Chapter 6. The Network Layer

Transcription

1 Chapter 6 The Network Layer 1

2 Network Layer Design Isues Store-and-Forward Packet Switching Services Provided to the Transport Layer Implementation of Connectionless Service Implementation of Connection-Oriented Service Comparison of Virtual-Circuit and Datagram Subnets 2

3 Store-and-Forward Packet Switching fig 5-1 The environment of the network layer protocols. 3

4 4

5 5

6 Implementation of Connectionless Service Routing within a diagram subnet. 6

7 Implementation of Connection-Oriented Service Routing within a virtual-circuit subnet. 7

8 8

9 Comparison of Virtual-Circuit and Datagram Subnets 5-4 9

10 Routing algorithms can be grouped into two classes: nonadaptive and adaptive 10

11 Routing Algorithms The Optimality Principle Shortest Path Routing Flooding Distance Vector Routing Link State Routing Hierarchical Routing Broadcast Routing Multicast Routing Routing for Mobile Hosts 11

12 Routing Algorithms (2) Conflict between fairness and optimality. 12

13 13

14 The Optimality Principle (a) A subnet. (b) A sink tree for router B. 14

15 15

16 Shortest Path Routing The first 5 steps used in computing the shortest path from A to D. The arrows indicate the working node. 16

17 Flooding Flooding Algorithm: Every incoming packet is sent out on every outgoing line except the one its arrived on Methods for damming the flood (a) hop count (b) keeping track of which packets have been flooded (list of sequence numbers) Modified flooding algorithm: Selective flooding Application of the flooding algorithm (a) military (b) distribution of database (c) wireless networks (d) bench mark for other algorithm (shortest path) 17

18 Distance Vector Routing 1. Each router maintains a routing table, whose entry contains two parts: the preferred outgoing line to use for that destination and an estimate of the time or distance to that destination 2. The router is assumed to know the distance to each of its neighbors 3. Each router sends the routing table (destination, distance) to each neighbor ihb 4B 4. Based on the collected routing tables from the neighbors, ihb the router computes the shortest distance for each destination and update the routing table 18

19 Distance Vector Routing (a) A subnet. (b) Input from A, I, H, K, and the new routing table for J. 19

20 Distance Vector Routing g( (2) Good news propagates fast Routers may take infinite number of exchanges to know bad news (a router was down) The count-to-infinity problem. 20

21 Link State Routing Each router must do the following: 1. Discover its neighbors, learn their network addresses. 2. Measure the delay or cost to each of its neighbors. 3. Construct a packet telling all it has just learned. 4. Send this packet to all other routers. 5. Compute the shortest path to every other router. 21

22 Learning about the Neighbors The LAN is modeled as a node (artificial) (a) Nine routers and a LAN. (b) A graph model of (a). 22

23 Measuring Line Cost A special ECHO packet is sent to the other side to measure the delay (with / without taking the load into account) A subnet in which the East and West parts are connected by two lines. 23

24 24

25 25

26 26

27 27

28

29 Distributing the Link State Packets Flooding is used to distribute the link state packet Over a router has accumulates a full set of link state packet, it can construct the entire subnet graph and compute the shortest path. The packet buffer for router B in the previous slide (Fig. 6-13). 29

30 Hierarchical Routing Hierarchical routing. 30

31 31

32

33 Multicast Routing (a) Anetwork network. (b) A spanning tree for the leftmost router. (c) A multicast tree for group 1. (d) A multicast tree for group 2. 33

34 Broadcast Routing Reverse path forwarding. (a) A subnet. (b) a Sink tree. (c) The tree built by reverse path forwarding. 34

35

36 36

37 37

38 Routing for Mobile Hosts (2) Packet routing for mobile users. 38

39 Routing in Ad Hoc Networks Possibilities i when the routers are mobile: 1. Military vehicles on battlefield. No infrastructure. 2. A fleet of ships at sea. All moving all the time 3. Emergency works at earthquake. The infrastructure destroyed. 4. A gathering of people with notebook computers. In an area lacking

