DOC

ccie-3

By Alfred White,2014-01-27 01:23
9 views 0
ccie-3 ccie-3 ccie-3 ccie-3 ccie-3 ccie-3

    INTERNATIONAL NETWORK SERVICES

    CCIE Study Materials

    Module 3

    Bridging vs. Routing

    & IP Fundamentals

    Table of Contents

4.0 BRIDGING VS. ROUTING ............................................................................................................ 3

    1 BA4.SIC DEFINITIONS ......................................................................................................................... 3 4.2 MAJOR DIFFERENCES....................................................................................................................... 3 4.3 ROUTING ........................................................................................................................................ 6

    4.4 BRIDGING ADVANTAGES.................................................................................................................. 8 4.5 INTEGRATED SOLUTIONS ................................................................................................................. 8 5.0 IP FUNDAMENTALS ..................................................................................................................... 9

    5.1 DOTTED DECIMAL, BINARY AND HEX ............................................................................................... 9 5.2 IP ADDRESSING BASICS ................................................................................................................. 10 5.3 SEGMENTS AND SUBNETS .............................................................................................................. 12 5.4 NETMASKS .................................................................................................................................... 13

    5.5 DEFAULT GATEWAY ...................................................................................................................... 14 5.6 BROADCASTS ................................................................................................................................ 15

    5.7 POINT-TO-POINT NETWORKS ......................................................................................................... 17 5.8 VARIABLE LENGTH SUBNET MASKS ............................................................................................... 17 5.9 IP UNNUMBERED ........................................................................................................................... 18

    5.10 MULTICAST ADDRESSING ............................................................................................................ 18

4.0 Bridging vs. Routing

    4.1 Basic Definitions

    Bridging is defined at the data link layer and routing is defined at the network layer of the OSI reference model. When a network is bridged at the data link layer, a single logical network is the result. When routed at the network layer, two or more logical networks are interconnected. In a routed network, it may be necessary for a packet to traverse multiple hops and cross many logical networks to get from one destination to another. In a bridged network, a packet may also traverse multiple hops. In a bridged network, by definition, all packets remain within the same logical network, which might consist of many segments.

    4.2 Major Differences

    Bridges use raw MAC addresses to identify end nodes and are often used to extend LANs. Bridges are also used to filter network traffic between bridged segments. This feature is extremely useful when partitioning a group of heavy talkers from the rest of the network. Consider the following diagram:

    3

    International Network ServicesINS Confidential Materials

    Segment BB

    195.148.10.0

    RouterHost 1Host 2Host A

    Segment AA

    195.148.12.0Host 3Host 4Host B

    Figure 4.2a

    In Figure 4.2a, Hosts 1-4 are part of an extremely network intensive workgroup. Unfortunately for other hosts on the network, these four systems use the bulk of the available network bandwidth so that users on Host A and Host B cannot consistently access their remote database on Segment BB because the ethernet is too heavily utilized.

    One solution is to break Segment AA into two segments as shown in Figure 4.2b and connect them with a bridge. The bridge will learn that Hosts 1-4 only talk among themselves and that Hosts A-B do not talk to Hosts 1-4. As a result, the network traffic from Hosts 1-4 will remain on Segment A unless one of the hosts needs to talk to Host A or Host B or to non-local hosts across the router. Likewise, since Hosts A and B only talk across the router, Hosts 1-4 will not see network traffic from Hosts 1-4, except for broadcasts.

    4

    International Network ServicesINS Confidential Materials

    Segment BB

    195.148.10.0

    RouterHost 1Host 2Host A

    Bridge

    Segment ASegment AA

    195.148.12.0

    Host 3Host 4Host B

    Figure 4.2b

    Bridges do not filter broadcast traffic. In the example above, broadcast traffic from Hosts A-B are seen by Hosts 1-4 and vice versa. This characteristic of bridged networks tends to limit their scalability, particularly with network intensive protocols like NetWare 3.11 with it's numerous RIP updates and SAP broadcasts.

