lecture08.ppt
- 1. 15-441 Computer Networking Lecture 8 – IP Packets, Routers
- 2. Lecture 8: 9-20-01 2 Outline • IP Packet Format • NAT • IPv6 • Router Internals • Route Lookup
- 3. Lecture 8: 9-20-01 3 IPv4 Header – RFC791 (1981) ver length 32 bits data (variable length, typically a TCP or UDP segment) 16-bit identifier Header checksum time to live 32 bit source IP address header length type of service flags fragment offset Protocol 32 bit destination IP address Options (if any) 0 4 16 24 32 8 19 Padding (if any)
- 4. Lecture 8: 9-20-01 4 IP Header Fields • Version 4 for IPv4 • Header length (in 32 bit words) • Minimum value is 5 (header without any options) • Length of entire IP packet in octets (including header) • Identifier, flags, fragment offset used primarily for fragmentation • Time to live • Must be decremented at each router • Packets with TTL=0 are thrown away • Ensure packets exit the network
- 5. Lecture 8: 9-20-01 5 IP Header Fields • Protocol • Demultiplexing to higher layer protocols • TCP = 6, ICMP = 1, UDP = 17… • Header checksum • Ensures some degree of header integrity • Relatively weak – 16 bit • Source/Dest address • Options • E.g. Source routing, record route, etc. • Performance issues • Poorly supported
- 6. Lecture 8: 9-20-01 6 IP Type of Service • Typically ignored • Values • 3 bits of precedence • 1 bit of delay requirements • 1 bit of throughput requirements • 1 bit of reliability requirements • Replaced by DiffServ
- 7. Lecture 8: 9-20-01 7 ICMP: Internet Control Message Protocol • Used by hosts, routers, gateways to communication network-level information • Error reporting: unreachable host, network, port, protocol • Echo request/reply (used by ping) • Network-layer “above” IP: • ICMP msgs carried in IP datagrams • ICMP message: type, code plus first 8 bytes of IP datagram causing error Type Code description 0 0 echo reply (ping) 3 0 dest. network unreachable 3 1 dest host unreachable 3 2 dest protocol unreachable 3 3 dest port unreachable 3 6 dest network unknown 3 7 dest host unknown 4 0 source quench (congestion control - not used) 8 0 echo request (ping) 9 0 route advertisement 10 0 router discovery 11 0 TTL expired 12 0 bad IP header
- 8. Lecture 8: 9-20-01 8 Fragmentation • IP packets can be up to 64KB • Different link-layers have different MTUs • Split IP packet into multiple fragments • IP header on each fragment • Intermediate router may fragment as needed
- 9. Lecture 8: 9-20-01 9 IP Fragmentation & Reassembly • Network links have MTU (max.transfer size) - largest possible link-level frame. • different link types, different MTUs • Large IP datagram divided (“fragmented”) within net • one datagram becomes several datagrams • IP header bits used to identify, order related fragments fragmentation: in: one large datagram out: 3 smaller datagrams reassembly
- 10. Lecture 8: 9-20-01 10 Reassembly • Where to do reassembly? • End nodes • Avoids unnecessary work where large packets are fragmented multiple times • Dangerous to do at intermediate nodes • How much buffer space required at routers? • What if routes in network change? • Multiple paths through network • All fragments only required to go through destination
- 11. Lecture 8: 9-20-01 11 Fragmentation Related Fields • Length • Length of IP fragment • Identification • To match up with other fragments • Flags • Don’t fragment flag • More fragments flag • Fragment offset • Where this fragment lies in entire IP datagram • Measured in 8 octet units (13 bit field)
- 12. Lecture 8: 9-20-01 12 IP Fragmentation and Reassembly ID =x offset =0 fragflag =0 length =4000 ID =x offset =0 fragflag =1 length =1500 ID =x offset =1480 fragflag =1 length =1500 ID =x offset =2960 fragflag =0 length =1040 One large datagram becomes several smaller datagrams
- 13. Lecture 8: 9-20-01 13 Fragmentation is Harmful • Uses resources poorly • Forwarding costs per packet • Best if we can send large chunks of data • Worst case: packet just bigger than MTU • Poor end-to-end performance • Loss of a fragment • Reassembly is hard • Buffering constraints
- 14. Lecture 8: 9-20-01 14 Path MTU Discovery • Hosts dynamically discover minimum MTU of path • Algorithm: • Initialize MTU to MTU for first hop • Send datagrams with Don’t Fragment bit set • If ICMP “pkt too big” msg, decrease MTU • What happens if path changes? • Periodically (>5mins, or >1min after previous increase), increase MTU • Some routers will return proper MTU • MTU values cached in routing table
- 15. Lecture 8: 9-20-01 15 Outline • IP Packet Format • NAT • IPv6 • Router Internals • Route Lookup
- 16. Lecture 8: 9-20-01 16 IP Address Utilization (‘98) • Address space depletion • In danger of running out of classes A and B • 32-bit address space completely allocated by 2008 • Two solutions • NAT • IPv6
