|
Conferences and Events | Online Resources | Programs | Security | Services | Shared Network | Technical Support | Training Home » Technical Support » Networks and Workstations » Network Design and Planning » Web Servers |
|||||||||||||
Web ServersScopeThis document is intended to provide general information on Web server software and will answer questions like "what is a Web server?," "how does a Web server communicate?" and "how can you determine which Web server is right for you?" IntroductionIn the last few years, the World Wide Web (WWW) has become a popular and efficient way to provide information and services to people and areas that might once have been unreachable. The uses and benefits of the Web are virtually limitless. Regardless of physical location, individuals, educational institutions, communities, businesses and organizations have access to an almost limitless amount of information and services. Through the Web, people can access class assignments and materials, enroll in specific courses, make reservations for transportation or accommodations, track merchandise delivery or check account information at financial institutions. The possibilities increase endlessly. In this document we look at the mechanism that makes providing information and services on the Web possible — the Web server. What is a "Web Server?"Almost any computer regardless of its hardware configuration/size (desktop, midsize, mainframe), platform (Intel, RISC) or operating system (Windows Macintosh, NetWare, UNIX), can function as a Web server. The hardware configuration, platform and operating system you choose for your Web server should be determined by the content and intended audience of your website. Those issues will be discussed later in this document. What is a Web server? Essentially, a Web server is a computer that has specific Web server software installed and is connected to a network using physical cables, telephone lines, wireless communication devices, etc. To be more specific, the network must use the Transmission Control Protocol/Internet Protocol (TCP/IP) to communicate, and the Web server software must communicate using the Hypertext Transfer Protocol (HTTP). To understand how a network uses TCP/IP to communicate and how Web server software "speaks" using HTTP, we will examine two models used in computing — the client/server computing model and the Open Systems Interconnection (OSI) model. The client/server model illustrates the different functions of the sending and receiving systems. It defines the process used to pass information from one system to another. The OSI model illustrates how information is transported over the network. It breaks the network up into seven distinct layers and assigns each layer a specific task. These tasks are controlled by protocols. Protocols are the guidelines or rules used to manage the communication process. They ensure that the information is communicated in a manner that will be understood by both the sending and the receiving systems. The Client/Server Computing ModelThe client/server computing model has two basic components, a client and a server. The term "client" refers to the program that interfaces with the user and sends a request for information to a server. The term "server" can refer to a specific software package whose purpose is to serve information to clients — file server, print server, Web server, etc. — or it can refer to the hardware on which the software package is running. (For the purpose of this document, the term "server" will refer to software.) The server is responsible for interpreting the clients' requests and sending back or "serving" the requested information to the client. The client/server model relies on the OSI model to be able to send information between systems that are not alike.
The Open Systems Interconnection ModelThe OSI Reference Model is a layered model that defines how information is passed from one system to another. Each layer of the OSI model performs unique functions or services and is associated with specific protocols. Protocols have defined responsibilities to help with the delivery of information. During the delivery process, control of the information is passed from one layer to the next. The process starts at the top layer (application) of the sending system and is passed down through each layer until it reaches the bottom layer (physical). It then crosses the network, is intercepted by the destination or receiving system and travels back up through the layers.
