Internetworking Part 2

Transcription

1 CMPE 344 Computer Networks Spring 2012 Internetworking Part 2 Reading: Peterson and Davie, 3.2, /08/2018 1

2 Aim and Problems Aim: Build networks connecting millions of users around the globe spanning networks based on any technology Problems: heterogeneity and scalability heterogeneity: need to support different LANs, point-to-point technologies, switched networks, different addressing formats scalability: addressing (management and configuration) and routing must be able to handle millions of hosts We will examine the (original) IP protocol (IPv4), IP addressing, packet forwarding, subnetting, and IPv6. 2

3 Outline Internet architecture IP service model IP forwarding Address translation (ARP) Automatic host configuration (DHCP) and error reporting (ICMP) Virtual Private Networks (VPNs) Subnetting Supernetting: Classless routing (CIDR) IPv6 3

4 Terminology The Internet and IP network = network based on one technology internet = network of networks The Internet = internet using IP routers = nodes connecting networks IP = Internet Protocol, current version IPv4 (IP Version 4) 4

5 What is internetwork Internetworking An arbitrary collection of networks interconnected to provide some sort of host-host to packet delivery service A simple internetwork where H represents hosts and R represents routers

6 What is IP Internetworking IP stands for Internet Protocol Key tool used today to build scalable, heterogeneous internetworks It runs on all the nodes in a collection of networks and defines the infrastructure that allows these nodes and networks to function as a single logical internetwork A simple internetwork showing the protocol layers

7 IP principles Each host has a globally unique IP address must be reachable by everyone and from everywhere Simple packet forwarding Network nodes simply forward packets Keeping the routers as simple as possible was one of the original design goals of IP 7

8 Internet architecture Recall Internet architecture from Chapter 1 Application layer: FTP, HTTP, (Last Chapter) Transport layer: TCP (reliable byte transfer) and UDP (unreliable datagram delivery) provide logical channels to applications (Next Chapter) Network or IP layer: IP protocol interconnects multiple networks into a single logical network (This Chapter) Link layer: wide variety of LAN and point-to-point protocols FTP HTTP NV TFTP TCP UDP IP NET 1 NET 2 NET n 8

9 IP Service Model Packet Delivery Model Connectionless model for data delivery Best-effort delivery (unreliable service) packets are lost packets are delivered out of order duplicate copies of a packet are delivered packets can be delayed for a long time Format of datagram (next slide) Global Addressing Scheme Provides a way to identify all hosts in the network

10 IP packet (datagram) format Format aligned at 32 bit words simplifies packet processing in software Header fields Version(4): version 4 (IPv4) or version 6 (IPv6) Hlen(4): header length, in 32-bit words (min 5) TOS(8): type of service (not widely used), used to give priorities to packets (QoS issue) Length(16): data+header length, in bytes Ident (16): used by fragmentation Flags/Offset (16): used by fragmentation TTL(8): time to live, no of times packet allowed to be forwarded (no of hops), default 64, detects packets caught in routing loop Protocol(8): identifies upper layer protocols, TCP=6, UDP =17 Checksum(16): header only, erroneous packets discarded DestAddr & SrcAddr (32): global Internet addresses Options: rarely used 10

11 Fragmentation/reassembly Each network has an MTU (Maximum Transmission Unit) e.g., Ethernet 1500 bytes, PPP 532 bytes Strategy Fragmentation occurs in a router when it receives a datagram that it wants to forward over a network which has (MTU < datagram. In general, try to avoid fragmentation: Hosts are encouraged to perform path MTU discovery Fragments are self-contained datagrams All the fragments carry the same identifier in the Ident field Flags (M-bit) and Offset used to guide fragmentation process Offset measured in 8 byte units Fragmented packet can be again re-fragmented Reassembly is performed only at destination host Reassembly does not try to recover lost fragments 11

12 Fragmentation/reassembly example Assume H5 wants to send a 1400 byte datagram to H8 (see slide 5): 1400 bytes + 20 bytes (IP header) MTU: and Ethernet 1500 bytes, PPP 532 bytes IP datagrams traversing the sequence of physical networks 12

13 Example (continued) Header fields used in IP fragmentation. (a) Unfragmented packet: H5 R1 R2 (b) Fragmented packets: R2 R3 H8 Can you think of a strategy that will abuse fragmentation to carry out a denial-of-service attack? How can one prevent such an attack? 13

