1. Ethernet Switching Md.Ariful Islam(Shohag) B.Sc(Hons) in CS Microland E-Mail: sohagcs@yahoo.comsohagcs@yahoo.com Mob : +880175100015

2. Objectives Ethernet switching Collision domains and broadcast domains

3. Objective 1.Shared Ethernet works extremely well under ideal conditions. 2.When number of users increase, the number of collision can significantly reduce the performance of the network. 3.Switch and Bridge are evolved to overcome this situation. 4.We need to control collision and broadcast.

4. Layer 2 Bridging Problem: 1.Ethernet is a shared media. Only one node can transmit data at a time. 2.The addition of more nodes increases the demands on the available BW and places load on the media. 3.Thus increase the probability of collisions, which results in more retransmission. Solution: 1.Break the large segment into parts and separate it into isolated collision domains. 2.To accomplish this a bridge keeps a table of MAC addresses and the associated ports. 3.The bridge forwards or discards frames based on the table entries.

5. Bridge Operation Explain A ping B A ping C … B ping A

6. Layer 2 Switching 1.Generally, a bridge has only two ports, divides the collision domain into two parts. 2.Decisions are based on MAC addresses. 3.It is no effect on broadcast or logical domain. 4.If networks has no router the entire network will share the same logical or broadcast domain. 5.A bridge will create more but will not add collision domainbroadcast domain 6.A switch is essentially fast, multi-port bridge that can contain dozens of ports. 7.Each ports create its own collision domain. 8.A switch dynamically builds and maintains a content-addressable memory table (CAM), which holds all of the necessary MAC information for each port.

7. Switch Operation - Microsegment 1.A switch is simply a bridge with many ports. 2.When only one node is connected to a switch port  Collision domain contains two nodes Switch port Host  This small physical segment is called microsegment

8. Switch Operation-Full Duplex 1.In a network with Twisted-Pair cabling, one pair is used to carry the signal and other pair is used for receive signal. 2.It is possible for signals to pass through both pairs simultaneously. 3.The ability to communicate in both directions is known as Full Duplex. 4.Most switches are capable of supporting full duplex,as are most NICs. 5.Thus no contention for the media. 6.A collision domain no longer exists.

9. Switch Operation-CAM 1.CAM is a memory 2.Works backward compared to conventional memory 3.When data is entered into the memory it will return the associated address. 4.CAM allows the switch to find the port that is associated with MAC address without search algorithm. 5.ASIC (Application Specific Integrated Circuit) allows some Software operations to be dome in hardware. 6.Theses technology greatly reduced the delays caused by software and enable a switch to keep up high bit rates.

10. Switches Review

11. Switch Benefits

12. Switching Table

13. Latency 1.Latency is the delay between the time a frame first starts to leave the source device and the time the first part of the frame reaches its destination. 2.Causes of delay 1.Media delay 2.Circuit delay caused by the electronics that process the signal along the path 3.Software delay caused by the decision software must take to implement switching decision 4.Delays caused by the content of the frame and location of the frame switching decisions.

14. Switch Modes How a frame is switched to the destination port?  Latency  Reliability Cut-through switching: A switch can start to transfer the frame as soon as the destination MAC address is received. No Error checking is possible Store-and-forward switching: 1.The switch can receive the entire frame before sending it out the destination port. 2.This gives the switch software an opportunity to verify the frame check sum (FCS). Trade off

15. Switch Modes Fragment Free: 1.It is a compromise between Cut-through and Store-and-Forward. 2.It reads 1 st 64 bytes, it includes Frame header 3.Starts to send out before the entire data and FCS are read. 4.This mode verifies reliability of the addresses and LLC protocol information to ensure the data will be handled properly and arrive at the correct destination. 5.Ensures that frame is not a fragment. No runt. Symmetric Switching: When Cut-through switching is used, the source and destination ports must have the same bit rate to keep the frame intact. This is called symmetric. Asymmetric Switching: 1.If the bit rates are not same, frame must be stored at one bit rate and send out at other bit rate. 2.Suitable for client server application

16. S1S2 F Initial Frame F2F2 F1F1 Switching Loops Loops can occur when extra switches and bridges are added to provide redundant paths for reliability and fault tolerance.