OSI LayersThe application layer is an interface which allows applications to connect to the network. The name is misleading since the actual application processes or programs which interface with the user do not reside in this layer. It is this layer that allows the HTTP protocol to send and receive information. The presentation layer is responsible for formatting information so the receiving system can understand it. It establishes a common format or "language" between the sending and receiving systems. The session layer establishes the connection between the sending and receiving systems over the network. It provides for the synchronization and management of information exchange between the systems. The transport layer is responsible for checking the size of the information "package," detecting any errors and ensuring the package is delivered to the appropriate upper level application. If an information "package" is too large to send across the network, the transport layer on the sending system divides the information into smaller pieces before passing it down to the next layer. On the receiving system, the transport layer puts the pieces back together and then passes it up to the next layer. On systems that receive information from multiple protocols at the same time, the transport layer ensures that the information is delivered to the correct upper layer process (that is, if a request from a Web client is received — Web clients communicate using HTTP — the request is delivered to the application associated with HTTP or the Web server). On the sending system, the transport layer is also responsible for making the destination system aware that information is going to be delivered, it checks to make sure the information was received, and it ensures that the speed of delivery is appropriate and if errors occur that the appropriate portions are retransmitted. The network layer is responsible for assigning addresses and routing information between networks/systems. At this layer, the exchange of information is connectionless meaning that there is no continuous or dedicated dialog or exchange of information between the sending and receiving systems. In other words, if the information was broken down into smaller pieces by the transport layer, the network layer may receive the pieces out of order or even mixed in with pieces of information from other requests. The data link layer is responsible for physically passing information. Its function is to provide the upper layers access to the physical media (network). It can detect physical errors, send notification of errors, and establish and terminate logical links between systems. The data link layer actually transmits packets from one network interface card to another, based on the physical address — Media Access Control (MAC) address — of the interface cards. (MAC addresses are unique numbers assigned to your computer's networking hardware.) The physical layer is responsible for putting information onto the network and taking information off. It defines the physical network media the system is connected to (ethernet, coaxial cable, fiber, etc.). This layer encodes and decodes the information into digital bits (1s and 0s) between the sending and receiving network interfaces. Now that we've seen how information is passed from one system to another using the client/server and OSI models, let's take that information and relate it to the previous observation — Web servers must be connected to TCP/IP-based networks and they must communicate using HTTP. Transmission Control Protocol/Internet Protocol (TCP/IP)The two layers of the OSI model that TCP/IP encompasses are the transport layer (TCP) and the network layer (IP). Transmission Control Protocol (TCP)Transmission Control Protocol (TCP) is the transport layer protocol responsible for breaking information into smaller pieces, detecting errors and ensuring that the information is delivered to the correct upper level applications. TCP was originally developed to provide a standardized open and independent form of connectivity for communication between networks. To ensure correct delivery, each upper level application that intends to receive or send information is identified with a port number — a unique number associated with that particular application. (Note: standard or default port numbers have been defined for the various applications. For example, the default port number for Web servers is 80). Any incoming information sent to that port on that system is delivered to the corresponding application. Internet Protocol (IP)IP (Internet Protocol) is the network layer protocol which allows systems to communicate with each other by associating logical addresses with each system. These logical addresses are used for the delivery of information. To be able to send and receive information using IP, each system on a TCP/IP network must be assigned a unique 32-bit IP address. This address contains both a network portion and a host portion which can be written in binary format (for example, 11111111 11111111 11111111 11111111), or dotted-decimal format (for example, 255.255.255.255). IP addresses contain the information used to route information between systems. Note: An in-depth look at IP addressing is beyond the scope of this paper. You can get additional information on IP addressing from the following MOREnet documents. IP Addressing Domain Name Services and Naming Guide So what does TCP/IP have to do with Web servers? Web servers must be connected to TCP/IP-based networks because they use the components of TCP/IP to request and send information. IP is used to assign specific addresses to systems and to route requests and responses to those addresses. TCP is used to break the client requests or server responses into pieces that can be sent across the network. TCP checks the information for errors and makes sure that all the pieces have been received and put back in the proper order. It also makes sure that the information is communicated or delivered to the appropriate upper level application protocol. The application level protocol associated with Web servers is HTTP. How do Web Servers Communicate?Web servers communicate using the Hypertext Transfer Protocol (HTTP). HTTP is an application layer (top layer) protocol which has been in use by the Web since 1990. HTTP defines how files (text, graphic, audio, video, multimedia, etc.) are exchanged and what actions Web servers and Web clients take in response to various requests or commands. Most clients on the Web are called browsers — two well known browsers are Netscape Communicator and Microsoft Internet Explorer. Web servers and Web clients use HTTP to send and receive information. For the "communication" process to be successful on the Web, an application must specify which transport layer protocol to use (TCP), which network layer protocol to use for identifying the destination system (IP), and which port the destination application is using (port 80).