    Bridged LANs are said to have a "flat" network topology. In Figure 4.2b this is illustrated by the fact that the IP subnet range didn't change when Segment AA was broken up into two segments. In other words, the aggregation of MAC addresses has no topological meaning. While the network structure has changed from Figure 4.2a to Figure 4.2b, it's topology has remained the same from a routing point of view.

    One big drawback of bridged networks is that in the event of a network error condition, such as a streaming adapter, an entire bridged network can crash extremely fast and hard. In a network that is segmented using routers, a part of the network may be compromised but the overall network can operate in a business as usual manner.

    Routing is often used to limit broadcast traffic between local LAN segments or to support topologies that consist of many logical internetworks. By limiting the propagation of broadcast traffic, routing allows an internetwork to scale well beyond that of a bridged network and still be manageable.

    5

    International Network ServicesINS Confidential Materials

    Broadcast storms are the bane of bridging and often disable entire bridged networks. This is because each bridge forwards every broadcast through the entire internetwork. By using a router all broadcasts are blocked by default.

    Bridging has less protocol overhead. The header associated with each frame at the data link level has fewer fields than network layer packets. Examples of fields provided in network layer packets include fragmentation and reassembly information, source and destination addresses, hop counts, etc. The real world "cost" of this additional overhead becomes apparent when measuring the percentage of throughput decrease as a protocol is either bridged or routed across multiple device hops. Router hop degradation is generally greater than the hop degradation incurred from traversing multiple bridges. It should be noted that the router hop degradation issue is becoming less of an issue as routers approach wire speed transfer of packets between networks. 4.3 Routing

    When reachability information changes, routers can reconfigure their topologies much more quickly than bridges. Use of routers, therefore, provides higher availability of networked resources when compared to bridging. Routers also have the capability to choose the best path between a source and destination. A relatively new feature of routers is policy-based routing. This is important from a management and security point of view as certain packets can be directed onto desired paths based on cost or security or both.

    Depending to a large extent on the routing protocol that is employed, routers are subject to routing loops. Like routers, source-route bridges can use multiple paths, but spanning tree bridges are limited to a specific path through the network. In addition, loops are avoided in an SRB environment as each bridge verifies that the output network segment has not been traversed before transmitting packets. With transparent bridging, loops pose a substantially greater risk than with either a routed internetwork or a SRB network.

    The total number of connected systems that can be supported in a routed internetwork is essentially unlimited. With bridges, the maximum number of connected systems is constrained to the low thousands. Two thousand nodes on a bridged network are a lot of nodes. Routers can aggregate large numbers of individual routes into hierarchical groupings. This provides the potential to

    6

    International Network ServicesINS Confidential Materials

support an extremely large address space.

    If a packet is too large, a bridge will simply drop it. When a bridge drops a packet, it does not inform the originating node that the packet was dropped. In contrast, routers fragment and reassemble large packets. Many network layer protocols, such as IP, fragment and reassemble packets at intermediate nodes in the network. The data link layer has not been defined to perform this function. Routers are also capable of providing feedback to end stations when traffic is heavy and congestion occurs. Bridges do not perform this function.

Attribute Bridge Router

    Contains Broadcasts No Yes

    OSI Layer Data Link Network

    Protocol Independent? Yes No

    Price Per Port Low High to Medium

    Complexity Low High

    Support Skill Level Needed Low High

    Address Type MAC Network (IP, IPX, etc.)

    Logical Subnets? No Yes

    Topology Flat Can Be Hierarchical

    Scalability Fair to OK Excellent

    Availability Medium High

    Address space size In the 1000's Virtually Unlimited

    Congestion Notification? No Yes

    Large Packet Handling Drops it With No Notification Fragments and Sends It

    Potential routing loops? Transparent-Yes; SRB-No Yes in Some Cases

    Policy-Based Routing No Yes

    Device Hop Degradation Low Medium to Medium-Low

    7

    International Network ServicesINS Confidential Materials

4.4 Bridging Advantages

    The choice of whether to use bridges or routers is sometimes very difficult. In some environments, bridges are clearly the best choice over routers. In other environments, routing is the best choice. On a per port basis, routers are more expensive than bridges. Routers have more packet overhead to handle, and tend to introduce more latency than bridges in general. The newest generation of routers are quite fast, however, and are capable of providing near wire speed performance. The newest generation of bridges are really network switches and are just now beginning to mature. Network layer issues are extremely complex. As a result, routers are difficult and complex to configure. Bridges do not participate in layer three protocols. As a result, bridges are less complex and exhibit more ease of configuration than routers. In some situations, a bridge can be taken out of the box, powered up, attached to the network and it will work as advertised. Routing has developed to the degree it has in large part because of the limitations inherent in bridging. There is a place for both.

