1. Introduction to telecommunications
1. Communication Protocols

• syntax
• semantics

• timing.

1. Protocol Stack

• segmentation and re-assembly
• addressing
• connection control
• sequencing
• error recovery
• flow control
• multiplexing
• transmission services

1. Connection Control

 

 
 
 
 
 

There are two fundamental types of protocols: connection-oriented and connection-less. The path that every PDU has to travel in a connection-oriented communication is the same. The connection, which is called a virtual circuit (VC), is maintained for the duration of the communication. It closely resembles a conversation over a telephone line (and hence the name). Protocol handlers can use the connection identifier as an address for routing the data.

Each virtual circuit can be additionally organized to carry a number of virtual channels, so multiple connections can be handled over a single VC.

In the connection-less communication, each PDU is handled independently and may be routed through a unique path. At every node, the message, which is called a datagram, is sent further according to a certain routing algorithm. There are a number of routing strategies involving either static or dynamic (adaptive) algorithms.
 

2. Addressing

 

 
 
 
 
 

There may be many computer nodes involved in a communication process between two parties. It is analogous to a telephone conversation that may be carried out over a number of telephone exchanges. Therefore, each message must indicate its target, so at each node it can be routed toward that target. A part of the control information in the header specifies the destination of the PDU. Very often, the source also is included along with other related information like addressing level (e.g., IP address), scope (e.g., global) and mode (unicast, multicast or broadcast).
 

• unicast – sending messages to one party (one target)
• multicast – sending messages to many parties (multiple targets)
• broadcast – sending messages to anybody, who is listening.

1. Segmentation and re-assembly

 

 
 
 
 
 

Very often, long messages are split into smaller chunks to increase reliability and efficiency, and to accommodate available services. For example, a human translator must focus on translating relatively short phrases. In the extreme, no translation could occur. Imagine translating a book after hearing it just once!
 

2. Sequencing

 

 
 
 
 
 

To maintain a coherent communication the parties have to exchange information in an orderly fashion. If you ask two questions about time and temperature, you want to get the time first and the temperature next. Even more so, if the messages are segmented. A human translator must translate every phrase in order, because otherwise the whole speech would be hard to understand. Therefore, PDUs are assigned sequence numbers that can be used to ensure that the flow of information is chronological.
 

3. Error recovery

 

 
 
 
 
 

A sequence number is one of several types of control information that might be used to detect and recover from errors. A simplest example might be a request to re-send the message that has not managed to arrive before another message with a higher sequence number. There are many elaborated schemes involving message buffering that built upon this simple strategy. For example, a number of messages can be buffered for a period of time to give each message a window of opportunity to arrive on time. Only if a message is still missing after the maximum capacity of the buffer has been reached or the time allocated has been exceeded will a request for a re-transmission be issued.

It is hard to imagine a human translator buffering many phrases in an online translating process. Nevertheless, an analogous scenario can arise if the phrases are delivered in writing on numbered sheets. If one of the sheets were missing, then the source would be asked to make another copy of the missing phrase.
 

4. Flow control

 

 
 
 
 
 

If the consumption of the messages at the target is slower than the arrival rate, then the source might be requested to slow or stop the transmission. Similar scenario may occur between any two communicating nodes.

A corresponding event in the example with translating a speech might be a request from the translator to slow the speech.

The generalization of the flow control issue is congestion control. The problem arises when the capacity to handle messages is lower than the number of arriving messages. There are a number of elaborate schemes for congestion control, which use control information from message headers.
 

5. Multiplexing

1. OSI Model

 

 
 
 
 
 

There are seven layers in the OSI model of a communication stack, which are shortly described in the table. The OSI model strictly forbids the use of services provided by layers other than the next lower layer. It makes the model very open, but also awkward at times.
 

7	Application	Application layer provides for application-oriented protocols that the applications use to process information.
6	Presentation	Presentation layer provides the application layer with data structuring and manipulation services.
5	Session	Session layer provides the higher layers with services for maintaining a communication session between the communicating parties.
4	Transport	Provides services that ensure coherent message transfer between the source and the target.
3	Network	Provides services to route messages between two nodes in a network.
2	Data	Provides services for reliable transmission of data over a physical medium.
1	Physical	Specifies the exchange rules over a physical media like fiber, copper, wireless, satellite, etc.

