Boise State University, College of Business & Economics

Page updatedTuesday, January 19, 1999 04:10 PM 

  IS380 Telecommunications

Spring 1999 

Home Schedule Homework Nicenet PROJECT presentations Overheads Text Resources Mags

Chapter 8: High Speed LAN and Backbone Networks


Backbone networks are closely related to many internetworking topics, and internetworking is arguably the fastest growing specialty area within data communications and networking.

Backbone network component terminology can be very confusing, but the OSI model provides one way to reduce the confusion. Here are the four most common backbone network components in terms of how they transfer network traffic from one end system to another:

Repeaters operate at layer 1, the physical layer. They simply receive bits and re-transmit them. This can be useful in extending the size and distance limitations of LANs.

Hubs, such as those commonly used with 10BaseT Ethernet, are multi-port repeaters. They commonly have 8, 12, 24, or more ports and make a logical bus into a physical star by having all cables make a "home run" from workstation to hub.

Repeaters generally are not helpful in reducing the size of a collision domain.

Bridges operate at layer 2, the data link layer. They forward and filter frames as necessary. This can be useful in reducing the size of collision domains (improving performance) and enhancing security.

Switching hubs, such as those now commonly used for Ethernet backbones, are multi-port bridges. They provide a kind of circuit switching between communicating devices to avoid the contention problems of shared media (in effect reducing the collision domain to just the two communicating devices).

Routers operate at layer 3, the network layer. They open a data link layer frame, look at the network layer address, and then create another data link layer frame to send out. Because of this, they can connect networks with different data link layers, such as Ethernet and Token Ring networks.

Multi-protocol routers can provide routing for several network-layer protocols at once, such as IP (from TCP/IP) and IPX (from NetWare), and are the central component in many present-day collapsed backbone network architectures.

Layer 3 switching, also called IP switching, is a recent internetworking development.

Gateways generally operate higher than the network layer, often at the application layer. They are used to interface radically different communications architectures, such as PCs on a LAN interfacing with legacy mainframe SNA applications.

Gateways commonly have to perform functions at the session layer (establishing and maintaining sessions) and the presentation layer (code conversion, etc.) as well as the application layer.

The term gateway is often used differently in various contexts; e.g., routers are sometimes called gateways in TCP/IP and Internet terminology.


Network Components

Device Layer Messages Physical Layer Data Link Layer Network Layer
Repeater Physical All Transferred Same/Different Same Same
Hub Physical All Transferred Same/Different Same Same
Bridge Data Link Filtered Same/Different Same Same
Switch Data Link Routed Same/Different Same Same
Router Network Routed Same/Different Same/Different Same
Brouter D.L. & N.W. Filtered & Routed Same/Different Same/Different Same
Gateway N.W. & Transport Routed Same/Different Same/Different Same/Different

Questions for class discussion:

  1. Media speed is often not the bottleneck (limiting factor) in network speed. When internetworking devices such as bridges and routers are involved, what bottlenecks might they introduce?
  2. Data of almost any kind can be carried from one endpoint to another across a network through a technique called encapsulation (also called tunneling). How is this done?
  3. Data of almost any kind can also be carried across a network through protocol translation. How is this different from encapsulation?
  4. A 4MB packet comes from network A to a router and is destined for network B which can handle a maximum packet size of 1MB. What will the router do?
  5. Sometimes network performance can be improved by replacing workstation hardware, NICs, cable, hubs, software, or server hardware. How do you decide which should get attention?
  6. Suppose network response time appears to be unacceptably slow in a computer lab on campus. Describe several options you would investigate to improve the situation.
  7. Draw a typical architecture involving a collapsed backbone, switching hubs, and shared media (ordinary) hubs.
  8. When a network manager needs to decide among competing technologies that accomplish similar goals (e.g, various 100Mbps Ethernet alternatives), on what should the decision be based?
  9. Additional questions and discussion issues from the class are welcome.

Terms:

  1. 10/100 switched Ethernet
  2. 100BaseT
  3. 100VG-AnyLAN
  4. Asynchronous Transfer Mode (ATM)
  5. Backbone network, collapsed
  6. Backbone network, dispersed
  7. Bridge
  8. Brouter
  9. Encapsulation
  10. Fast Ethernet
  11. Fiber distributed data interface (FDDI)
  12. Gateway, original Internet definition
  13. Gateway, OSI definition
  14. Hub (ordinary, "smart," and switching)
  15. Internetworking
  16. Multiprotocol router
  17. Router
  18. Switch
  19. Switched Ethernet
  20. Switching hub

Other terms important to this chapter were introduced and/or listed with prior chapters.


Last modified on November 10, 1998 12:30 PM.