14 IP global addressing Properties globally unique, 32 bits hierarchical: network part + host part address identifies interface end host has one interface router has many interfaces IP address domain name Original classful addressing class A, B and C networks defines different sized networks idea: small no of WANs, modest no of campus networks, large no of LANs Dotted Decimal Notation 32 bit addresses represented as groups of 8 bit integers e.g., Format A : to B : to C : to

15 Classful Adressing IP addr.= ; = Class: A (consider first bit - 0) Net ID: Host ID: IP addr.= ; = Class: C (consider first three bits 110) Net ID: Host ID: Class IP address Net Host Net Host # of PCs ID No bit # bit # in network A w x.y.z = 16,777,214 B w.x y.z =65534 C w.x.y z =

16 Network and Broadcast Numbers 0 indicates a network 255 broadcast address; all PCs in a network Class C address ; Network number: A device that wants to send a message to all the PCs in this network will use the following address; Class B address ; Network number: A device that wants to send a message to all the PCs in this network will use the following address;

17 Forwarding vs. routing Forwarding: Process of taking a packet from input interface, looking at its destination address, consulting a forwarding table, and sending the packet over an output interface determined by that table Routing: Process by which forwarding tables are constructed Routing must create optimal or efficient paths that the packets should follow 17

18 IP forwarding Preliminaries Every datagram contains destination s address Every router has a forwarding table Each host has a default router configured Routers maintain forwarding tables with multiple entries (constructed via routing process) Forwarding table maps network number into next hop router number or local interface number Strategy A router receiving a packet checks destination network address of datagram and if directly connected to destination network, then forwards directly to host need to map IP address to physical LAN address (ARP) if not directly connected to destination network, then forwards to next hop router 18

19 IP Datagram Forwarding Algorithm if (NetworkNum of destination = NetworkNum of one of my interfaces) then deliver packet to destination over that interface else if (NetworkNum of destination is in my forwarding table) then deliver packet to NextHop router else deliver packet to default router For a host with only one interface and only a default router in its forwarding table, this simplifies to if (NetworkNum of destination = my NetworkNum)then deliver packet to destination directly else deliver packet to default router

20 IP forwarding example 1 H1 H3: forwarding on the same network H5 H8: forwarding via R1 and R2 Router R2 20

21 IP forwarding example 2 Suppose, A wants to send a datagram to B. Thus, A knows I B = Step 1: A and B are on different networks, so to reach , A routes a datagram to R 1 (address ) Step 2: Router R 1 sees (from its table) that to reach , it is necessary to send a datagram to R 2 (address ). Step 3: Router R 2 sends to R 3 (address ). Step 4: Router R 3 is on the same network as B, so R 3 delivers the datagram to B directly. 21

22 Routers vs. bridges Bridges are used for interconnecting LANs of the same type whereas routers interconnect networks of different (or the same) types Bridge (+/-) + bridge operation is simple, requires less processing + transparent (no configuration needed when new nodes added to LAN) restricted topology (forwarding determined by a spanning tree) LANs use a flat addressing space (no hierarchical network structure) Router (+/-) + arbitrary topologies, enables use of efficient routing algorithms for distributing traffic (helps traffic management) + hierarchical addressing enables scalability: scalability requires minimization of address info stored in routers routing based on network numbers forwarding tables contain info on all networks, not all nodes requires IP address configuration packet processing more demanding Summary: bridges do well in small (~ 100 hosts) networks while routers are used in large networks (1000s of hosts) 22

23 Address translation What to do when router/host notes that it is connected directly to the network where an arriving packet is destined? Need to map IP addresses into physical LAN addresses destination host next hop router Techniques encode physical LAN address in host part of IP address not scalable! (not implemented) table-based (maintain IP address, PHY address pairs) ARP 23

24 Address translation Algorithm of direct routing Start Hosts A and B are on the same network. A wants to send a datagram to B. ARP is used here!! Knowing I B of B, learn P B Encapsulate the datagram in a data-link frame I B IP addr. of B P B phys. addr. of B Send the frame using P B End To know that A and B are on the same network: it is enough to compare netid s I A and I B. 24

25 ARP details ARP (Address Resolution Protocol) utilizes LAN s broadcast capabilities each node maintains table of IP to physical LAN address bindings broadcast request if address not in table table entries are discarded if not refreshed target machine responds with its physical LAN address ARP request contains also source addresses (physical and IP) all interested parties can learn the source address Node (host/router) actions: table entries timeout in about 10 minutes if node already has an entry for source, refresh timer if node is the target, reply and update table with source info if node not target and does not have entry for the source, ignore source info 25