Not all nodes are required to implement a complete communication stack. For example, a repeater implements only the lowest two layers of the stack because its role is to just forward the signal. A router requires the lowest three layers to unpack the message up to the network header and use the information from it for further routing.

In this course, we are interested in the application layer, but the transport and network layers are relevant as well. We will be using the TCP/IP model, which is the basis of the Internet and many intranets and extranets.
 

2. TCP/IP

 

 
 
 
 
 

The TCP/IP model has become a de facto standard, because it was ready when needed, Department of Defense (DoD) sanctioned its use and the Internet adopted it. The TCP/IP model is managed by the Internet Architecture Board (IBA) through its subsidiary, the Internet Engineering Task Force (IETF). IETF issues Requests for Comments (RFCs) to obtain proposals for standards from the Internet society. RFC goes through a series of stages before it becomes a standard.

In the TCP/IP model, the upper three layers of the OSI model are collapsed into one application layer. Services provided by any lower layer can be used, as is the case with the ICMP protocol. In contrast to the OSI model, the protocols at the same level do not have to provide the same services. Accordingly, protocols cannot be substituted at will. If protocols share the set of support services provided by protocols at a certain layer, then they are considered to belong to the next higher layer.

In this model, when an application (such as an FTP, Telnet, SMTP or HTTP client) wants to exchange data with a remote peer, it usually uses the services of the transport layer. It sends individual datagrams using User Datagram Protocol (UDP) or establishes connection and sends messages over the connection using Transmission Control Protocol (TCP). In any event, all messages are divided into packets that are routed by the handlers of the Internet Protocol (IP). The packets are routed to the destination node, where they are reassembled and passed to the corresponding transport layer. Packets can be further divided into smaller pieces on their way to the target, but it is done transparently to the transport layer and to the application.

There are a number of routing protocols used to move packets towards its target. Border Gateway Protocol (BGP) is widely used to route messages between routers on different networks. Open Shortest Path First (OSPF) is predominant as an interior routing protocol. The protocols are used to exchange routing information between routers.
 

1. Internet Protocol (IP)

 

 
 
 
 
 

Internet Protocol is the glue that keeps the networks connected into one global super network, the Internet. It is a connection-less protocol designed to route data packets between nodes independent of the networks to which they belong. To achieve its objective, IP uses control information encapsulated in the header, which is added to every routed message.

From the perspective of this course, the most important part of the header is the address information. It determines the source and the target of the message. The source is added automatically, so is of lesser interest to us. The target has to be provided by the application that wants to communicate. The IP address can be derived from more abstract naming schema provided by name servers; e.g., Domain Name System (DNS).
 

2. IPv4 Addressing

 

 
 
 
 
 

The current version (4) of IP uses 32-bit addresses. Dotted decimal notation is usually used to annotate addresses. It divides the 32-bit address into four groups of 8 bits, and then translates the values of each group from binary to decimal. For example, all zeros are represented by 0.0.0.0, while all ones will be 255.255.255.255. All IP addresses fall somewhere in between.

The addresses are divided into the following categories that define certain address spaces:
 

 

Class A	Directed to networks with many nodes. Few networks can be accommodated. 126.
Class B	Directed towards medium size networks. 16382.
Class C	Accommodates plenty networks with small number of nodes. 2097150.
Class D	Multicast.
Class E	Reserved for future use.


Network addresses are assigned by Network Information Center (NIC). NIC has to be contacted each time a new network is requested or an address is added or changed. To increase flexibility of the addressing scheme, networks may be divided into subnetworks. Managing addressing within subnetworks does not require NIC’s intervention, so it has become popular with large organizations. The scheme minimizes the sizes of the routing tables by addressing subnets rather than all hosts in the network. A router has to know hosts only in the subnet to which it belongs. Addressing other subnets is easy if the address is logically ANDed with a subnet mask.

There are five classes of IP addresses.



