Carges, Т. J. Dwyer, and S. D. Felts, The Tuxedo System, Addison Wesley, Reading, Massachusetts, 1996.
BBN, "A History of the ARPANET: The First Decade," Tech. Rep., Bolt, Beranek, and Newman, Massachusetts, 1981.
G. Bell and J. Gray, "The Revoluti on Yet to Happen," Beyond Calculation: The Next Fifty Years of Computing, P. J. Denning and R. M. Metcalfe, eds., Copernicus Springer-Verlag, New York, 1997.
Дэниел А. Монаске, Виргилио А. Ф. Алмейда, Производительность Web-служб.Анализ, оценка и планирование,СПб:ООО «ДиаСофтЮП», 2003