The History Of Hypertext Transfer Protocol

Published Date: 02 Nov 2017

With reference to the OSI model, Domain Name Server (DNS) is a hierarchical naming system which operates at the application layer and enables a user to input a named address (rather than a more complex IP address which is more difficult for people to remember) in to a browser to locate a node. In this scenario, the DNS query originates from the laptop which has an empty DNS resolver cache and host file when it is turned on initially. When the laptop user opens up a web browser and enters the address of, or points to, a website the laptop sends a DNS query to the DNS server which initiates the process of connecting to the webpage.

Web addresses generally take the form of www.companyname.com. The DNS system then resolves this domain name in to an IP address (in the form of a binary number) that enables any network node or computer to recognise it.

The scenario detailed in this report depicts the second routers involvement when the DNS packet query is received. This DNS packet is addressed to the DNS server and is a request to verify the domain name. This process is known as address resolution.

Now that it possesses the IP address, the DNS lookup uses the router table to identify the destination address and the associated traffic at that point. After crossing the 100BaseFX Ethernet, the DNS lookup arrives at the switch with the IP address. After the switch identifies that the DNS server is connected to port 5, it sends the correct port to the DNS server and awaits a response. Throughout this process the DNS lookup changes as MAC addresses alter with each data link layer connection.

The response to the DNS request will be resolved, providing a Fully Qualified Domain Name (FQDN). The DNS in turn will respond with a return packet (to the laptop in this scenario) via the previous route, mapping the host name to a definite location. The laptop can is then in a position to send a GET request (discussed in more detail later) to obtain the webpage from the webserver.

Hypertext Transfer Protocol

Successful resolution of the DNS query and subsequent mapping of the hostname to an actual location means that the laptop in this scenario will be able to contact the web server via the OSI layer 7 end-to-end protocol, HTTP. This protocol is often referred to as ‘stateless’ (Seebach, 2008) meaning that the web server has no storage facility for information pertaining to the clients whom it provides contents to.

Various updates to HTTP have culminated in a total of nine commands which specify the particular action to be performed during the transfer of data over the internet, specifically;

GET – most commonly used, retrieves information from resources specified by the URI

PUT – requests that entities are stored under the supplied URI

POST – requests the acceptance of the entity

DELETE – deletes the identified resource

HEAD – asks for an identical response to a GET request, however without the response body

TRACE – received requests are echoed back to clients to check for any alterations made by intermediate servers

CONNECT –requested connection is converted into a TCP/IP tunnel

OPTIONS –http methods that the server supports for a specified URL are returned

PATCH – applies partial modifications to a resource

The webserver responds to these commands through the use of status codes which indicate the success or failure of any client requests.

Encapsulation

Encapsulation occurs when a packet for transfer is contained within an out IP packet for the purposes of delivery. The Transport, Network and Data link layers of the OSI model all provide services to the DNS and HTTP with the final Physical layer also involved.

Transport (Layer 4)

The transport layer controls data flow between users and incorporates error control. If data is being sent from more than one source then the transport layer will combine the data into a single stream and allows the identification and retrieval of any data which has been lost or become damaged between users/devices.

HTTP and DNS use different protocols within the same layer. HTTP uses TCP (Transmission Control Protocol) in the transport layer, whilst DNS uses UDP (User Datagram Protocol). TCP can be used by DNS for response messages which exceed 512 bytes in size. These larger web pages/messages cannot be allowed to fail because of errors and TCP provides the necessary, reliable, delivery connection and incorporates both error checks and flow control.

Alternatively, smaller messages can be handled via the connectionless protocol UDP. This is fast method for messages of less than 512 bytes, however it is more unreliable then TCP and provides no flow control. The only error control facility which UDP provides is a checksum which offers comparatively poor protection (Eggert & Fairhurst, 2008) and means that corrupt or lost packets are not repaired but are rather simply re-sent.

Network (Layer 3)

The network layer allows the transfer of data between locations and/or across networks by defining both routing and addressing (Odom, 2003).

Within this layer both TCP and UDP transmit communications (packets or datagrams), via Internet Protocol (IP). This system identifies both the source and destination IP addresses at the network layer to determine which devices are communicating with each other.