Assigned Classes of Internet Addresses

Note: Two numbers out of each of the class A, class B and class C network numbers, and two host numbers out of every network are pre-assigned: the ``all bits 0'' number and the ``all bits 1'' number. These are discussed below in Special IP Addresses.
 

• Class A addresses use 7 bits for the network number giving 126 possible networks (we shall see below that out of every group of network and host numbers, two have a special meaning). The remaining 24 bits are used for the host number, so each networks can have up to 2(superscript 24)-2 2 to the power 24 minus 2 (16,777,214) hosts.
• Class B addresses use 14 bits for the network number, and 16 bits for the host number giving 16382 networks each with a maximum of 65534 hosts. (The Three Bears Problem)
• Class C addresses use 21 bits for the network number and 8 for the host number giving 2,097,150 networks each with up to 254 hosts.
• Class D addresses are reserved for multicasting, which is used to address groups of hosts in a limited area.
• Class E addresses are reserved for future use.


It is clear that a class A address will only be assigned to networks with a huge number of hosts, and that class C addresses are suitable for networks with a small number of hosts. However, this means that medium-sized networks (those with more than 254 hosts or where there is an expectation that there may be more than 254 hosts in the future) must use Class B addresses. The number of small- to medium-sized networks has been growing very rapidly in the last few years and it was feared that, if this growth had been allowed to continue unabated, all of the available Class B network addresses would have been used by the mid-1990s. This is termed the IP Address Exhaustion problem. The problem and how it is being addressed are discussed in The IP Address Exhaustion Problem below

One point to note about the split of an IP address into two parts is that this split also splits the responsibility for selecting the IP address into two parts. The network number is assigned by the InterNIC, and the host number by the authority which controls the network. As we shall see in the next section, the host number can be further subdivided: this division is controlled by the authority which owns the network, and not by the InterNIC.
 

3. Subnets

 

 
 
 
 
 

Due to the explosive growth of the Internet, the use of assigned IP addresses became too inflexible to allow easy changes to local network configurations. These changes might occur when:
 

• A new physical network is installed at a location.
• Growth of the number of hosts requires splitting the local network into two or more separate networks.


To avoid having to request additional IP network addresses in these cases, the concept of subnets was introduced.

The host number part of the IP address is sub-divided again into a network number and a host number. This second network is termed a subnetwork or subnet. The main network now consists of a number of subnets and the IP address is interpreted as:

<network number><subnet number><host number>

The combination of the subnet number and the host number is often termed the ``local address'' or the ``local part''. ``Subnetting'' is implemented in a way that is transparent to remote networks. A host within a network which has subnets is aware of the subnetting but a host in a different network is not; it still regards the local part of the IP address as a host number.

The division of the local part of the IP address into subnet number and host number parts can be chosen freely by the local administrator; any bits in the local part can be used to form the subnet accomplished. The division is done using a subnet mask which is a 32 bit number. Zero bits in the subnet mask indicate bit positions ascribed to the host number, and ones indicate bit positions ascribed to the subnet number. The bit positions in the subnet mask belonging to the network number are set to ones but are not used. Subnet masks are usually written in dotted decimal form, like IP addresses.

The special treatment of ``all bits zero'' and ``all bits one'' applies to each of the three parts of a subnetted IP address just as it does to both parts of an IP address which has not been subnetted. See Special IP Addresses. For example, a subnetted Class B network, which has a 16-bit local part, could use one of the following schemes:
 

• The first byte is the subnet number, the second the host number. This gives us 254 (256 minus 2 with the values 0 and 255 being reserved) possible subnets, each having up to 254 hosts. The subnet mask is 255.255.255.0.
• The first 12 bits 15 are used for the subnet number and the last four for the host number. This gives us 4094 possible subnets (4096 minus 2) but only 14 hosts per subnet (16 minus 2). The subnet mask is 255.255.255.240. There are many other possibilities.


While the administrator is completely free to assign the subnet part of the local address in any legal fashion, the objective is to assign a number of bits to the subnet number and the remainder to the local address. Therefore, it is normal to use a contiguous block of bits at the beginning of the local address part for the subnet number because this makes the addresses more readable (this is particularly true when the subnet occupies 8 or 16 bits). With this approach, either of the subnet masks above are ``good'' masks, but masks like 255.255.252.252 and 255.255.255.15 are not.