    4.5 Integrated Solutions

    As if network design using bridges and routers isn't already complex enough, a new generation of product is hitting the market. Network switches are used in a manner similar to bridges. With switches, each network host uses a dedicated port on the switch. Network designers now have greater flexibility than ever with respect to using routers, bridges or switches to solve internetworking problems.

    Routers, bridges and switches all have their place. Bridges may be a good fit for a remote site for their ease of implementation and maintenance. The connection to the WAN could use a router to provide a barrier against potential broadcast storms in the remote LAN networks. Switches can be used to alleviate traffic problems on heavily utilized networks that need to maintain a "flat" topology. The unique requirements of each individual environment are endless and provide a rich opportunity for creative network design.

    8

    International Network ServicesINS Confidential Materials

5.0 IP Fundamentals

    5.1 Dotted Decimal, Binary and Hex

    An IP address is made up of four bytes or 32 bits. There are several shorthand ways to represent this sequence of bits. Some of the more common ones are hex, dotted decimal, octal and binary. In IP shops, the most popular is dotted decimal, followed by hex. An example of a dotted decimal IP address is 167.148.3.245.

    When using hex, a 32-bit computer word is broken up into eight 4-bit sequences. The smallest value is 0000 and the largest is 1111. Since 1111 is equal to 2**0 + 2**1 + 2**2 + 2**3 for a grand total of 15 it quickly becomes apparent that a shorthand method is necessary in order to represent the integers 10-15 (as the digits 0-9 don’t change). Consider the following table:

    Decimal Hex Binary

    10 A 1010

    11 B 1011

    12 C 1100

    13 D 1101

    14 E 1110

    15 F 1111

    Example 1:

    Write “1101001010100111 0110101011101111Ó in Hex.

    Answer:

     D2A7 6AEF where (D=1101, 2=0010, etc.)

    Example 2:

    Write 167 and 148 in binary.

    Answer. First express the first eight bits in base 2:

    9

    International Network ServicesINS Confidential Materials

Two to the “x” power 2**7 2**6 2**5 2**4 2**3 2**2 2**1 2**0

    Decimal Equivalent 128 64 32 16 8 4 2 1

    Then convert to decimal:

    128 +0 +32 +0 +0 +4 +2 +1 = 167

    1 0 1 0 0 1 1 1 = 167

    128 +0 +0 +16 +0 +4 +0 +0 = 148

    1 0 0 1 0 1 0 0 = 148

    Example 3:

    Write “11010010 10100111 01101010 11101111” in dotted decimal.

    Answer: 210.167.106.239

    Note: Eight bits constitute the range of bits that are used to calculate the decimal equivalent. 5.2 IP Addressing Basics

    An IP address is written as a single value, but in reality it consists of two parts, the network ID and the host ID. There are several classes of IP address that are specified. The most common classes of IP address are Class B and Class C addresses. A Class B IP address has two octets that specify the network ID and two octets that specify the host ID as shown in the following table. Class A networks are used by organizations with a very large number of hosts. The first bit of a Class A network number is 0, leaving seven bits to be used for the network ID. This specifies a total of 126 separate networks. Twenty-four bits are used for the host ID, which allows for 17 million hosts.

    Class B addresses are intended for medium-sized networks. The first two bits of the network number for Class B addresses are 10. Fourteen bits are used for the network ID, leaving sixteen for the host ID for a total of 16,384 networks and 65,000 hosts.

    10

    International Network ServicesINS Confidential Materials

Report this document

For any questions or suggestions please email
cust-service@docsford.com