CS Networks and Distributed Systems. Lecture 8: Inter Domain Routing

Transcription

1 CS 3700 Networks and Distributed Systems Lecture 8: Inter Domain Routing Revised 2/4/2014

2 Network Layer, Control Plane 2 Data Plane Application Presentation Session Transport Network Data Link Physical Function:! Set up routes between networks Key challenges:! Implementing provider policies! Creating stable paths RIP OSPF BGP Control Plane

3 3 Outline BGP Basics Stable Paths Problem BGP in the Real World

4 ASs, Revisited 4 AS-1 AS-3 Interior Routers AS-2 BGP Routers

5 AS Numbers 5 Each AS identified by an ASN number! 16-bit values (latest protocol supports 32-bit ones)! are reserved Currently, there are > ASNs! AT&T: 5074, 6341, 7018,! Sprint: 1239, 1240, 6211, 6242,! Northeastern: 156! North America ASs! ftp://ftp.arin.net/info/asn.txt

6 Inter-Domain Routing 6 Global connectivity is at stake!! Thus, all ASs must use the same protocol! Contrast with intra-domain routing

7 Inter-Domain Routing 6 Global connectivity is at stake!! Thus, all ASs must use the same protocol! Contrast with intra-domain routing What are the requirements?! Scalability! Flexibility in choosing routes Cost Routing around failures

8 Inter-Domain Routing 6 Global connectivity is at stake!! Thus, all ASs must use the same protocol! Contrast with intra-domain routing What are the requirements?! Scalability! Flexibility in choosing routes Cost Routing around failures Question: link state or distance vector?

9 Inter-Domain Routing 6 Global connectivity is at stake!! Thus, all ASs must use the same protocol! Contrast with intra-domain routing What are the requirements?! Scalability! Flexibility in choosing routes Cost Routing around failures Question: link state or distance vector?! Trick question: BGP is a path vector protocol

10 BGP 7 Border Gateway Protocol! De facto inter-domain protocol of the Internet! Policy based routing protocol! Uses a Bellman-Ford path vector protocol

11 BGP 7 Border Gateway Protocol! De facto inter-domain protocol of the Internet! Policy based routing protocol! Uses a Bellman-Ford path vector protocol Relatively simple protocol, but! Complex, manual configuration

12 BGP 7 Border Gateway Protocol! De facto inter-domain protocol of the Internet! Policy based routing protocol! Uses a Bellman-Ford path vector protocol Relatively simple protocol, but! Complex, manual configuration! Entire world sees advertisements Errors can screw up traffic globally

13 BGP 7 Border Gateway Protocol! De facto inter-domain protocol of the Internet! Policy based routing protocol! Uses a Bellman-Ford path vector protocol Relatively simple protocol, but! Complex, manual configuration! Entire world sees advertisements Errors can screw up traffic globally! Policies driven by economics How much $$$ does it cost to route along a given path? Not by performance (e.g. shortest paths)

14 BGP Relationships 8

15 BGP Relationships 8 Provider Customer pays provider Customer

16 BGP Relationships 8 Provider Customer pays provider Customer

17 BGP Relationships 8

18 BGP Relationships 8 Peers do not pay each other Peer 1 Peer 2 Peer 3

19 BGP Relationships 8 Peer 1 Peer 2 Peer 3

20 BGP Relationships 8 Peer 1 Peer 2 Peer 3

21 BGP Relationships 8 Peer 2 has no incentive to route 1! 3 Peer 1 Peer 2 Peer 3

22 BGP Relationships 8 Provider Customer Customer

23 BGP Relationships 8 Provider Customer Customer

24 Tier-1 ISP Peering 9 Inteliquent Centurylink AT&T Verizon Business Level 3 Sprint XO Communications

25

26 Peering Wars 11 Peer Reduce upstream costs Improve end-to-end performance May be the only way to connect to parts of the Internet Don t Peer You would rather have customers Peers are often competitors Peering agreements require periodic renegotiation

27 Peering Wars 11 Peer Reduce upstream costs Improve end-to-end performance May be the only way to connect to parts of the Internet Don t Peer You would rather have customers Peers are often competitors Peering agreements require periodic renegotiation Peering struggles in the ISP world are extremely contentions, agreements are usually confidential