With the growth of the Internet, the name space based on a 32-bit address is running out. Therefore, the IETF issued an RFC for new generation of the IP protocol. The result is known as IP new generation, IP version 6 or IPv6.
 

4. The IP Address Exhaustion Problem

 

 
 
 

The number of networks on the Internet has been approximately doubling annually for a number of years. However, the usage of the Class A, B and C networks differs greatly: nearly all of the new networks assigned in the late 1980s were Class B, and in 1990 it became apparent that if this trend continued, the last Class B network number would be assigned during 1994. On the other hand, Class C networks were hardly being used.

The reason for this trend was that most potential users found a Class B network to be large enough for their anticipated needs, since it accommodates up to 65534 hosts, whereas a class C network, with a maximum of 254 hosts, severely restricts the potential growth of even a small initial network. Furthermore, most of the class B networks being assigned were small ones. There are relatively few networks that would need as many as 65,534 host addresses, but very few for which 254 hosts would be an adequate limit. In summary, although the Class A, Class B and Class C divisions of the IP address are logical and easy to use (because they occur on byte boundaries), with hindsight they are not the most practical because Class C networks are too small to be useful for most organizations while Class B networks are too large to be densely populated by any but the largest organizations.
 

5. Private Internets

 

 

Another approach to conservation of the IP address space is described in RFC 1597 - Address Allocation for Private Internets. Briefly, it relaxes the rule that IP addresses are globally unique by reserving part of the address space for networks which are used exclusively within a single organization and which do not require IP connectivity to the Internet. There are three ranges of addresses which have been reserved by IANA for this purpose:

     10 A single Class A network
     172.16 through 172.31 16 contiguous Class B networks
     192.168.0 through 192.168.255 256 contiguous Class C networks

Any organization may use any addresses in these ranges without reference to any other organization. However, because these addresses are not globally unique, they cannot be referenced by hosts in another organization and they are not defined to any external routers. Routers in networks not using private addresses, particularly those operated by Internet service providers, are expected to quietly discard all routing information regarding these addresses. Routers in an organization using private addresses are expected to limit all references to private addresses to internal links; they should neither advertise routes to private addresses to external routers nor forward IP datagrams containing private addresses to via external routers. Hosts having only a private IP address do not have IP-layer connectivity to the Internet. This may be desirable and may even be a reason for using private addressing. All connectivity to external Internet hosts must be provided with application gateways.
 

6. Network Address Translation

 

 

Network address translation (NAT) may be used to achieve the following:
 

• automatic local area network protection
• a transparent connection of the network (or its part) to the Internet using a single registered IP address


When NAT is employed, the local area network does not use registered IP addresses. Because of this, the internal structure of the network is hidden and not directly accessible from the Internet. A mediator is needed to access the LAN from without. The NAT module takes care of that. Since it remembers all communication initiated from the protected network it only allows the packets which are an answer to the initiated communication to enter the protected network. Other packets are blocked.

The connection of an entire network using a single registered IP address is made possible since the NAT module rewrites the source address in the packets sent from computers in the local area network with the address of the computer WinRoute is running on.

The connection to the Internet is transparent, which means that the computers in the local network use WinRoute as their gateway (router). From the point of view of the local computers it looks as if they were connected to the Internet using registered addresses. Thus, most applications work with the NAT without the need to setup anything on the application's or server's side. This is the main feature which makes NAT to differ significantly from various proxy servers and application-level gateways that will in principle never be able to support some protocols.
 

7. How NAT Works

 

 

The NAT module maintains a table, which records information about each connection. The main pieces of information are: source IP address and port, target IP address and port, IP address and port used to modify packets.

We may demonstrate the way NAT works using the following example:

Let us consider a computer in a protected network. The IP address of the computer is 192.168.1.22. The computer decides to communicate from port 7658 with a WWW server in the Internet, the IP address of which is 194.196.16.43 and its port number is 80. The communication passes through WinRoute, which uses the address 195.75.16.75 on its outer interface.

