1. 1 26-Jun-16 Switches CCNA Exploration Semester 3 Chapter 2-Part 1 Warning – horribly long! Taken from Slides provided by S Ward Abingdon and Witney College and St. Clair College in Windsor, Ontario

2. 2CSE490, SKZ CCNA Exploration Semester 3 LAN DesignBasic Switch Concepts VLANs VTP STP Inter-VLAN routing Wireless

3. 3CSE490, SKZ Topics Key Elements of Ethernet/802.3 Networks Switches and how they forward frames Configure a switch Basic security on a switch

4. 4CSE490, SKZ Key Elements of Ethernet /802.3 Networks

5. 5CSE490, SKZ CSMA/CD Shared medium Physical shared cable or hub. Ethernet was designed to work with collisions. Uses carrier sense multiple access collision detection (CSMA/CD) for media access.

6. 6CSE490, SKZ CSMA/CD reminder Device needs to transmit. It “listens” for signals on the medium. If finds signals – it waits. If clear – it sends. Carry on listening. If it receives while sending the first 64 bytes of the frame then collision. Stop sending frame, send jam signal. Wait for random time (backoff) Try again – listen for signals etc.

7. 7CSE490, SKZ NOTES: No collisions Fully switched network with full duplex operation = no collisions. Higher bandwidth Ethernet does not define collisions – must be fully switched. Cable length limited if CSMA/CD needed. Fibre optic – always fully switched, full duplex. (Shared medium must use half duplex in order to detect collisions.)

8. 8CSE490, SKZ Ethernet Communications Efficiency is typically rated at 50 to 60 percent of the 10-Mb/s bandwidth. 100 percent efficiency in both directions.

9. 9CSE490, SKZ Switch Port Settings AUTO: Auto-negotiation of duplex mode. The two ports communicate to determine the best mode. Default for FastEthernet and 10/100/1000 ports. Auto is fine if both devices are using it. Potential problem if switch uses it and other device does not. Switch defaults to half. FULL: Full-duplex mode. Default for 100BASE-FX ports. HALF: Half-duplex mode. Full one end and half the other – errors.

10. 10CSE490, SKZ Auto-MDIX feature: Command makes switch detect whether cable is straight through or crossover and compensate so you can use either. Depends on IOS version Enabled by default from 12.2(18)SE on Switch Port Settings Cross Over CableStraight Through Cable

11. 11CSE490, SKZ Ethernet Communications

12. 12CSE490, SKZ Communication types reminder Unicast – to a single host address e.g. most user traffic: http, ftp, smtp etc. Broadcast – addressed to all hosts on the network e.g. ARP requests. Multicast – to a group of devices e.g. routers running EIGRP, group of hosts using videoconferencing. IP addresses have first octet in range 224 – 239.

13. 13CSE490, SKZ Ethernet Communications Ethernet Frame: Minimum 64 bytes, Maximum 1518 bytes Preamble/SOFD: To synchronize. Destination Address: MAC Address of destination device. Source Address: MAC address of source device. Length/Type: Length of frame or protocol type code. Data: Encapsulated data from OSI Layers 7 to 3. FCS: Frame Check Sequence.

14. 14CSE490, SKZ MAC address 48-bits written as 12 hexadecimal digits. Format varies: 00-05-9A-3C-78-00, 00:05:9A:3C:78:00, or 0005.9A3C.7800. MAC address can be permanently encoded into a ROM chip on a NIC - burned in address (BIA). Some manufacturers allow the MAC address to be modified locally.

15. 15CSE490, SKZ Ethernet Communications MAC Address: 12 hexadecimal digits Broadcast: Indicates a broadcast or multicast frame. Local: indicates whether the address can be modified locally. OUI Number: Manufacturer of the NIC, allocated by IEEE Vendor Number: Unique identifier for port on device, vendor assigned number.

16. 16CSE490, SKZ Switch MAC Address Table Table matches switch port with MAC address of attached device Built by inspecting source MAC address of incoming frames Destination MAC address checked against table, frame sent through correct port If not in table, frame flooded Broadcasts flooded

17. 17CSE490, SKZ Switch MAC Address Table Example Step 1: The switch receives a broadcast frame from PC 1 on Port 1. The switch receives a broadcast frame from PC 1 on Port 1.