- 17. Lecture 8: 9-20-01 17 Network Address Translation (NAT) • Possible solution to address space exhaustion • Kludge (but useful) • Sits between your network and the Internet • Translates local network layer addresses to global IP addresses • Has a pool of global IP addresses (less than number of hosts on your network) • Uses special unallocated addresses (RFC 1597) locally • 10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16
- 18. Lecture 8: 9-20-01 18 NAT Illustration Global Internet Private Network Pool of global IP addresses • Operation: Source (S) wants to talk to Destination (D): • Create Sg-Sp mapping • Replace Sp with Sg for outgoing packets • Replace Sg with Sp for incoming packets • How many hosts can have active transfers at one time? P G Dg Sp Data NAT Destination Source Dg Sg Data
- 19. Lecture 8: 9-20-01 19 Problems with NAT • What if we only have few (or just one) IP address? • Use Network Address & Port Translator (NAPT) • NAPT translates: • Translates addrprivate + flow info to addrglobal + new flow info • Uses TCP/UDP port numbers • Potentially thousands of simultaneous connections with one global IP address
- 20. Lecture 8: 9-20-01 20 Problems with NAT • Hides the internal network structure • Some consider this an advantage • Some protocols carry addresses • E.g., FTP carries addresses in text • What is the problem? • Must update transport protocol headers (port number & checksum) • Encryption • No inbound connections
- 21. Lecture 8: 9-20-01 21 Outline • IP Packet Format • NAT • IPv6 • Router Internals • Route Lookup
- 22. Lecture 8: 9-20-01 22 IPv6 • Primary objective bigger addresses • Addresses are 128bit What about header size!!! • Simplification • Header format helps speed processing/forwarding • Header changes to facilitate QoS • Removes infrequently used parts of header • 40byte fixed size vs. 20+ byte variable
- 23. Lecture 8: 9-20-01 23 IPv6 Changes • IPv6 removes checksum • Relies on upper layer protocols to provide integrity • IPv6 eliminates fragmentation • Requires path MTU discovery • Requires 1280 byte MTU
- 24. Lecture 8: 9-20-01 24 IPv6 Header Source Address Destination Address 0 4 16 24 32 Version Class Flow Label Payload Length Next Header Hop Limit 12 19
- 25. Lecture 8: 9-20-01 25 IPv6 Changes • TOS replaced with traffic class octet • Flow label • Identify datagrams in same “flow.” (concept of“flow” not well defined) • Help soft state systems • Maps well onto TCP connection or stream of UDP packets on host-port pair • Easy configuration • Provides auto-configuration using hardware MAC address to provide unique base
- 26. Lecture 8: 9-20-01 26 IPv6 Changes • Protocol field replaced by next header field • Support for protocol demultiplexing as well as option processing • Option processing • Options are added using next header field • Options header does not need to be processed by every router • Large performance improvement • Makes options practical/useful • Additional requirements • Support for security • Support for mobility
- 27. Lecture 8: 9-20-01 27 Transition From IPv4 To IPv6 • Not all routers can be upgraded simultaneous • No “flag days” • How will the network operate with mixed IPv4 and IPv6 routers? • Two proposed approaches: • Dual Stack: some routers with dual stack (v6, v4) can “translate” between formats • Tunneling: IPv6 carried as payload n IPv4 datagram among IPv4 routers
- 28. Lecture 8: 9-20-01 28 Dual Stack Approach
- 29. Lecture 8: 9-20-01 29 Tunneling IPv6 inside IPv4 where needed
- 30. Lecture 8: 9-20-01 30 Outline • IP Packet Format • NAT • IPv6 • Router Internals • Route Lookup
- 31. Lecture 8: 9-20-01 31 Router Architecture Overview Two key router functions: • Run routing algorithms/protocol (RIP, OSPF, BGP) • Switching datagrams from incoming to outgoing link
- 32. Lecture 8: 9-20-01 32 What Does a Router Look Like? • Line cards • Network interface cards • Forwarding engine • Fast path routing (hardware vs. software) • Usually on line card • Backplane • Switch or bus interconnect • Processor • Handles routing protocols, error conditions
- 33. Lecture 8: 9-20-01 33 Router Processing • Packet arrives arrives at inbound line card • Header processed by forwarding engine • Forwarding engine determines output line card/destination • Checksum updated but not checked • Packet copied to outbound line card • Odd situations sent to network processor
- 34. Lecture 8: 9-20-01 34 Network Processor • Runs routing protocol and downloads forwarding table to forwarding engines • Performs “slow” path processing • ICMP error messages • IP option processing • Fragmentation • Packets destined to router
- 35. Lecture 8: 9-20-01 35 Three Types of Switching Fabrics
- 36. Lecture 8: 9-20-01 36 Switching Via Memory First generation routers: • Packet copied by system’s (single) CPU • Speed limited by memory bandwidth (2 bus crossings per datagram) Input Port Output Port Memory System Bus Modern routers: • Input port processor performs lookup, copy into memory • Cisco Catalyst 8500