First, the computer 192.168.1.22 sends a packet from port 7358 to computer 194.196.16.43, port 80. The packet passes through WinRoute, which checks its table to see if it contains an appropriate entry. If so, the existing entry is used, otherwise WinRoute creates a new one. Then it modifies the packet so that it replaces the source address to its own address. It also changes the source port. Thus the source address will be 195.75.16.65, and the port number will for example be 61001. After the changes the packet is sent on. When an answer arrives, it contains 195.75.16.65 as the target address and 61001 as the target port. WinRoute searches its table by the port number 61001 and finds the entry for the connection. According to the entry, it changes the target address and port, back to 192.168.1.22 and 7658, respectively.



Please note:
Port numbers in the packets passing through WinRoute must be modified, since if two or more stations in the protected network start to communicate from the same port number, we need to identify to which of the stations a packet belongs. The NAT module assigns port numbers from the range of 61000 through 61600. A unique port is allocated for each connection.

NAT Critical Points

Applications work with NAT without any problems if the communication is initiated from the protected network. This is the case with most applications. However, there are applications which are not designed correctly and do not comply with the client-server model entirely. Such applications may not work through NAT, or some of their functions may be restricted. The reason is that these applications use more than one connection and the additional connections are initiated by the server (located somewhere in the Internet). Naturally enough, NAT blocks such connections.

Security Policies, Firewalls and Packet Filtering ...
 

8. IPv6

 

 
 
 

Addressing is only one of the modifications to the IP protocol. Many other things have been changed. The address space not only has been extended but the flexibility of addressing has been increased by adding anycasting and scope to multicasting. There is also a capability to add IPv6 addresses dynamically. A new field in the main header assigns a particular packet to a flow, which may be a basis for allocating special resources. The standard includes provisions for increased security through a support for authentication and privacy. Instead of trailing options in an IPv4 packet, IPv6 has a number of optional headers in addition to the main header.
 
 

9. IPv6 Addressing

 

 
 
 

The new standard defines a new addressing space based on a 128-bit address. The addressing schemes are more elaborated than in IPv4 to accommodate increasingly sophisticated uses of the protocol.

It would be awkward to use the doted decimal notation with 128-bit addresses. For example, imagine dealing with addresses like the following:

105.220.136.100.255.255.255.255.0.0.18.128.140.10.255.255

Instead, the colon hexadecimal notation is used, in which groups of 16 bits are put together as a hex number. The above address looks simpler:

69DC:8864:FFFF:FFFF:0:1280:8C0A:FFFF

A sequence of zeros can be substituted by a double colon, so:

FF0C:0:0:0:0:0:0:B1

becomes:

FF0C::B1
 

10. TCP

 

 
 
 

Although TCP is a connection-oriented protocol, it is carried out transparently by a connection-less IP. IP packets are transmitted unreliably; i.e., there is no guarantee that they will be delivered to the destination. It is up to TCP to provide this kind of reliability. TCP provides connection management, error recovery, multiplexing and flow control.

TCP communication is based on connections between sockets. A socket is a number consisting of the IP address and a 16-bit port number (service access point). The IP address refers to the host (node), while the port number indicates the application. It is used for multiplexing, as a socket can accept multiple connections at the same time. All ports with numbers below 1024 are reserved for standard applications. These ports are called well-known ports. For example, FTP uses port 21, Telnet – 23, HTTP – 80, SNMP – 161, Mobile Agents – 434.

To establish a connection, one of the parties, the server, has to use the LISTEN primitive for anybody or for a specific source. The other end, the client, has to invoke CONNECT to request establishing of a connection. The parameters include the IP address, the port, maximum size of transmitted segments (Maximum Transfer Unit, MTU) and optional data like a password. If the listening process is willing to satisfy the request for communication, then it uses the ACCEPT primitive to send a positive acknowledgement to the client.

Any of the ends can use the SEND primitive to send messages. If a segment is too long, then it is divided into smaller parts, sent and then reassembled as the destination host. A sliding window is used to ensure a proper flow of information; i.e., a timer is started for each sent segment. If the segment is not acknowledged by the time the timer expires, then the segment is resent. The maximum number of allowed outstanding acknowledgements is determined by the size of the window. There are number of algorithms that attempt to optimize this basic scheme.