28 Two Types of BGP Neighbors 12

29 Two Types of BGP Neighbors 12 IGP Exterior routers also speak IGP

30 Two Types of BGP Neighbors 12 ebgp ebgp

31 Two Types of BGP Neighbors 12 ebgp ibgp ibgp ebgp

32 Full ibgp Meshes 13

33 Full ibgp Meshes 13 ibgp ebgp

34 Full ibgp Meshes 13 ibgp ebgp Question: why do we need ibgp?! OSPF does not include BGP policy info! Prevents routing loops within the AS

35 Full ibgp Meshes 13 ibgp ebgp Question: why do we need ibgp?! OSPF does not include BGP policy info! Prevents routing loops within the AS ibgp updates do not trigger announcements

36 Path Vector Protocol 14 AS-path: sequence of ASs a route traverses! Like distance vector, plus additional information Used for loop detection and to apply policy Default choice: route with fewest # of ASs AS /16 AS 2 AS /16 AS /16 AS /16: AS 2! AS 3! AS /16: AS 2! AS /16: AS 2! AS 5

37 BGP Operations (Simplified) 15 Establish session on TCP port 179 AS-1 Exchange active routes Exchange incremental updates BGP Session AS-2

38 Four Types of BGP Messages 16 Open: Establish a peering session. Keep Alive: Handshake at regular intervals. Notification: Shuts down a peering session. Update: Announce new routes or withdraw previously announced routes.

39 Four Types of BGP Messages 16 Open: Establish a peering session. Keep Alive: Handshake at regular intervals. Notification: Shuts down a peering session. Update: Announce new routes or withdraw previously announced routes. announcement = IP prefix + attributes values

40 BGP Attributes 17 Attributes used to select best path! LocalPref Local preference policy to choose most preferred route Overrides default fewest AS behavior

41 BGP Attributes 17 Attributes used to select best path! LocalPref Local preference policy to choose most preferred route Overrides default fewest AS behavior! Multi-exit Discriminator (MED) Specifies path for external traffic destined for an internal network Chooses peering point for your network

42 BGP Attributes 17 Attributes used to select best path! LocalPref Local preference policy to choose most preferred route Overrides default fewest AS behavior! Multi-exit Discriminator (MED) Specifies path for external traffic destined for an internal network Chooses peering point for your network! Import Rules What route advertisements do I accept?! Export Rules Which routes do I forward to whom?

43 Route Selection Summary 18 18

44 Route Selection Summary Highest Local Preference Enforce relationships

45 Route Selection Summary Highest Local Preference Enforce relationships Shortest AS Path Lowest MED Lowest IGP Cost to BGP Egress Traffic engineering

46 Route Selection Summary Highest Local Preference Enforce relationships Shortest AS Path Lowest MED Lowest IGP Cost to BGP Egress Lowest Router ID Traffic engineering When all else fails, break ties

47 Shortest AS Path!= Shortest Path 19 Source Destination

48 Shortest AS Path!= Shortest Path 19 Source?? Destination

49 Shortest AS Path!= Shortest Path 19 4 hops 4 ASs? Source? Destination

50 Shortest AS Path!= Shortest Path 19 4 hops 4 ASs? Source? 9 hops 2 ASs Destination

51 Shortest AS Path!= Shortest Path 19 4 hops 4 ASs? Source? 9 hops 2 ASs Destination

52 Hot Potato Routing 20 Source Destination

53 Hot Potato Routing 20 Source?? Destination

54 Hot Potato Routing 20? Source? 3 hops total, 3 hops cost Destination

55 Hot Potato Routing 20 5 hops total, 2 hops cost? Source? 3 hops total, 3 hops cost Destination

56 Hot Potato Routing 20 5 hops total, 2 hops cost? Source? 3 hops total, 3 hops cost Destination

57 Importing Routes 21

58 Importing Routes 21 ISP Routes

59 Importing Routes 21 ISP Routes From Customer

60 Importing Routes 21 ISP Routes From Peer From Peer From Customer

61 Importing Routes 21 From Provider ISP Routes From Peer From Peer From Customer

62 Exporting Routes 22

63 Exporting Routes 22 To Customer Customers get all routes

64 Exporting Routes 22 Customer and ISP routes only To Peer To Peer To Customer Customers get all routes

65 Exporting Routes 22 To Provider Customer and ISP routes only To Peer To Peer To Customer Customers get all routes

66 Exporting Routes 22 $$$ generating routes To Provider Customer and ISP routes only To Peer To Peer To Customer Customers get all routes

