Bài giảng Network+ Certification - Chapter 3, Network Connections

Chapter 3, Network Connections |1| Chapter Overview A. Bridging B. Switching C. Routing Chapter 3, Lesson 1 Bridging |2| 1. What Is a Bridge? A. A data-link layer device that 1. Connects networks 2. Filters packets a. Propagates only packets destined for the other side of the bridge b. Can reduce network traffic and collisions c. Can lessen delays caused by high traffic levels |3| 2. Connecting LANs with a Bridge A. Usually a hardware device with two ports B. Connects two LANs or splits an existing LAN in half C. Bridges operate in promiscuous mode, reading and processing all of the packets transmitted over the network segments. 1. Network interface adapters normally process only the packets that are addressed to them and discard the rest. D. Bridge functions 1. Bridges can read the source and destination addresses in data-link layer protocol headers. a. If the destination address of a packet arriving from one network segment is that of a computer on the other segment, the bridge transmits it out through the other port. b. If the destination address is that of a computer on the same network segment as the computer that generated it, the bridge discards the packet. 2. The bridge eliminates unnecessary traffic from both segments. 3. Bridges cannot operate any higher than the data-link layer and cannot read any part of the packet other than the data-link layer protocol header. |4| E. Bridges and collisions 1. A collision domain is a network (or part of a network) that is constructed so that a collision occurs when two computers transmit packets at precisely the same time. 2. Adding a hub to a LAN increases its size but maintains a single collision domain because hubs relay signals immediately at the physical layer, without filtering them. 2 Outline, Chapter 3 Network+ Certification, Second Edition a. Hubs do not relay packets immediately; they wait until the entire packet is received. 3. Because the bridge delays propagation, computers on opposite sides of the bridge transmitting at the same time do not cause a collision. 4. Adding a bridge to a LAN splits it into two collision domains, resulting in fewer collisions and increased efficiency. F. Bridges and broadcasts 1. Terms a. A broadcast message is a packet transmitted to all of the computers on the network. b. A unicast message is a packet transmitted to one computer. c. A multicast message is a packet transmitted to a group of computers. d. A broadcast domain is a group of computers that all receive broadcast transmissions sent by any computer in the group. |5| 2. Broadcasts are essential to LANs. a. Used by Address Resolution Protocol (ARP), Dynamic Host Configuration Protocol (DHCP), and other protocols to locate specific computers on the network |6| 3. Bridges propagate all broadcast packets without filtering them. 4. The network segments on either side of a bridge are in the same broadcast domain, and are therefore still part of the same LAN. |7| G. Transparent bridging 1. Transparent bridging is a method for automatically compiling a bridge’s address tables. 2. How do bridges know which computers are located on each network segment? a. Bridges maintain an address table for each of the network segments. b. After reading the source and destination addresses of each packet, the bridge compares them to the tables to determine whether to propagate them. 3. Where do the address tables come from? a. Originally, administrators had to create the tables manually. b. Today’s bridges use a technique called transparent bridging to compile their own address tables. c. As the bridge processes a packet, it adds its source address to the table for the segment from which it arrived. d. When a bridge lacks the information needed to filter a packet, it propagates the packet by default. H. Spanning tree algorithm (STA) 1. STA is a protocol that bridges use to select one of the bridges to process packets. 2. Using multiple bridges to connect two network segments provides fault tolerance. 3. Redundancy can cause data loss when bridges have conflicting information in their tables. Outline, Chapter 3 3 Network+ Certification, Second Edition 4. Multiple bridges can forward broadcasts endlessly, which is called a bridge loop. 5. The STA lets the bridges work together to select one bridge to function, while the others remain idle. I. Source route bridging 1. Token Ring systems use source route bridging instead of transparent bridging and the STA. 2. Workstations themselves select the bridge they will use. a. Each computer transmits an All Rings Broadcast (ARB) message. b. Each bridge receiving the ARB adds a route designator to it, identifying the bridge and port through which it received the packet. c. Each computer receiving an ARB returns it to its source. 3. Bridges use the route designators to avoid sending packets to the same bridge twice. 