A TCP connection is a byte stream. An application sends messages by writing to the output stream associated with the connection (similarly to writing to files). The size of the transmission segments is determined by TCP. The PUSH flag in the header can be used to prevent TCP from delaying transmission while waiting for more bytes to pack into the current segment. The receiver obtains the data from connection by reading from the input stream. The URGENT flag can be used to signal to the receiver the reception of the message. In such case, the process would be interrupted to give it an opportunity to read the urgent data.
 

11. UDP

 

 
 
 

UDP is a simple, connection-less transport protocol. It is an unreliable protocol, because it does not guarantee delivery of the messages. Any kind of reliability has top be implemented at the application layer. It is targeted at applications that need to exchange small numbers of messages without the overhead of establishing a connection.

For example, UDP is used by Simple Network Management Protocol (SNMP), because in a failing network establishing a connection might not be possible or not reliable.
 

12. DNS

 

 
 
 

Initially, it was relatively easy to manage a limited number of IP addresses that a given host was to communicate with. A simple, local text file was used to translate meaningful mnemonics to dotted decimal notation. This solution is not a feasible solution for the communication involving large numbers of nodes.

At the application layer of the communication protocol, Domain Name System (DNS) is used to provide mapping from text based names that can be better understood by humans, to actual IP addresses that can be used by the IP. A hierarchical, domain-based naming scheme is fundamental for functioning of the network of DNS servers. DNS servers are capable of resolving names sent to them via UDP by requesting applications.

The whole Internet is divided into domains, which in turn are divided into subdomains, and so on. There are plenty of generic and country top-level domains.

A single DNS server would not be practical for the whole Internet, so usually one or two servers cover one of the non-overlapping zones. If a DNS server is contacted with a request to resolve a name in the same zone as the server’s, then the name is in the server’s registry, and a corresponding address is returned. If the name is outside of the zone, a request is sent to the DNS server for the top-level domain of the name to resolve. This request can be forwarded to children zones in the domain, until it reaches the server for the target zone. It is either resolved there, and the address is returned, or the request fails. The returned address can be kept for some time in the local cache.

Certain abbreviations are allowed, but they are not handled by servers. A local name resolver can be programmed to try several name suffixes with the submitted name. For example, we do not need to use a full name if the receiver of our email is in the same domain.
 

3. HTTP

 

 
 
 
 
 

Hypertext Transport Protocol has been designed to transfer contents of a World Wide Web page from the server to the client browser. Other protocols can also be used to transfer files, but they are not suited to deal with file contents. The choice of the protocol is decided by the client through specifying Universal Resource Locator (URL).
 

1. URL

 

 
 
 
 
 

URL specifies the location of the document that the client is attempting to access and the protocol to use. For example:

http:/www.sce.carleton.ca/courses/

refers to the Web server running at www.sce.carleton.ca. The HTTP protocol will be used to access the repository of courses. Servers may listen on various ports for incoming requests, so there is a provision for explicit port specification for non-standard ports. For example, the following URL:

http://www.sce.carleton.ca:8080/dummy/

indicates that the request should be directed to the port 8080 on to the Web server www.sce.carleton.ca. Only if the server is listening on port 8080 will such a request reach it. By default, Web servers listen on port 80.

Another example:

ftp:/ftp.sce.carleton.ca/paper.ps.gz

refers to the ftp server ftp.sce.carleton.ca. The FTP protocol will be used to obtain a copy of the specified document.

Note that

http://www.sce.carleton.ca/paper.ps.gz

is also a valid URL. In this case, the HTTP protocol is used. The advantage of this is that the presentation of the document can be interpreted. For example, a program capable of handling compressed files is invoked. Multipurpose Internet Mail Extensions (MIME) has been invented by IETF to handle such translations.
 

2. MIME

 

 
 
 
 
 

MIME has been originally designed to allow sending binary files as email messages. It has been adopted a standard way of sending files over the Internet. In this scheme, the original file is encoded as a text file and sent with an extra header that describes the type of the file. The receiver needs a definition of handling for a specific type of binaries. In our example, it might be:

MIME-Version: 1.0
Content-Type: application/x-gzipped

and have a decompressing application associated with it (e.g., UNIX gzip or Windows WinZip).
 