67 Modeling BGP 23 AS relationships! Customer/provider! Peer! Sibling, IXP Gao-Rexford model! AS prefers to use customer path, then peer, then provider Follow the money!! Valley-free routing! Hierarchical view of routing (incorrect but frequently used)

68 Modeling BGP 23 AS relationships! Customer/provider! Peer! Sibling, IXP Gao-Rexford model! AS prefers to use customer path, then peer, then provider Follow the money!! Valley-free routing! Hierarchical view of routing (incorrect but frequently used) C-P P-P

69 Modeling BGP 23 AS relationships! Customer/provider! Peer! Sibling, IXP Gao-Rexford model! AS prefers to use customer path, then peer, then provider Follow the money!! Valley-free routing! Hierarchical view of routing (incorrect but frequently used) C-P P-P

70 Modeling BGP 23 AS relationships! Customer/provider! Peer! Sibling, IXP Gao-Rexford model! AS prefers to use customer path, then peer, then provider Follow the money!! Valley-free routing! Hierarchical view of routing (incorrect but frequently used) C-P P-P P-P P-C

71 Modeling BGP 23 AS relationships! Customer/provider! Peer! Sibling, IXP Gao-Rexford model! AS prefers to use customer path, then peer, then provider Follow the money!! Valley-free routing! Hierarchical view of routing (incorrect but frequently used) C-P P-P P-P P-C

72 Modeling BGP 23 AS relationships! Customer/provider! Peer! Sibling, IXP Gao-Rexford model! AS prefers to use customer path, then peer, then provider Follow the money!! Valley-free routing! Hierarchical view of routing (incorrect but frequently used) P-P P-P C-P P-C P-C P-P

73 Modeling BGP 23 AS relationships! Customer/provider! Peer! Sibling, IXP Gao-Rexford model! AS prefers to use customer path, then peer, then provider Follow the money!! Valley-free routing! Hierarchical view of routing (incorrect but frequently used) P-P P-P C-P P-C P-C P-P

74 Modeling BGP 23 AS relationships! Customer/provider! Peer! Sibling, IXP Gao-Rexford model! AS prefers to use customer path, then peer, then provider Follow the money!! Valley-free routing! Hierarchical view of routing (incorrect but frequently used) P-P P-P C-P P-C P-C P-P

75 AS Relationships: It s Complicated 24 GR Model is strictly hierarchical! Each AS pair has exactly one relationship! Each relationship is the same for all prefixes

76 AS Relationships: It s Complicated 24 GR Model is strictly hierarchical! Each AS pair has exactly one relationship! Each relationship is the same for all prefixes In practice it s much more complicated! Rise of widespread peering! Regional, per-prefix peerings! Tier-1 s being shoved out by hypergiants! IXPs dominating traffic volume

77 AS Relationships: It s Complicated 24 GR Model is strictly hierarchical! Each AS pair has exactly one relationship! Each relationship is the same for all prefixes In practice it s much more complicated! Rise of widespread peering! Regional, per-prefix peerings! Tier-1 s being shoved out by hypergiants! IXPs dominating traffic volume Modeling is very hard, very prone to error! Huge potential impact for understanding Internet behavior

78 Other BGP Attributes 25 AS_SET! Instead of a single AS appearing at a slot, it s a set of Ases! Why?

79 Other BGP Attributes 25 AS_SET! Instead of a single AS appearing at a slot, it s a set of Ases! Why? Communities! Arbitrary number that is used by neighbors for routing decisions Export this route only in Europe Do not export to your peers! Usually stripped after first interdomain hop! Why?

80 Other BGP Attributes 25 AS_SET! Instead of a single AS appearing at a slot, it s a set of Ases! Why? Communities! Arbitrary number that is used by neighbors for routing decisions Export this route only in Europe Do not export to your peers! Usually stripped after first interdomain hop! Why? Prepending! Lengthening the route by adding multiple instances of ASN! Why?