40 Route Discovery a) (a) Range of A's broadcast. b) (b) After B and D have received A's broadcast. c) (c) After C, F, and G have received A's broadcast. d) (d) After E, H, and I have received A's broadcast. Shaded nodes are new recipients. Arrows show possible reverse routes. 40

41 Route Discovery y( (2) Format of a ROUTE REQUEST packet. 41

42 Route Discovery y( (3) Format of a ROUTE REPLY packet. 42

43 Route Maintenance (a) D's Ds routing table before G goes down. (b) The graph after G has gone down. 43

44 Node Lookup in Peer-to-Peer Networks ()A (a) set of 32 node identifiers arranged in a circle. The shaded d ones correspond to actual machines. The arcs show the fingers from nodes 1, 4, and 12. The labels on the arcs are the table indices. (b) Examples of the finger tables. 44

45 Congestion Control Algorithms General Principles of Congestion Control Congestion Prevention Policies Congestion Control in Virtual-Circuit Subnets Congestion Control lin Datagram Subnets Load Shedding Jitter Control 45

46 Congestion When too much traffic is offered, congestion sets in and performance degrades sharply. 46

47 General Principles of Congestion Control 1. Monitor the system. detect when and where congestion occurs. 2. Pass information to where action can be taken. 3. Adjust system operation to correct the problem. 47

48 Congestion Prevention Policies (timeout interval) 5-26 (input /output queue) (timeout interval) (selective repeat) (response time) Policies that affect congestion. 48

49 Congestion Control in Virtual-Circuit Subnets a. Admission control (no more new VC will be setup) b. New VCs are allowed only to bypass the congestion areas c. Resources (bandwidth, buffer,..) reservation (a) A congested subnet. (b) A redrawn subnet, eliminates congestion and a virtual circuit from A to B. 49

50 Congestion control in datagram subnets a. Each router estimates the availability of each output line. unew = auold + ( 1 a) a: updating period f: utilization of the line f b. Whenever u moves above the threshold, h some action will be taken c. The warning bit method d. Choke packets (to source) e. hop-by-hop choke packets 50

51 Hop-by-Hop Choke Packets (a) A choke packet that affects only the source. (b) A choke packet that affects each hop it passes through. h 51

52 Load shedding (discarding packets) a. wine policy (minimize i i retransmission) i b. milk policy (for multimedia, new is better) c. priority (discarding low priority packets) d. Random Early Detection (for wired subnet, the router discards the packets randomly, and the source should response lost packets by slowing down the transmission rate) 52

53 Jitter Control Jitter control is important for audio and video applications a. Routers adjust the transmission order of output packets b. For video on demand packets can be buffered at the receiver in advance (a) High jitter. (b) Low jitter. 53

54 Quality of Service Requirements Techniques for Achieving Good Quality of Service Integrated Services Differentiated Services Label Switching and MPLS 54

55 Requirements 5-30 How stringent the quality-of-service requirements are. 55

56 ATM networks classify flows into four categories: 1. Constant bit rate (telephony) 2. Real-time variable bit rate ( compressed video) 3. Non-real-time variable bit rate (movie over Internet) 4. Available bit rate (file transfer) 56

57 Techniques for Achieving Good QOS 1. Over provisioning (over supply) 2. Buffering 3. Traffic shapping (agreement on certain traffic pattern) 4. The leaky bucket algorithm 5. The token bucket algorithm 6. Resource reservation 7. Admission control 8. Proportional routing 9. Packet scheduling 57

58 Buffering Smoothing the output stream by buffering packets. 58

59 59

60 60

61 61

62 The Leaky Bucket Algorithm (a) Input to a leaky bucket. (b) Output from a leaky bucket. Output from a token bucket with capacities of (c) 250 KB, (d) () 500 KB, (e) () 750 KB, (f) Output from a 500KB token bucket feeding a 10-MB/sec leaky bucket. 62