3. HTTP

 

 
 
 
 
 

In spite of what the name may suggest, Hypertext Transfer Protocol is used to transfer any type of information. It includes plain text, Hypertext Markup Language (HTML) documents, video, audio and other types that might be defined using MIME.

HTTP is a client/server transaction-oriented, stateless protocol. The exchange of data occurs over a TCP connection. Each transaction is carried out by an independent connection; i.e., the connection is dropped after a single request has been satisfied. The client, a Web browser, sends specially formatted request messages to the Web server. The server responds with response messages that include a response line with a response code.

The request line of a request message includes the URL of the involved resource and the method; i.e., the command to be performed by the Web server on the resource. They are called methods rather than commands, to accommodate object-orientation in interactions between the client and the server. Any name can be used for a method. If the server implements the method, then it will be executed. Otherwise, an error code will be sent back in the status line of the response message. There are a number of standard methods specified in the HTTP RFC. They must be implemented by every HTTP server.

The server may respond with a request for authentication, if the access to the specified resource is limited. In that case, the client is presented a challenge that indicates what authentication scheme is used and what parameters are needed. Usually the client has to re-send a request message with the authentication information that consists of a user ID and a password. If the specified authentication information passes the security checks, then the requested method is executed.

The request message may specify the date and time, which is used to retrieve the specified document. The date and time describe the age of the version of the resource that had been received in the past and cached locally. If the resource on the server is newer than the cached version, then it is sent by the server. Otherwise, the client will use the version from the local cache.

Very often, there are intermediate nodes involved in the HTTP communication. A proxy can be used on the client side of the firewall to the external networks. In this case, the server must authenticate itself to the firewall before a connection between the proxy and the server is allowed. All requests from the client are handled by the proxy, which acts as an intermediary. The content carried over the connection might be controlled and statistics can be collected. A similar security proxy can be installed on the server side of the firewall.

Another intermediary can be a gateway, which can provide transparent services by other servers such as FTP or Gopher. If a request is directed (for example) to FTP (as specified in the URL), then the gateway will contact the FTP server. The retrieved documents are converted to the format acceptable to the HTTP protocol.

The RFCs for HTTP use the term Universal Resource Identifier (URI), but at this moment, URL is used for practical purposes. URI is a generalization of a WWW identifier. URI specifies the resource without the location or the protocol to obtain it. Generally, the resource is of interest, and not how and from it is retrieved. URL is a type of URI with a specific access protocol and an Internet address of the host.
 

4. HTML

Note: Most people use "Composers" to write HTML

<STYLE>

.greengiant { color: green;

font-size: 24pt;
              margin-left: 5px;
            }

.littlered
{ color: red;
              font-size: 12pt;
            }


</STYLE>

<DIV
CLASS = "littlered">
Riding Hood
</DIV>

<STYLE>

A { color: green;

font-size: 24pt;
    text-decoration: none;

}

</STYLE>

<HEAD>
<STYLE>

.hilda { position:      absolute;

top:            100px;
         left:          50px;
         visibility:
        visible;
         z-index:       1;
         color:
        #008000;
         font-family:   times;
         font-size:
        72px;
       }

</STYLE>
</HEAD>

<BODY>
<DIV
CLASS = "hilda">Hello
Hilda</DIV>
</BODY>

<DIV CLASS =
"smallgreen" ID = "smallbig">
Small is
Big
</DIV>

function slideElement(theElement, from, to) {
    if (from
< to) {
      theElement.style.top = (from += 10);

setTimeout('slideElement(' + theElement + ',' + from + ',' + to + ')',
50);
    }
}

function showElement(theElement) {

element.style.visibility = 'visible';
}

function hideElement(theElement)
{
  theElement.style.visibility = 'hidden';
}

<A HREF
= "home.html" 
 onMouseOver = "showObject(document.all.desc)" 
 onMouseOut
= "hideObject(document.all.desc)">Home</A>

var isNS =
(navigator.appName == 'Netscape' && 

parseInt(navigator.appVersion) >= 4);