4. The original source system uses the returned packets to determine which bridge provides the most efficient route through the network to a given destination. |8| 3. Bridge Types A. The three bridge types are local, translation, and remote. B. Local bridge 1. Connects network segments of the same type and at the same location 2. Does not modify the data in the packets; it simply passes on the packets or discards them |9| C. Translation bridge 1. Connects network segments at the same location that use different media or different protocols 2. Strips the data-link layer frame off the packets and packages them in a new frame for transmission on the other segment 3. Disadvantages: a. Slower than local bridging b. Hardware more expensive 4. Rarely used D. Remote bridge 1. Connects network segments at remote locations, using a wide area network (WAN) link 2. Advantage: reduces the amount of traffic passing over the WAN link Chapter 3, Lesson 2 Switching |10| 1. Switch Characteristics A. Data-link layer device B. Replaces bridges, and in many cases, routers C. Similar in appearance to a hub 4 Outline, Chapter 3 Network+ Certification, Second Edition 1. Some manufacturers make hubs and switches that are identical except for their markings. D. A switch forwards incoming traffic out to the destination port only, as opposed to a hub, which forwards traffic out through all ports. E. Converts a shared network medium to a dedicated one 1. Each packet takes a dedicated path from the source computer to the destination. 2. Each pair of computers forms its own collision domain. 3. The switch forwards broadcast messages out through all ports, but not unicasts or multicasts. 4. No computers receive messages destined for other systems. 5. No collisions occur during unicast transmissions. F. Advantages 1. Reduces network traffic congestion more efficiently than a bridge does 2. Allows each pair of computers to use the full bandwidth of the network G. Some switches can operate in full-duplex mode. 1. Computers can transmit and receive at the same time. 2. Effectively doubles the bandwidth of the network H. In general, switches are more expensive than hubs and less expensive than routers. 2. Installing Switches A. Switches are typically used to replace routers and bridges. 1. A switch generally is not needed to replace a hub on a small LAN. |12| B. Replacing the routers on a backbone internetwork with switches can dramatically reduce traffic congestion. 1. On a routed network, the backbone carries all Internet traffic. 2. On a switched network, any computer can open a dedicated connection to any other computer. 3. Collisions are dramatically decreased. C. It is not necessary to replace all of the routers on a network with switches at once. 1. You can replace either a workgroup hub or a backbone router with a switch, depending on whether the network has more intranetwork or internetwork traffic. D. Replacing routers with switches on an internetwork creates one large broadcast domain. 1. Each computer must process more unnecessary broadcasts. E. There are several ways to address the broadcast problem, including |13| 1. Virtual LANs (VLANs) a. A VLAN is a subnet on a switched network that exists only in the switches. b. You specify the addresses of computers that belong in a specific subnet. Outline, Chapter 3 5 Network+ Certification, Second Edition c. Computers in a VLAN can be located anywhere because the subnet is independent of the physical network. d. Broadcasts transmitted by a computer go only to the other computers in the VLAN. e. All communications within a VLAN are switched. f. Communications between subnets can be routed or switched. |14| 2. Layer 3 switching a. A variation on the VLAN concept that minimizes the amount of routing needed between the VLANs b. When communication between systems on different VLANs is required, a router establishes a connection between the systems and then the switches take over. c. Routing occurs only when absolutely necessary. |15| 3. Switch Types A. Cut-through switch 1. Forwards packets immediately a. Does not wait for the entire packet to arrive 2. Most cut-through switches use a technique called matrix switching or crossbar switching. a. Uses a grid of input/output (I/O) circuits that enable data to enter and leave the switch through any port 3. Advantages: a. Relatively inexpensive b. Minimizes latency (the delay that occurs when the switch processes the packets) B. Store-and-forward switch 1. Waits until the entire packet arrives before forwarding it 2. Can be either a. A shared-memory switch, which uses a common memory buffer to store incoming data from all ports, or b. A bus architecture switch, which uses individual buffers for each port, connected by a bus 3. While the data is buffered, the switch a. Performs a cyclical redundancy check (CRC) b. Checks for data-link layer protocol faults such as runts, giants, and jabber 4. Disadvantages: a. Slower b. More expensive than cut-through switching 6 Outline, Chapter 3 Network+ Certification, Second Edition Chapter 3, Lesson 3 Routing |16| 1. Router Characteristics A. Routers are hardware or software devices that connect networks to form an internetwork. 1. Are network layer devices 2. Connect LANs running different data-link layer protocols, as long as they are running the same network-layer protocol B. IP performs most of the routing on networks today. C. Transmission Control Protocol/Internet Protocol (TCP/IP) documents often use the term “gateway” when referring to a router. 1. A gateway can also refer to a hardware or software device that operates at the application layer to provide an interface between two programs. |17| 2. Router Functions A. Routers forward packets to the destination network. 1. When a computer on a LAN wants to transmit data to a computer on another LAN, the system sends its packets to a router on the local network and the router forwards them to the destination network. 2. If the destination system is not located on an adjacent network, the router must forward the packets to another router. 3. On a large internetwork, such as the Internet, packets might pass through a dozen or more routers on the way to their destination. B. Routers strip the data-link layer frame off incoming packets and add a new frame for the outgoing network interface. C. Routers fragment packets when necessary, depending on the properties of the outgoing network. 