Basically, HTTP consists of a "send me this file" request from the Web client, and a "here it is" reply from the Web server. The Web client's job is to make a connection to the Web server, retrieve the requested information, interpret the content and present the information to the user. To accomplish this, the client or browser builds an HTTP request and sends it to the Web server's address. Note: a Web server address is generally referred to by a Uniform Resource Locator (URL), for example, http://www.whatever.com. The process of associating the URL with the 32-bit IP address is beyond the scope of this paper. Each request made to a Web server initiates a connection — that is, a request for information is made to the Web server, a connection is opened, the information is transferred and the connection is closed. Currently, most Web clients and servers support HTTP 1.1. which allows for persistent connections. This means that once a client connects to a Web server, it can receive multiple responses through the same connection. The Web server's job is to listen for incoming requests, interpret those requests and then return the requested information. A vast majority of the information available from Web servers is in the form of Hyper Text Markup Language (HTML) pages. Note: HTML is a standard which covers how webpages are formatted and displayed. Webpages generally contain text, graphics and links to other information. Although much of the information provided on the Web is static, today's technology makes it possible to provide sound, video and dynamic webpages (pages that are generated "on the fly"). A number of server-side technologies (Server Side Includes — SSIs) can be used to create and manage dynamic pages. SSIs are processes that are executed or run using the resources of the server (CPU, memory, disk space, etc.). Some of the more common examples of these technologies include Common Gateway Interface (CGI) scripts, Java (a programming language) and Active Server Pages (ASPs). SSIs can provide a website with more functionality and interactivity. However, implementing these new technologies can also have drawbacks. One drawback is that some of the new technologies can have a significant impact on the performance of your Web server. Server response time can be diminished because of the additional load placed on the processor, for example, database applications which rely on server-side includes can generate a high number of disk accesses which will take up the available resources and as a result, slow response time. In addition to performance issues, SSIs can also be considered an added security risk because essentially users are being allowed to execute commands on the server. There are many things to consider when you begin the Web server selection process. Being able to weigh the advantages and disadvantages of available technologies, and being able to determine which ones are right for you are important aspects of the selection process. How Can You Determine Which Web Server is Right for You?Almost any can function as a Web server regardless of its hardware configuration/size, platform or operating system. Some of the things that will be factors in your choice are content, audience, administrative overhead and hardware costs. Web server software can provide many services. Some questions you should answer are: What information do we want to make available? How much information (five pages, 10,000 pages) will we make available? Who should have access to this information? What technologies do we want to include? Some Web server software/hardware combinations are more appropriate for large volume sites and some are more appropriate for low volume sites. While the vast majority of Web server software is available for free, there are associated costs in hardware, programming, administration, etc. If you plan on making a large amount of information available, you will want to make sure you have plenty of drive space for storage. If you intend to use new technologies that are resource intensive, you will want to make sure you have plenty of memory and a fast processor. You will also need to pay close attention to the features each server software package supports. Not all software packages are capable of handling all types of SSIs. Security issues must be addressed: How secure does your information need to be? How secure does your hardware need to be? Will your Web server provide information to the Intranet (the clients on your Local Area Network), or will you provide information to the Internet (the global network connecting millions of computers.) Make sure you are aware of the security risks associated with each software/hardware configuration. Consider technical expertise issues: Are you going to manage/administer the server or will you contract for support? Will your server be providing more than just Web services to clients on the network (file and print services, mail services, etc.)? In some instances, it might not be financially or physically possible to install a dedicated Web server. If so, take a close look at the resources each service will utilize and plan accordingly. Do you need to allow for growth? Some server packages scale better than others. Web server software does not need to be installed on the largest or fastest machine. Each software package will have its own hardware requirements for optimal performance. Pick a software package that best fits your environment. Just because it's right for someone else, doesn't necessarily mean it's right for you. In brief, before you make a decision, plan your site and what information you want to provide, pick the method you want to use to make the information available, decide what kind of financial investment you want to make and then find the software/hardware configuration that fits your requirements. Bibliography/ResourcesWorld Wide Webhttp://serverwatch.internet.com/Webservers.html
(11/98) BooksAblan, Jerry and Yanoff, Scott. Web Site Administrator's Survival
Guide. Indianapolis, IN: |
|||||||||||||
|
Copyright © 1999-2003 MOREnet. All rights reserved.
Reviewed January 24, 2003.
Contact techsupp@more.net. DMCA and other copyright information. Site Information: Copyright, accessibility, privacy and other information about this site. PageMinder: Receive an e-mail notice when this page updates.
|