1. LAN Wiring & Physical Interfaces Lecture 6

2. Network Interfaces Every device has some form of a network interfaces. A PC has a Network Interface Card (NIC). The Cisco 2501 routers have an Attachment Unit Interface (AUI) connection.

3. NICs NICs are dependent on what type of topology exists in the LAN. For example, an Ethernet NIC can’t be used with a Token Ring NIC, and vice- versa.

4. NICs (cont.) A NIC performs most of the network-level operations in order to reduce CPU overhead. Receives incoming packets Storage of frames Actively listens on the network A NIC is exactly the same as any other I/O device to the CPU. Read and write from the data bus with certain addressing.

5. NIC (cont.) Ethernet NIC RJ-25 BNC (British Naval Connector) AUI ST/SC for fiber The NICs pictured at the right have RJ-45 (top) and ST (bottom) connectors.

6. NIC (cont.) In addition to providing the physical connection to the network, the NIC also deals with hardware addressing. All LAN traffic ultimately deals with hardware addresses. The NIC will only read (copy) packets destined for it’s own address. Broadcasted packets are also read.

7. Physical Hardware Addressing Three separate types of addressing occur: Static – hardware manufacture specifies the address, and it is unchangeable.  Advantages: Ease of use, and permanence.  Disadvantages: manufactures must keep track of usage. Configurable – the end user can modify the address through software or switches on the card.  Advantages & Disadvantages: A mix of above and below. Dynamic – random numbers are tried until one is found that is unused.  Advantages: smaller HW addresses!  Disadvantages: lack of permanence, potential conflict.

8. Broadcasting Broadcasting refers to the concept that a certain type of packet can be generated that is received by all hardware attached on the network. Useful for finding printers, file shares, etc. Broadcasting usually uses a specific broadcast address for such schemes.

9. Thick Ethernet Wiring

10. Thick Ethernet Specifications 50 Ohm coaxial cable, 0.4” thick. 10base5 – 10Mbps, 500m segments. Ends must be terminated. AUI connectors. “Vampire” taps / Transceivers.

11. Thin Ethernet Wiring

12. Thin Ethernet Specifications 50 Ohm coaxial cable, 0.2” thick. 10base2 – 10Mbps, 185m segments. Ends must be terminated. BNC connectors. Transceivers are integrated into the NIC.

13. Twisted-Pair Ethernet Wiring

14. Twisted-Pair Ethernet Specifications Twisted-pair cable. 10baseT- 10Mbps, 100m segments. 100baseT – 100Mbps, 100m segments. RJ-45 connectors. Hubs/Switches required.

16. Extending the LAN Hardware has been developed that “boosts” the signal so you can create larger LAN segments. Fiber Extensions Repeaters Bridges Switches Hubs

17. Why Extend the LAN? Distance limitation is the most common hurdle for LAN design. You always want more. More computers, more people, more access to shared resources. Users want access to shared resources. It doesn’t make economical sense to have a single printer for each segment, for example.

18. Fiber Modems Fiber modems are used to connect a single remote computer to an established LAN. Effective, since fiber is low latency, high bandwidth.

19. Repeaters Repeaters are just as they sound. They connect two segments together as one, and just repeat the traffic of one segment and forward it on to the second. Repeaters are unintelligent. What you get on one segment is what you get on the other. No more, no less.

20. Repeaters (cont.) A repeater as it’s commonly used. Hubs and switches have repeaters built-in, so you can connect multiple segments together. IEEE designates how many segments may be connected together.

21. Repeaters (cont.) Repeaters have several drawbacks: They expand the collision domain. Noise from one segment flows on to the second segment. Any electrical problems that occur on one segment will be sent to the second segment!

22. Another Common Use of Repeaters

23. Bridges Bridges are smarter repeaters. Main difference: bridges verify that the frame is intact and valid before it is forwarded on to the second segment. This provides a higher quality of service. Problems on one segment do not affect the other segment.

24. Bridges (cont.) Bridges also have the attractive feature of “frame filtering.” A frame will only be forwarded to the second frame if the destination computer is not on the first frame. This reduces chatter on the combined segments. Bridges that learn which computers are on which segments are called “adaptive” or “learning” bridges.

25. Adaptive Bridges When a bridge boots, it does not have any information about computers on the segments that it connects. After a long time, the bridge reaches a steady-state where it has developed a mapping of which computers are on which segments. At this steady-state, extra frames on the segments is cut to a minimum. In this state, maximum parallelism is achieved.

26. Bridging Between Buildings

27. Switching

28. Switches (cont.) Switches allow for independent, parallel communications between ports on the switch. Switches are combined with hubs for a cost-effective method of providing equal-access for all users.