63 63

64 Resource reservation Three different kinds of resources can be reserved Bandwidth Buffer space CPU cycles 64

65 Because of random arrivals, and random service time distribution, queues can build up and delays can occur. Consider that packets arrives with Poisson distribution with mean packet/sec, the service time is exponential with μ packet/sec. The CPU can be modeled as an M/M/1 queue. The delay is given as ρ T = λ ρ = λ ( 1 ρ ) μ = 1 1 μ 1 ρ For example = λ = packets μ = packets / ρ = 0.95 T = 20μ sec onds / sec sec λ 65

66 Admission Control When the incoming traffic from some flow is well shaped, the capacity can be reserved in advance on the routers along the path. When a flow with flow specification (produced by the sender) is offered, the router along the router examines the flow specification and modifies the parameters as need be. Finally the acceptable flow specification (acceptable parameters) can be established. 66

67 An example of flow specification 5-34 maximum sustained rate average over a long time interval maximum capacity of the bucket maximum tolerated t transmission rate related related to protocols and processing time A router will map the parameters to a set of specific resource reservations 67

68 Proportional Routing To provide a high quality of service, the traffic can be split over multiple routes. A simple method is to divide the traffic equally or in proportion to the capacities of the outgoing links. 68

69 Packet Scheduling 1. Fair queueing algorithm The router has separate queues for each output line, one for each flow. When a line becomes idle, the router scans the queues round robins, taking the first packet on the nest queue. 2. Weighted fair queueing algorithm * for priority service (e.g. giving some servers more bandwidth) (a) A router with five packets queued for line O. (b) Finishing times for the five packets. 69

70 Integrated Services It was aimed at both unicast (streaming video) and multicast (digital television broadcasting) applications In many multicast applications groups can change membership dynamically. Under these conditions, the approach of having the senders reserve bandwidth in advance dose not work well. If the number of receivers are large, the dynamical grouping dose not work at all. 70

71 Integrated Services (flow-based algorithms) RSVP-The ReSerVation Protocol (a) Anetwork network, (b) The multicast spanning tree for host 1. (c) The multicast spanning tree for host 2. 71

72 RSVP-The ReSerVation Protocol (2) Already reserved When Host5 requests, new reservation is needed d (a) Host 3 requests a channel to host 1. (b) Host 3 then requests a second channel, to host 2. (c) Host 5 requests a channel to host 1. 72

73 Differentiated services (class-based) Differentiated Service (DS) can be offered by a set of routers forming an administrative domain (e.g. an ISP or a Telco). The administration defined a set of service classes with corresponding forwarding rules. Some classes (e.g. premium service) with higher requirements may has better service than others. DS scheme does not require advance setup, resource reservation or time consuming end-to-end negotiation for each flow. This makes DS relatively easy to implement. For example, the operator reserves resources for all IP-phone users (the IP-phone class), on contrary, a flow-based scheme, each phone call gets its own resources and 73 guarantees.

74 Differentiated Services (class-based service) Expedited Forwarding Two queues for each output line, one for expedited packets and one for regular packets Expedited packets experience a traffic-free network. 74

75 Assured Forwarding The assured forwarding scheme which manages the service classes, specifies three packet discard d probabilities bili i and four priority classes, each class having its own resources. Taken together, three two factors define 12 service classes. Implementation procedure 1. Classify the packets into one of the four classes (on the sending host or the ingress router) 2. Mark the packets (type A service in IP, which is specified in the type of service field) 3. Pass the packets through a shaper/dropper filter (e.g. leaky buckets or token buckets) 75

76 Assured Forwarding A possible implementation of the data flow for assured forwarding. 76

77 Label Switching and MPLS This scheme adds a label in front each packet and does the routing based on the label rather than on the destination address. Making the label an index into an internal table makes finding the correct output line becomes just a matter of table lookup. (X, 25, ATM, frame relay and virtual circuits i did the similar il thing) 77

78 Label Switching and MPLS A new MPLS header is added in front of the IP header. On a router-to-router line using PPP as the framing protocol MPLS is layer 2.5 s: stacking multiple labels in hierarchical networks TTL: Time To Live Transmitting a TCP segment using IP, MPLS, and PPP. 78