At this layer, the network protocol DHCP (Dynamic Host Configuration Protocol), uses UDP on ports 67 and 68 to allow the configuration of devices on the network to allow communication via the use of IP.

There are a number of RFC documents (Request for Comments) which deal with the topic of IP within the network layer, including; RFC 5405 which relates to UDP and RFC 2488 and RFC 3155 which both deal with TCP.

There are other protocols available which work at layer 3 of the OSI model, including; Internet Protocol Security (IPSec), Internet Control Message Protocol (ICMP), and Routing Information Protocol (RIPv1 and RIPv2). Each of the protocols available allow for the successful communication of data between locations, utilising different methods to achieve the same outcome.

Data Link (Layer 2)

The data link layer communicates with both the physical and network layers. It is concerned with how datagrams are sent and segments them in to more manageable frame sizes for easier transfer.

There are two lower layers available; Media Access Control (MAC) and Logical Link Control (LLC). The latter contains the functions needed to both establish and control logical links between devices connected over a network. Whilst MAC provides a unique identification address of (for example) a physical machine, router or network, in order to ensure that the data gets to the correct destination.

A Cyclic Redundancy Check (CRC) may be used to ensure that the data has been received without error.

Similarly to Network layer 3, there are a number of protocols which carry out these functions, including; Wireless LAN, Address Resolution Protocol (ARP) and Point-to-Point Protocol (PPP).

Physical (Layer 1)

The physical layer of the OSI model is where the message is actually sent out over the network, carrying the signals for all of the layers above it (Microsoft, 2002). It defines hardware specifications and transforms data from bits in to transmittable signals which are then sent and received at this layer. This is the layer where physical attributes are identified, including; voltages, physical mediums and connectors.

Similarly to the above two layers, there are various protocols which can be employed at this level, including; Fibre Distribution Data Interface (FDDI) and Ethernet Physical Layer (100Base-T, TX or FX).

Stages of Scenario

In this scenario there are various pieces of hardware (illustrated below in Figure 1) and types of Fast Ethernet connection employed, which are detailed below.

Figure 1 Scenario Hardware in Reference to the OSI Model

Router

A router is used to forward data packets between networks via the most efficient route and maintains a networks connection to the internet. Data is transferred between locations on a single network or from one network to another. The second router in this single network scenario connects the first router and the servers, via the switch.

Information contained with the IP packet header is identified by the router and is used to determine the most appropriate route for the data to be transmitted. The router identifies how busy each possible pathway is and assigns the most efficient route to the packet. All data link layers have been removed at this stage as MAC addresses are no longer required and the router now adds a new link layer protocol header to direct the data to the next destination.

100BaseFX

100BaseFX is a version of Fast Ethernet over optical fibre which connects the second router in this scenario to the switch. HP (2007) highlights the benefits of 100BASEFX:

Supports extension of the network to longer distances than copper cabling can support – can extend up to 2 kilometres in length

Immunity of fibre to external noise interference

Increased security – harder to tap in to fibre or eavesdrop

Electrical immunity – the are no issues with grounding of fibre

Switch

A switch acts as a control which allows devices on a network to communicate with each other in an efficient manner. The switch does not amend the packets which travel through it in any manner. It simply identifies the address of the final destination of the packet and establishes a connection which allows the packet to reach this endpoint.

In this scenario there are two possible destinations for the packets; the DNS Server and the Web Server. The DNS and HTTP frames which are passed to the switch are analysed and the appropriate connection established with either server to allow them to complete their journey.

100BaseTX

100BaseTX is one of the most commonly used forms of Fast Ethernet, running over two copper wires, rather than optical fibre as with 100BaseFX. Cabling can extend up to a maximum of 100metres (significantly shorter then optical fibre can extend) and can provide up to 100Mbits/s of throughput.

DNS Server and Web Server

When the packet reaches its final destination at one of the two end servers (DNS or Web), it is decapsulated and the final DNS or HTTP request can complete the process for which it was originally intended.

In this scenario, the user has pointed a browser at an address which represents a page on the web server. When this request reaches the second router it is passed through the above phases before the HTTP request is received and the appropriate web page is sent back for viewing within the web browser.