3. Packet Routing A. Routers do not forward broadcasts, except in specific circumstances. B. Routers forward packets based on the destination address in the network layer protocol header. C. Routers use routing tables to determine where to send packets. 1. Every router has an internal table called a routing table, which contains information about nearby networks. 2. When a packet is addressed to a computer on a network the router is directly connected to, the router transmits it to the destination. 3. When a packet is addressed to a computer on a distant network, the router looks up the destination in its routing table to locate an appropriate router for the next hop. 4. Routers select the most efficient path to the destination. |18| D. Routing process example 1. The computers on each segment use the router connecting the segment to the backbone as their default gateway. Outline, Chapter 3 7 Network+ Certification, Second Edition 2. The computers transmit all of the packets they generate to either a specific system on the local network or the default gateway. 3. The default gateway router strips the data-link layer frame off each packet and reads the destination address from its network layer protocol header. 4. Using the information in its routing table, the gateway determines which router it must use to access the network the destination system is located on. 5. The gateway then constructs a new frame for the packet, using the backbone’s data-link layer protocol and specifying the router leading to the destination network as the data-link layer destination address. 6. When the packet reaches the next router, the process repeats itself, except that the next router’s table indicates that the destination system is on the segment the router is attached to. a. The router can therefore construct a frame that transmits the packet directly to the destination system. |19| E. Each router that processes a packet is referred to as a hop. 1. The efficiency of a given path through the internetwork is often calculated by the number of hops required to reach the destination. F. Routers also can connect networks at remote locations by using WAN links. 1. Because only packets destined for the other network pass over the WAN, WAN traffic is minimized. 4. Routing to the Internet A. The most common use for WAN routers is connecting a network to an Internet service provider (ISP), giving the computers on the network access to the Internet. 1. To connect your LAN to the Internet, you install a router that can connect to an ISP, using any type of WAN link. 2. The router is configured to forward all traffic not destined for the local network to the ISP, which relays it to the Internet. |20| 5. Understanding Routing Tables A. Routing tables are essential to the router operation and functions. 1. Unlike bridges and switches, routers cannot compile their own routing tables from the incoming packets. B. Static routing 1. Manual creation of a routing table by an administrator 2. Practical on small networks only a. On a large network, static routing would require too much work. 3. Cannot accommodate changes in network status C. Dynamic routing 1. Automatic creation of routing table entries by using a specialized routing protocol a. Routers use the specialized protocols to exchange routing table information. 8 Outline, Chapter 3 Network+ Certification, Second Edition b. Routers have direct information only about the networks they are connected to. c. Sharing information with other routers enables all of the routers to build a picture of the network. 2. Routing protocols enable routers to a. Build their tables automatically, with no participation by an administrator b. Update their tables when the network changes, such as when a router malfunctions 6. Routing Metrics A. Networks often have more than one route to a particular destination. B. Routers use a metric to evaluate the efficiency of each route. C. Routers choose the most efficient route available by selecting the one with the lowest metric. 1. The metric can be based on the number of hops or on more complex computations. 2. The lowest metric is best. |21| 7. Router Types A. Routers can be stand-alone hardware devices or software running on a computer. B. Hardware routers 1. Large enterprise networks use expensive rack-mounted routers to connect LANs into an internetwork. 2. Lower-priced router devices enable you to connect a small network to the Internet. C. Software routers include 1. Microsoft Windows 2000 Internet Connection Sharing (ICS) 2. Windows 2000 Routing and Remote Access |22| Chapter Summary A. Bridging 1. Bridges selectively relay packets between network segments, depending on their data-link layer destination addresses. 2. Bridges maintain a single broadcast domain and create separate collision domains. B. Switching 1. Switches improve on the function of bridges by forwarding packets only to their destination systems. 2. Switches reduce the collisions on a network and increase the bandwidth available to each computer. C. Routing 1. Routers are used to connect networks together at the network layer of the OSI reference model. Outline, Chapter 3 9 Network+ Certification, Second Edition 2. Routers strip away the data-link layer frame of incoming packets and build a new frame using the data-link layer protocol of the outgoing network. 3. Routing tables can be created manually by a network administrator or automatically by a routing protocol.