18. 18CSE490, SKZ Switch MAC Address Table Example Step 2: The switch enters the and the into the address table. The switch enters the source MAC address and the switch port that received the frame into the address table.

19. 19CSE490, SKZ Example Step 3: Because the destination address is a broadcast, the switch Because the destination address is a broadcast, the switch floods the frame to all ports, except the port on which it received the frame. Switch MAC Address Table

20. 20CSE490, SKZ Example Step 4: The destination device replies to the broadcast with a The destination device replies to the broadcast with a unicast frame addressed to PC 1. Switch MAC Address Table

21. 21CSE490, SKZ Example Step 5: The switch enters the source MAC address of PC 2 and the port number of the switch port that received the frame into the address table. The switch enters the source MAC address of PC 2 and the port number of the switch port that received the frame into the address table. Switch MAC Address Table

22. 22CSE490, SKZ Example Step 6: The switch can now forward frames between source and destination devices because it has entries in the address table that identify the associated ports. The switch can now forward frames between source and destination devices because it has entries in the address table that identify the associated ports. Switch MAC Address Table

23. 23CSE490, SKZ Design Considerations – Collision Domains A of Ethernet is. A major disadvantage of Ethernet is collisions. to either eliminate or reduce collisions. A hub offers no mechanisms to either eliminate or reduce collisions. Shared medium – same collision domain. The more devices – the more collisions. A Switch (+ full duplex) dedicated link each way 100% bandwidth in each direction. Link regarded as an individual collision domain if you are asked to count them.

24. 24CSE490, SKZ How many collision domains?

25. 25CSE490, SKZ How many collision domains? 11

26. 26CSE490, SKZ Broadcast Domains Layer 2 switches flood broadcasts. Devices linked by switches are in the same broadcast domain. A layer 3 device (router) splits up broadcast domains, does not forward broadcasts Destination MAC address for broadcast is all 1s, that is FF:FF:FF:FF:FF:FF (We ignore VLANs here – they come later.)

27. 27CSE490, SKZ Interconnecting switches extends the broadcast domain. Broadcast Domains

28. 28CSE490, SKZ How many broadcast domains? No VLANs

29. 29CSE490, SKZ How many broadcast domains?

30. 30CSE490, SKZ Design Considerations – Network Latency is the time a frame or a packet takes to travel from the source to the final destination. Latency is the time a frame or a packet takes to travel from the source to the final destination. NIC Delay Propagation Delay Intermediate Devices Delay NIC Delay

31. 31CSE490, SKZ Design Considerations – Network Congestion : Most common causes: More powerful PCs can send and process more data at higher rates. Increasing use of remote resources (servers, Internet) generates more traffic. More broadcasts, more congestion. Applications make more use of advanced graphics, video etc. Need more bandwidth.

32. 32CSE490, SKZ Design Considerations – Network Congestion Solution: Segmenting LANs into smaller parts LANs are segmented into a number of smaller and using routers and switches. LANs are segmented into a number of smaller collision and broadcast domains using routers and switches. The is to isolate traffic and to achieve better use of bandwidth per user. The primary reason is to isolate traffic and to achieve better use of bandwidth per user.

33. 33CSE490, SKZ Design Considerations – Network Congestion HubHub No LAN Segmentation:

34. 34CSE490, SKZ HubHub JAMJAMJAMJAM JAMJAMJAMJAM JAMJAMJAMJAM JAMJAMJAMJAM Design Considerations – Network Congestion No LAN Segmentation:

35. 35CSE490, SKZ Broadcast Domain SwitchSwitch Collision Domains LAN Segmentation: Design Considerations – Network Congestion Solution

36. 36CSE490, SKZ Broadcast Domains LAN Segmentation: RouterRouter Collision Domains Design Considerations – Network Congestion Solution

37. 37CSE490, SKZ Controlled Collision and Broadcast Domains

38. 38CSE490, SKZ Design Considerations – Network Congestion Control latency Consider the latency caused by each device on the network. Consider the latency caused by each device on the network. Removing Bottle Necks Use a faster link. Have several links and use link aggregation so that they act as one link with the combined bandwidth.