79 1. MPLS headers are not part of the network layer packet or the data link layer frame 2. It is possible to build MPLS switches that can forward both IP packets and ATM cells 3. In the MPLS router, the label is used as an index into a table to determine the outgoing line to use and the new label to be used (label swapping) 4. The router may group multiple flows that end at a particular router (LAN) and use a single label l (Forwarding Equivalence Class, FEC) 5. No setup is needed for MPLS routing 79

80 6. Two ways to create the forwarding table. Dt Data-driven: di The hit router asks the router downstream to generate a label for the flow (for ATM subnet) Control-driven: di The router checks to see for which hihrouters it is the final destination. It then creates one or more FECs for them, allocates a label for each one, and passes the labels in their forwarding table and send new ones to their neighbors. 7. A packet may carry an entire stack of labels. The S bit is set to 1 for the bottom label and 0 for all the other labels. 80

81 Internetworking How Networks Differ How Networks Can Be Connected Concatenated Virtual Circuits Connectionless Internetworking Tunneling Internetwork Routing Fragmentation 81

82 82

83 Connecting Networks A collection of interconnected networks. 83

84 How Networks Differ 5-43 Some of the many ways networks can differ. 84

85 85

86 How Networks Can Be Connected (a) Two Ethernets connected by a switch. (b) Two Ethernets connected by routers. 86

87 87

88 88

89 89

90 90

91 an Autonomous System (AS) Each network in an internetwork is independent and is often referred to as 91

92 92

93 Fragmentation (a) Transparent fragmentation. (b) Nontransparent fragmentation. 93

94 Fragmentation (2) Fragmentation when the elementary data size is 1 byte. (a) Original packet, containing 10 data bytes. (b) Fragments after passing through a network with maximum packet size of 8 payload bytes plus header. (c) Fragments after passing through a size 5 gateway. 94

95 The Network Layer in the Internet The IP Protocol IP Addresses Internet Control Protocols OSPF The Interior Gateway Routing Protocol BGP The Exterior Gateway Routing Protocol Internet Multicasting Mobile IP IPv6 95

96 96

97 Design Principles for Internet 1. Make sure it works. 2. Keep it simple. 3. Make clear choices. 4. Exploit modularity. 5. Expect heterogeneity. 6. Avoid static options and parameters. (negotiation is preferred) 7. Look for a good design; it need not be perfect. 8. Be strict when sending and tolerant when receiving. 9. Think about scalability. 10. Consider performance and cost. 97

98 Collection of Subnetworks The Internet is an interconnected collection of many networks. 98

99 Fragment offset: Tell where in the current datagram this fragment belongs (A 13 maximum of 2 =8192 fragments per datagram) 99 Time to line: Packet lifetime 255 seconds

100 The IP Protocol (2) 5-54 Some of the IP options. 100

101 and Numbers) ICANN (Internet t Corporation for Assigned Names 101

102 102

103 Subnets Getting a new network address would be hard to do since network addresses are scarce. The problem is the rule that a single class A, B or C address refers to on network, not to a collection of LANs. To solve this problem is to allow a network to be split into several parts for internal use but still act like a single network to the outside world. In the Internet literature, the parts of the network (e.g. Ethernets) are called subnets. (Do not confuse with subnet used elsewhere) 103

104 Subnets A campus network consisting of LANs for various departments. 104

105 Subnets (2) To implement subnetting, the main router needs a subnet mask that indicates the split between network +subnet number and host. Subnet masks are also written in dotted decimal notation or (/number of bits). For example, the original 16-bit host number is split into 6-bit subnet number and 10-bit host number, allowing for up to 64 Ethernets, each with maximum of 1022 hosts (0 and -1 are not available) A class B network subnetted into 64 subnets. 105