81 26 Outline BGP Basics Stable Paths Problem BGP in the Real World

82 What Problem is BGP Solving? Underlying Problem Distributed Solution Shortest Paths RIP, OSPF, IS-IS, etc.??? BGP

83 What Problem is BGP Solving? Underlying Problem Distributed Solution Shortest Paths RIP, OSPF, IS-IS, etc.??? BGP Knowing??? can:! Aid in the analysis of BGP policy! Aid in the design of BGP extensions! Help explain BGP routing anomalies! Give us a deeper understanding of the protocol

84 The Stable Paths Problem 28 An instance of the SPP:! Graph of nodes and edges! Node 0, called the origin

85 The Stable Paths Problem 28 An instance of the SPP:! Graph of nodes and edges! Node 0, called the origin! A set of permitted paths from each node to the origin Each set contains the null path

86 The Stable Paths Problem 28 An instance of the SPP:! Graph of nodes and edges! Node 0, called the origin! A set of permitted paths from each node to the origin Each set contains the null path! Each set of paths is ranked Null path is always least preferred

87 A Solution to the SPP 29 A solution is an assignment of permitted paths to each node such that:! Node u s path is either null or uwp, where path uw is assigned to node w and edge u! w exists! Each node is assigned the higest ranked path that is consistent with their neighbors

88 A Solution to the SPP 29 A solution is an assignment of permitted paths to each node such that:! Node u s path is either null or uwp, where path uw is assigned to node w and edge u! w exists! Each node is assigned the higest ranked path that is consistent with their neighbors

89 A Solution to the SPP 29 A solution is an assignment of permitted paths to each node such that:! Node u s path is either null or uwp, where path uw is assigned to node w and edge u! w exists! Each node is assigned the higest ranked path that is consistent with their neighbors

90 A Solution to the SPP 29 A solution is an assignment of permitted paths to each node such that:! Node u s path is either null or uwp, where path uw is assigned to node w and edge u! w exists! Each node is assigned the higest ranked path that is consistent with their neighbors

91 A Solution to the SPP 29 A solution is an assignment of permitted paths to each node such that:! Node u s path is either null or uwp, where path uw is assigned to node w and edge u! w exists! Each node is assigned the higest ranked path that is consistent with their neighbors

92 A Solution to the SPP 29 A solution is an assignment of permitted paths to each node such that:! Solutions Node u s path need is either not use null or uwp, where path uw is assigned to the node shortest w and edge paths, u! or w exists form! Each node a spanning is assigned tree the higest ranked path that is consistent with their neighbors

93 Simple SPP Example

94 Simple SPP Example

95 Simple SPP Example

96 Simple SPP Example

97 Simple SPP Example

98 Simple SPP Example

99 Simple SPP Example Each node gets its preferred route 0 Totally stable topology

100 Good Gadget

101 Good Gadget

102 Good Gadget

103 Good Gadget

104 Good Gadget

105 Good Gadget Not every node gets preferred route Topology is still stable Only one stable configuration 0 No matter which router chooses first!

106 SPP May Have Multiple Solutions

107 SPP May Have Multiple Solutions

108 SPP May Have Multiple Solutions

109 SPP May Have Multiple Solutions

110 SPP May Have Multiple Solutions

111 SPP May Have Multiple Solutions

112 SPP May Have Multiple Solutions

113 Bad Gadget

114 Bad Gadget

115 Bad Gadget

116 Bad Gadget

117 Bad Gadget

118 Bad Gadget

119 Bad Gadget

120 Bad Gadget

121 Bad Gadget

122 Bad Gadget

123 Bad Gadget

124 Bad Gadget

125 Bad Gadget

126 Bad Gadget That was only one round of oscillation! This keeps going, infinitely Problem stems from: Local (not global) decisions Ability of one node to improve its path selection

127 SPP Explains BGP Divergence 34 BGP is not guaranteed to converge to stable routing! Policy inconsistencies may lead to livelock! Protocol oscillation Solvable Can Diverge Must Converge Must Diverge