39. 39CSE490, SKZ Remove bottlenecks

40. 40CSE490, SKZ Forwarding Frames Using a Switch Two Methods: Store and Forward. Cut Through. Fast Forward Fragment Free Cisco switches now all use Store and Forward. Some older switches used Cut Through.

41. 41CSE490, SKZ Store-and forward: Receives the entire frame. Receives the entire frame. Computes the CRC and checks the frame length. Computes the CRC and checks the frame length. If valid, checks the switch table for the destination address and forwards the frame. If valid, checks the switch table for the destination address and forwards the frame. If invalid, the frame is dropped. If invalid, the frame is dropped. Allows entry and exit at different bandwidths Switch Forwarding Methods DestinationSourceDataFCS 123896745 = 123896745 CRC Frame is Good Destination found in Switching Table

42. 42CSE490, SKZ Cut Through - Fast forward Read start of frame as it comes in, as far as end of destination MAC address (first 6 bytes after start delimiter) Look up port and start forwarding while remainder of frame is still coming in. No checks or discarding of bad frames Entry and exit must be same bandwidth Lowest latency Switch Forwarding Methods

43. 43CSE490, SKZ Cut Through - Fragment Free Read start of frame as it comes in, as far as end of byte 64 Look up port and start forwarding while remainder of frame (if any) is still coming in. Discards collision fragments (too short) but other bad frames are forwarded Entry and exit must be same bandwidth Compromise between low latency and checks Switch Forwarding Methods

44. 44CSE490, SKZ Symmetric: All ports are of the same bandwidth. All ports are of the same bandwidth. Optimized for a reasonably distributed traffic load. Optimized for a reasonably distributed traffic load. Symmetric and Asymmetric Switching

45. 45CSE490, SKZ Asymmetric: Provides switched connections between ports of unlike bandwidth. Provides switched connections between ports of unlike bandwidth. For example, more bandwidth can be assigned to a server to prevent bottlenecks. For example, more bandwidth can be assigned to a server to prevent bottlenecks. Symmetric and Asymmetric Switching

46. 46CSE490, SKZ A switch analyzes some or all of a packet before it forwards it to the destination host based on the forwarding method. A switch analyzes some or all of a packet before it forwards it to the destination host based on the forwarding method. It stores the packet for the brief time in a It stores the packet for the brief time in a memory buffer. Built into the hardware Built into the hardware Two types: Port based. Port based. Shared. Shared. Memory Buffering

47. 47CSE490, SKZ Port Based: Each incoming port has its own queue. Frames stay in buffer until outgoing port is free. Frame destined for busy outgoing port can hold up all the others even if their outgoing ports are free. Each incoming port has a fixed and limited amount of memory. Memory Buffering

48. 48CSE490, SKZ Shared: Deposits all frames into a common memory buffer that all the ports on the switch share. Deposits all frames into a common memory buffer that all the ports on the switch share. The amount of buffer memory required by a port is dynamically allocated. The amount of buffer memory required by a port is dynamically allocated. The frames in the buffer are linked dynamically to the destination port. The frames in the buffer are linked dynamically to the destination port. Allows the packet to be received on one port and then transmitted on another port, without moving it to a different queue. Allows the packet to be received on one port and then transmitted on another port, without moving it to a different queue. Memory Buffering

49. 49CSE490, SKZ Layer 2 and Layer 3 Switching Traditional Ethernet switches work at layer 2. They use MAC addresses to make forwarding decisions. They do not look at layer 3 information. Cisco Catalyst 2960 Series

50. 50CSE490, SKZ Layer 2 and Layer 3 Switching Layer 3 switches can carry out the same functions as layer 2 switches. They can also use layer 3 IP addresses to route between networks. The can control the spread of broadcasts. Cisco Catalyst 3560 Series

51. 51CSE490, SKZ Layer 3 switches do not completely replace the need for routers on a network. Routers perform additional Layer 3 services that Layer 3 switches are not capable of performing. Routers perform additional Layer 3 services that Layer 3 switches are not capable of performing. Routers and Switches

52. 52 26-Jun-16 Part 1- End