Some Background on the Internet


Introduction.  An internet is a network of networks.  The Internet is made up of thousands of networks spread all over the world.

The Evolution of the Internet

The Internet originated with ARPAnet which was accessible to only an elite group of government and university personnel.  One of the goals that the military had when designing this network was to develop it so it could survive war and even nuclear devastation and still provide communications.  The universities were generally more inclined to be interested in ways for researchers to quickly and more easily interact.

In the mid 1980s, the NSF - National Science Foundation began to develop the backbone for the Internet.  While the NSF still provides some services, the backbone development has been privatized and funding was distributed to regional networks.  The Internet infrastructure uses a hybrid mesh topology.  The following diagram represents some of the main transcontinental links within the US.



This sort of topology provides major alternative routes for packets if one link breaks down.

Today's Internet consists of two basic backbone components as outlined below.

  • The Commercial Backbone
    • this is provided by private corporations such as the following
      • AT&T
      • Sprint
      • UUNET
      • BBN Planet
      • Cable & Wireless USA
    • These backbone providers have established peering agreements to carry one another's network traffic
    • ISPs - Internet Service Providers connect to these commercial networks
    • the regional Bell operating companies own much of the physical cabling and lease it to providers
      • provider networks connect with T-1, T-3 or OC-3 lines
    • connected to NAPs - National Network Access Points
  • vBNS - Very High Speed Backbone Internet Service
    • this connects five supercomputer centers for scientific purposes
    • they are located at the following locations
      • NCSA - National Center for Supercomputing Applications in Urbana, Illinois
      • NPACI - National Partnership for an Advanced Computing Infrastrcuture in San Diego, California
      • Cornell Theory Center in Ithaca, New York
      • National Center for Atmospheric Research in Boulder, Colorado
      • Pittsburgh Supercomputing Center in Pittsburgh, Pennsylvania
    • NSF funded in a contract with MCI
    • links are OC-3 of at least 155 Mbps
    • connected to NAPs - National Network Access Points

Internet2.  The vBNS is a part of the Internet2 Project which is a joint effort of US universities and the US government.  It is aimed at developing a high speed backbone that can be used as a test bed for deployment of new technologies.  For example, Internet2 has deployed the following.

  • IPv6 - IP version 6
  • Qbone a QoS - Quality of Service initiative

Internet2 is not a separate physical network and it is not intended to replace the existing Internet.  It is a consortium of organizations that have banded together to develop and test new technologies before they are deployed on the commercial Internet.

The Components of the Internet

Internet Components.  In the following sections we outline a discussion of the following components of the Internet.

  • Local computer or LAN connected to the Internet
    • requires appropriate software and protocols
      • modem
        • requires dial up software
      • NIC on a LAN
    • TCP/IP is the protocol stack
  • ISP - Internet Service Provider
    • a computer becomes a remote client on the ISPs local network
    • ISP develops capability to provide a point of presence or access point
    • ISPs usually lease high speed dedicated lines from a telephone company
    • Large ISPs can have their own dedicated high speed connections
    • Small ISPs might actually go through a regional ISP to get to the Internet backbone
  • Regional Networks
    • some examples of regional providers
      • BARRNet - Located in north central California
      • Westnet - Covers western part of the US
      • NEARNET and NYSERNet - Covers northeastern US
      • MIDnet - Covers central part of US
      • SURAnet - Covers southeastern part of US
      • CICnet - Covers midwestern part of US
  • NAPs - National Network Access Points
    • originally developed and funded by NSF
    • located in following cities with operations by following companies
      • New York - Sprint
      • Washington DC - WorldCom
      • Chicago - Ameritech
      • San Francisco - Pacific Bell
    • they provide switching facilities
    • not all Internet traffic must go through them
  • MAEs - Metropolitan Area Exchanges
    • Point where ISPs connect to each other and traffic is switched between them
    • First Tier MAEs
      • WorldCom - Washington DC NAP = MAE EAST
      • WorldCom Silicon Valley = MAE WEST
    • Major ISPs such as Sprint and Netcom connect to the first tier MAEs
    • Second Tier MAEs
      • MAE Chicago
      • MAE Dallas
      • MAE Houston
      • MAE Los Angeles
      • MAE New York
      • these MAEs do not route the data
      • the ISPs connecting to the MAE do the routing
      • the MAEs do house the routers which belong to the ISPs
    • MAEs require huge routing tables
    • Switching platforms are owned by WorldCom linked via FDDI
    • MAE devices are administered through Sun workstations using SNMP
    • Regional and smaller ISPs connect to the second tier MAEs

How Internet Components Work Together.  It isn't enough to have the components.  they need to be able to work together.  The following diagram presents a simplified version of the sequence of events associated with sending messages from a home computer to the Internet.



The following outlines the sequence of steps involved in just sending an e-mail message to the Internet.
  1. Data is broken down into manageable chunks, called packets.
  2. Networking protocols add header and trailer information.
  3. The binary data is converted to electrical signals or light pulses to travel over appropriate physical medium.
  4. Data travels
    1. If your computer is on a LAN then it travels to a server or router that is connected to some sort of phone or dedicated line.
    2. If a modem is used the data is encapsulated into a link protocol such as PPP - Point to Point or SLIP - Serial Line Internet Protocol and the digital signals are modulated to travel over an analog phone line.
  5. The signal reaches the ISP's remote access server which is configured to accept dial-in connections.  Though you might have a direct link to the ISP.  You're very likely going to need to login to this server.
  6. The source computer then becomes a remote node on the ISP's local network.
  7. The data travels from the ISP's server to the regional network to which the ISP is connected.  This step may be skipped if you have a larger provider.
  8. Your data travels through the major NAPs, if necessary, and onto the commercial Internet backbone.
  9. At the other end the data is likeliest to go through another NAP, another regional network and an ISP at the receiving side.
  10. This ISP then delivers the data to the destination computer.
  11. The data is finally delivered to the appropriate user when the user's client machine does the appropriate login to the ISP's or organization's e-mail server.