128 SPP Explains BGP Divergence 34 BGP is not guaranteed to converge to stable routing! Policy inconsistencies may lead to livelock! Protocol oscillation Good Gadgets Solvable Can Diverge Bad Gadgets Must Converge Must Diverge Naughty Gadgets

129 Beware of Backup Policies

130 Beware of Backup Policies

131 Beware of Backup Policies BGP is not robust 0 It may not recover from link failure

132 BGP is Precarious

133 BGP is Precarious If node 1 uses path 1! 0, this is solvable

134 BGP is Precarious No longer stable

135 Can BGP Be Fixed? 37 Unfortunately, SPP is NP-complete

136 Can BGP Be Fixed? 37 Unfortunately, SPP is NP-complete Possible Solutions Dynamic Approach Extend BGP to detect and suppress policy-based oscillations?

137 Can BGP Be Fixed? 37 Unfortunately, SPP is NP-complete Possible Solutions Static Approach Dynamic Approach Automated Analysis of Routing Policies (This is very hard) Inter-AS coordination Extend BGP to detect and suppress policy-based oscillations?

138 Can BGP Be Fixed? 37 Unfortunately, SPP is NP-complete Possible Solutions Static Approach Dynamic Approach Automated Analysis of Routing Policies (This is very hard) Inter-AS coordination Extend BGP to detect and suppress policy-based oscillations? These approaches are complementary

139 38 Outline BGP Basics Stable Paths Problem BGP in the Real World

140 Motivation 39 Routing reliability/fault-tolerance on small time scales (minutes) not previously a priority Transaction oriented and interactive applications (e.g. Internet Telephony) will require higher levels of end-toend network reliability How well does the Internet routing infrastructure tolerate faults?

141 Conventional Wisdom 40 Internet routing is robust under faults! Supports path re-routing! Path restoration on the order of seconds BGP has good convergence properties! Does not exhibit looping/bouncing problems of RIP Internet fail-over will improve with faster routers and faster links More redundant connections (multi-homing) will always improve fault-tolerance

142 Delayed Routing Convergence 41 Conventional wisdom about routing convergence is not accurate! Measurement of BGP convergence in the Internet! Analysis/intuition behind delayed BGP routing convergence! Modifications to BGP implementations which would improve convergence times

143 Open Question 42 After a fault in a path to multi-homed site, how long does it take for majority of Internet routers to fail-over to secondary path? Customer Primary ISP Backup ISP

144 Open Question 42 After a fault in a path to multi-homed site, how long does it take for majority of Internet routers to fail-over to secondary path? Route Withdrawn Routing table convergence Primary ISP Customer Backup ISP

145 Open Question 42 After a fault in a path to multi-homed site, how long does it take for majority of Internet routers to fail-over to secondary path? Customer Primary ISP Route Withdrawn Routing table convergence Stable end-to-end paths Backup ISP Traffic

146 Bad News 43 With unconstrained policies:! Divergence! Possible create unsatisfiable policies! NP-complete to identify these policies! Happening today?

147 Bad News 43 With unconstrained policies:! Divergence! Possible create unsatisfiable policies! NP-complete to identify these policies! Happening today? With constrained policies (e.g. shortest path first)! Transient oscillations! BGP usually converges! It may take a very long time

148 Bad News 43 With unconstrained policies:! Divergence! Possible create unsatisfiable policies! NP-complete to identify these policies! Happening today? With constrained policies (e.g. shortest path first)! Transient oscillations! BGP usually converges! It may take a very long time BGP Beacons: focuses on constrained policies

149 16 Month Study of Convergence 44 Instrument the Internet! Inject BGP faults (announcements/withdrawals) of varied prefix and AS path length into topologically and geographically diverse ISP peering sessions! Monitor impact faults through Recording BGP peering sessions with 20 tier1/tier2 ISPs Active ICMP measurements (512 byte/second to 100 random web sites)! Wait two years (and 250,000 faults)

150 Measurement Architecture 45

151 Measurement Architecture 45 Researchers pretending to be an AS Researchers pretending to be an AS

152 Measurement Architecture 45

153 Measurement Architecture 45

154 Announcement Scenarios 46 Tup a new route is advertised Tdown A route is withdrawn! i.e. single-homed failure Tshort Advertise a shorter/better AS path! i.e. primary path repaired Tlong Advertise a longer/worse AS path! i.e. primary path fails