- 37. Lecture 8: 9-20-01 37 Switching Via Bus • Datagram from input port memory to output port memory via a shared bus • Bus contention: switching speed limited by bus bandwidth • 1 Gbps bus, Cisco 1900: sufficient speed for access and enterprise routers (not regional or backbone)
- 38. Lecture 8: 9-20-01 38 Switching Via An Interconnection Network • Overcome bus bandwidth limitations • Crossbar provides full NxN interconnect • Expensive • Banyan networks, other interconnection nets initially developed to connect processors in multiprocessor • Typically less capable than complete crossbar • Cisco 12000: switches Gbps through the interconnection network
- 39. Lecture 8: 9-20-01 39 Switch Design Issues • Suppose we have N inputs and M outputs • Multiple packets for same output – output contention • Switch contention – switching fabric cannot support arbitrary set of transfers • I.e, not a full crossbar • Solution – buffer packets when/where needed • What happens when these buffers fill up? • Packets are THROWN AWAY!! This is where packet loss comes from
- 40. Lecture 8: 9-20-01 40 Input Port Functions Decentralized switching: • Given datagram dest., lookup output port using routing table in input port memory • Goal: complete input port processing at ‘line speed’ • Needed if datagrams arrive faster than forwarding rate into switch fabric Physical layer: bit-level reception Data link layer: e.g., Ethernet
- 41. Lecture 8: 9-20-01 41 Output Ports • Queuing required when datagrams arrive from fabric faster than the line transmission rate
- 42. Lecture 8: 9-20-01 42 Switch Buffering • 3 types of switch buffering • Input buffering • Fabric slower than input ports combined queuing may occur at input queues • Can avoid any input queuing by making switch speed = N x link speed • Output buffering • Buffering when arrival rate via switch exceeds output line speed • Internal buffering • Can have buffering inside switch fabric to deal with limitations of fabric
- 43. Lecture 8: 9-20-01 43 Input Port Queuing • Which inputs are processed each slot – schedule? • Head-of-the-Line (HOL) blocking: datagram at front of queue prevents others in queue from moving forward
- 44. Lecture 8: 9-20-01 44 Output Port Queuing • Scheduling discipline chooses among queued datagrams for transmission • Can be simple (e.g., first-come first-serve) or more clever (e.g., weighted round robin)
- 45. Lecture 8: 9-20-01 45 Virtual Output Queuing • Maintain per output buffer at input • Solves head of line blocking problem • Each of MxN input buffer places bid for output • Challenge: map bids to schedule of interconnect transfers
- 46. Lecture 8: 9-20-01 46 Outline • IP Packet Format • NAT • IPv6 • Router Internals • Route Lookup
- 47. Lecture 8: 9-20-01 47 How To Do Variable Prefix Match 128.2/16 10 16 19 128.32/16 128.32.130/240 128.32.150/24 default 0/0 0 • Traditional method – Patricia Tree • Arrange route entries into a series of bit tests • Worst case = 32 bit tests • Problem: memory speed is a bottleneck Bit to test – 0 = left child,1 = right child
- 48. Lecture 8: 9-20-01 48 Speeding up Prefix Match - Alternatives • Content addressable memory (CAM) • Hardware based route lookup • Input = tag, output = value associated with tag • Requires exact match with tag • Multiple cycles (1 per prefix searched) with single CAM • Multiple CAMs (1 per prefix) searched in parallel • Ternary CAM • 0,1,don’t care values in tag match • Priority (I.e. longest prefix) by order of entries in CAM
- 49. Lecture 8: 9-20-01 49 Speeding up Prefix Match • Cut prefix tree at 16/24/32 bit depth • Fill in prefix tree entries by creating extra entries • Entries contain output interface for route • Add special value to indicate that there are deeper tree entries • Only keep 24/32 bit cuts as needed • Example cut prefix tree at 16 bit depth • 64K entries!! • Use a variety of clever techniques to compress space taken
- 50. Lecture 8: 9-20-01 50 Prefix Tree 1 0 1 1 1 5 5 X 7 3 3 3 3 X X 9 5 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Port 1 Port 5 Port 7 Port 3 Port 9 Port 5
- 51. Lecture 8: 9-20-01 51 Prefix Tree 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Subtree 1 Subtree 2 Subtree 3 1 1 1 1 5 5 X 7 3 3 3 3 X X 9 5
- 52. Lecture 8: 9-20-01 52 Speeding up Prefix Match • Scaling issues • How would it handle IPv6 • Other possibilities • Why were the cuts done at 16/24/32 bits?
- 53. Lecture 8: 9-20-01 53 Speeding up Prefix Match - Alternatives • Route caches • Packet trains group of packets belonging to same flow • Temporal locality • Many packets to same destination • Other algorithms • Bremler-Barr – Sigcomm 99 • Clue = prefix length matched at previous hop • Why is this useful?