106 106

107 CDR Classless InterDomain Routing The Internet is rapidly running out of IP addresses. One solution is CIDR which allocates the remaining IP addresses in variable-singed blocks. If a site needs 2000 addresses, it is given a block of 2048 addresses. For example, a set of IP addresses assignments is given as follows. A set of IP address assignments. 107

108 Dropping the classes makes forwarding more complicated. The network-based forwarding scheme is no longer applicable. A table look-up with address matching forwarding algorithm is devised. When the set of IP address is assigned to the three universities: Cambridge, Edinburgh and doxford. The router uses the mask to get the value which only matches Oxford address as The Oxford entry is used and the packet is sent along the line going to Oxford. A remote router, say Omaha, Nebraska, may use a single aggregate entry with a binary address and submask. 108

109 NAT Network Address Translation IP addresses are scarce. NAT is a quick fix. Each company is a assigned a few IP addresses for Internet traffic. Within the company, every computer gets a unique IP address for intramural traffic No packets containing these addresses may appear on the Internet. There are three reserved ranges. Most IP packets carry TCP or UPD payload, which have source port and destination port. NAT box will establish translation table based on the TCP or UPD source port and destination port. Placement and operation of a NAT box. 109

110 NAT can quickly solve the IP address shortage problem (even for ADSL, cable modem application) Here are some of the objections 1. NAT violates the architectural model of IP, i.e. every IP address identifies a single machine worldwide 2. NAT establish the translation table, which changes the Internet from connectionless to a kind of connection-oriented network 3. NAT violates the layer structure 4. NAT can only handle IP packets with TCP or UDP payload, not others 5. NAT will not work when IP addresses are inserted in the body of the text (e.g. FTP, H.323) 6. Each IP address can only map up to =61440 machines 7. Delay the implantation of IPV6 110

111 Internet Control Protocols (ICP) ICP includes ICMP, ARP, RARP, BOOTP, and DHCP Internet Control Message Protocol (I CMP) ICMP is used to test the internet and to report unexpected events Each ICMP message type is encapsulated in an IP packet. The principal ICMP message types. 111

112 112

113 When Host1 (E1) sends a packet to a user on Host2 ( ) by . The Domain Name System (DNS) will find the address corresponding to IP address

114 send all traffic for host 4 to the router, and this solution is called proxy ARP. 114

115 RARP, BOOTP, and DHCP Given an Ethernet address, what is the corresponding IP address? Reverse Address Resolution Protocol (RARP) allows a newlybooted workstation to broadcast its Ethernet address. The RARP server sees the request, looks up the file and sends back the corresponding IP address. RARP server is needed on each network, because the router will not forward the broadcost (all 1 s) packet. BOOTP uses UDP messages, however BOOTP requires manual configuration of tables mapping IP address to Ethernet address. 115

116 Dynamic Host Configuration Protocol (DHCP) DHCP allows both manual IP assignment and automatic assignment. agent Operation of DHCP. 116

117 OSPF-The Interior Gateway Routing Protocol Design principlesp 1. published algorithm 2. support variety of distance metrics, including physical distance, delay, etc. 3. dynamic, adapting to changes 4. support routing based on type of service 5. Do load balancing (load splitting) 6. support for hierarchical systems 7. security 117

118 6 118

119 119

120 120

121 121

122 122

123 123

124 124

125 125

126 126

127 127

128 128

129 129

130 130

131 131

132 132

133 Flow label will be used to allow a source and destination to setup a pseudo connection Next header tells which of the six extension headers Fig. The IPv6 fixed header (required) 133

134 IPv6 address 134

135 135

136 Extension Headers (2) To support of datagram exceeding 64KB The hop-by-hop extension header for large datagrams (jumbograms). 136

137 Extension Headers (3) The routing header lists one or more routers that must be visited The next header tells which of the six extension headers follow this one. If this header is the last IP header, then the Next header tells which transport protocol handler to pass the packet to The routing type field gives the format of the rest of the header. The segments left field keeps track of how many addresses have not yet been visited. When it hits 0, the packet is on its own (no more guidance). The extension header for routing. 137