155 Major Convergence Results 47 Routing convergence requires an order of magnitude longer than expected! 10s of minutes Routes converge more quickly following Tup/Repair than Tdown/Failure events! Bad news travels more slowly Withdrawals (Tdown) generate several more announcements than new routes (Tup)

156 Example 48 BGP log of updates from AS2117 for route via AS2129 One withdrawal triggers 6 announcements and one withdrawal from 2117 Increasing AS path length until final withdrawal

157 Why So Many Announcements? 49 Events from AS 2177 AS 2041 AS 3508 AS 1 AS 5696 AS 2129 AS 2117

158 Why So Many Announcements? 49 Events from AS 2177 AS 2041 AS 3508 AS 1 AS 5696 AS 2129 AS 2117

159 Why So Many Announcements? 49 Events from AS Route Fails: AS 2129 AS 2041 AS 3508 AS 1 AS 5696 AS 2129 AS 2117

160 Why So Many Announcements? 49 Events from AS Route Fails: AS Announce: AS 2041 AS 3508 AS 1 AS 5696 AS 2129 AS 2117

161 Why So Many Announcements? 49 Events from AS Route Fails: AS Announce: Announce: AS 2041 AS 3508 AS 1 AS 5696 AS 2129 AS 2117

162 Why So Many Announcements? 49 Events from AS Route Fails: AS Announce: Announce: Announce: AS 2041 AS 3508 AS 1 AS 5696 AS 2129 AS 2117

163 Why So Many Announcements? 49 Events from AS Route Fails: AS Announce: Announce: Announce: Announce: Route Withdrawn: 2129 AS 2041 AS 3508 AS 1 AS 5696 AS 2129 AS 2117

164 How Many Announcements Does it Take For an AS to Withdraw a Route? 50

165 How Many Announcements Does it Take For an AS to Withdraw a Route? 50 Answer: up to 19

166 BGP Routing Table Convergence Times 100 Cumulative Percentage of Events New Route Long->Short Fail-over Short->Long Fail-Over Failure Tup Tshort Tlong Tdown Seconds Until Convergence Less than half of Tdown events converge within two minutes Tup/Tshort and Tdown/Tlong form equivalence classes Long tailed distribution (up to 15 minutes)

167 Failures, Fail-overs and Repairs 52 Bad news does not travel fast Repairs (Tup) exhibit similar convergence as long-short AS path fail-over Failures (Tdown) and short-long fail-overs (e.g. primary to secondary path) also similar! Slower than Tup (e.g. a repair)! 80% take longer than two minutes! Fail-over times degrade the greater the degree of multihoming

168 Intuition for Delayed Convergence 53 There exists possible ordering of messages such that BGP will explore ALL possible AS paths of ALL possible lengths BGP is O(N!), where N number of default-free BGP routers in a complete graph with default policy

169 Impact of Delayed Convergence 54 Why do we care about routing table convergence?! It impacts end-to-end connectivity for Internet paths ICMP experiment results! Loss of connectivity, packet loss, latency, and packet reordering for an average of 3-5 minutes after a fault Why?! Routers drop packets when next hop is unknown! Path switching spikes latency/delay! Multi-pathing causes reordering

170 In real life 55 Discussed worst case BGP behavior In practice, BGP policy prevents worst case from happening BGP timers also provide synchronization and limits possible orderings of messages

171 Inter-Domain Routing Summary 56 BGP4 is the only inter-domain routing protocol currently in use world-wide Issues?! Lack of security! Ease of misconfiguration! Poorly understood interaction between local policies! Poor convergence! Lack of appropriate information hiding! Non-determinism! Poor overload behavior

172 Lots of research into how to fix this 57 Security! BGPSEC, RPKI Misconfigurations, inflexible policy! SDN Policy Interactions! PoiRoot (root cause analysis) Convergence! Consensus Routing Inconsistent behavior! LIFEGUARD, among others

173 Why are these still issues? 58 Backward compatibility Buy-in / incentives for operators Stubbornness

174 Why are these still issues? 58 Backward compatibility Buy-in / incentives for operators Stubbornness Very similar issues to IPv6 deployment