17. Spanning Tree Protocol 1.Solution to this difficulty for the bridges/switches is to communicate with each other and overlay the actual topology with a spanning tree that reaches every segment. 2.A switch sends special messages called bridge protocol data units (BPDUs) out all its ports to let other switches know of its existence. 3.The switches use a spanning-tree algorithm (STA) to resolve and shut down the redundant paths 4.The protocol used to resolve and eliminate loops is known as the Spanning Tree Protocol (STP). 5.STP is a standard-based routing protocol to avoid routing loops. Root? Algorithms continues to Run?

18. Spanning Tree Protocol

19. Root bridge Designated port Root port A B S1 S2 A B

20. Shared Media Environments 1.It is important to be able to identify a shared media environment, because collisions only occur in a shared environment. 2.Some networks are directly connected and all hosts share Layer 1: –Shared media –Extended shared media –Point-to-point network

21. Collision Domains 1.Collision domains are connected physical network segment where collision can occur. 2.Causes network to be inefficient 3.Every time a collision happens, all transmission stops for a period of time 4.The length of this period is determined by backoff algorithm. When two bits are propagated at the same time on the same network, a collision will occur.

22. Collision Domains 1.Types of devices that interconnect the media segments define collision domains. 2.They are classified as Layer 1 devices Layer 2 devices Layer 3 devices 3.Layer 2 and 3 devices breaks collision domains.

23. Shared Media, Repeaters, Hubs, and Collision Domains Shared media Extended by a repeater Extended by a hub Extended by a hub and repeater Each is a single collision domain!

24. The Four-Repeater Rule Legacy Ethernet network: No more than 4 repeaters or repeating hubs can be between any 2 computers on the network. –From A to B is 4 repeaters. A B

25. Four-Repeater Rule Example The 5-story building shown violates the four-repeater rule because host A and B are 5 repeaters apart. Hubs would cause the same result. Even if all servers were on the third floor, and A and B would never communicate directly; they are too far to hear each other transmit and can cause data collisions. If 4 repeater rule is violated, the maximum delay limit may exceeded. A late collision is when a collision happens after the first 64 bytes of the frame are transmitted. Late collision frames add delay that is called consumption delay. A B

26. Four-Repeater Rule Example Layer 1 Solution The hub added, which could be on any floor, allows us to comply with the four-repeater rule. No 2 hosts are more than 3 repeaters apart. It really wouldn’t matter if each floor connects to the hub. How many collision domains do we have? Still only one and getting bigger with each floor. A B Hub

27. Too Much Traffic / Too Many Hosts in a Collision Domain…

28. The 5-4-3-2-1 Rule This rule requires that following guidelines should not be exceeded: A.5 segments of network media B.4 repeaters or hubs C.3 hosts segments of the network D.2 link section with no hosts E. 1 Large Collision Domain

29. Segmentation One Important Skill for a Networking Professional is the ability to recognize collision domain. Layer 1 devices? Layer 2 devices? Layer 3 devices?

30. Layer 1 Devices Extend Collision Domains

31. Segmenting with Bridges

32. Segmenting with Switches

33. Segmenting with Routers

34. Limiting the Collision Domain The only devices that can segment collision domains are bridges, switches (both Layer 2), and routers (Layer 3).

35. Layer 2 Broadcasts 1.To communicate with all collision domains, Protocols use broadcast and multicast frame at layer 2 of the OSI model. 2.Broadcast Frame has a destination MAC 0xFFFFFFFFFFFF. It is an Address of every host. 3.Layer 2 devices must flood all broadcast or multicast traffic. The accumulation of broadcast and multicast traffic from each device in the network is referred to as broadcast radiation. 4.The circulation of broadcast radiation can saturate the network so that …? This is situation is called Broadcast Storm. 5.The probability of Broadcast Storms increases as the switched network grows. 6.A NIC must rely on the CPU to process each broadcast/multicast group it belongs to. Thus, broadcast radiation affects the performance of hosts in the network.

36. Layer 2 Broadcasts 1.A host does not benefited if it processes a broadcast when it is not the intended destination. 2.Three sources of broadcasts or multicasts in IP network a.Workstation – (ARP: MAC?) b.Routers Routing protocols Workstations can run RIP for redundancy and reachable policy Routing Table broadcasting in every 30 sec c.Multicast applications Non group members suffer due to shared media

37. Data Flow Through a Network

38. Network Segment 1.If the segment is used in TCP, it would be defined as a separate piece of the data. 2.If segment is being used in the context of physical networking media in a routed network, it would be seen as one of the parts or sections of the total network.

39. The End