26 ARP packet format HardwareType: type of physical network (e.g., Ethernet) ProtocolType: type of higher layer protocol (e.g., IP) HLen & PLen: length of physical and upper layer addresses Operation: request or response Physical/IP addresses of Source and Target 26

27 Mapping with ARP protocol A B IP = IP = MAC = 02C00011DE25 MAC=... Broadcast message 1.Who has IP = ? Frame from A Local Area Network Dest. MAC = FFFFFFFFFFFF Orig. MAC = 02C00011DE25 ARP Source MAC=02C00011DE25 Sorce IP = Target MAC =? Target IP = Unicast message C IP = MAC = 4B000D9182EA MAC address is a physical 2. IP address(unique locally). has MAC= 4B000D9182EA It is assigned to each network interface Frame from C card ( NIC ). Dest. MAC = 02C00011DE25 ARP features: 1. Works only in a LAN. 2. LAN must provide the broadcasting capability. 3. Broadcasting is used only to transmit ARP request messages, ARP reply messages are Orig. MAC = 4B000D9182EA ARP Source MAC=02C00011DE25 Sorce IP = Target MAC = 4B000D9182EA Target IP = unicast.

28 ARP Protocol Host 1 (datalink) Host 2 Host 3 Host N IP packet (datagram) to Host 3: phys. addr. of Host 3 -? ARP frame ARP frame (phys.addr. of Host3) Broadcast IP packet (in aframe) Time 28

29 Need for automatic host configuration! IP addresses need to be reconfigurable Ethernet addresses hardwired onto the network adapter. Ethernet addresses are configured into network by manufacturer and they are unique. IP addresses must be unique on a given internetwork but also must reflect the structure of the internetwork. IP address consists of network and host part. hosts can move between networks host gets new address in each network Need for automated host configuration hosts need other configuration info, e.g., the default router configuration manually impossible (too much work and errors) Automated Configuration Process is required Dynamic Host Configuration Protocol (DHCP) DHCP server is responsible for providing configuration information to hosts at least one DHCP server for each administrative domain DHCP server maintains a pool of available addresses centralized repository for configuration info two operation modes: administrator chooses host addresses and configures them to DHCP DHCP manages the addresses by allocating addresses dynamically 29 from a pool of available addresses (more sophisticated)

30 Server discovery: Newly booted or attached host sends DHCPDISCOVER message to IP broadcast address ( ) Message broadcasted only on same network If server on same network, host receives its IP address DHCP operation If not, message picked up by DHCP relay agent Relay agent knows address of DHCP server, forwards the message to DHCP server and host receives its IP address Use of DHCP relay agent makes it possible to have fewer DHCP servers (relay agent configuration simpler than DHCP server configuration) 30

31 Internet Control Message Protocol (ICMP) ICMP used for reporting errors in Internet Defines a collection of error messages that are sent back to the source host whenever a router or host is unable to process an IP datagram successfully Messages Echo (ping), Echo reply Redirect (from router to source host if router knows of a better route to packet s destination) Destination host unreachable due to link/node failure (protocol, port, or host) TTL had reached 0 - exceeded (so datagrams don t cycle forever) IP header checksum failed Reassembly failed Cannot fragment 31

32 Virtual private networks (VPN) Problem: group of isolated networks geographically distant from each other need to connect different networks together into a private network securely e.g., company with many branch offices Solution: VPN Methods Build a private network from leased lines (not shared) Create a virtual private network which is overlaid on top of public networks ATM and Frame Relay: Use virtual circuits IP: Use IP tunneling (usually coupled with secure IPsec) 32

33 VPN and IP tunneling Problem with IP: - not possible to connect to Internet via router without the whole Internet also knowing about your network. Tunneling virtual point-to-point link between two nodes separated by arbitrary no of networks created at R1 by providing it with address of R2 R1 encapsulates original packet in a new packet addressed to R2 packet forwarded normally inside IP network R2 receives packet and strips off packet header and notices payload contains an encapsulated packet addressed to some host inside network 2 33

34 Global Internet Scalability Issues IP hides hosts in address hierarchy, but... Inefficient use of address space class C network with 2 hosts (2/255 = 0.78% efficient) class B network with 256 hosts (256/65535 = 0.39% efficient) Too many networks today's Internet has tens of thousands of networks routing tables do not scale route propagation protocols do not scale 34

35 Subnetting Problem 1: Any network with more than 255 hosts, needs class B addresses, or should get many class C addresses Problem 2: Each new network implies additional entry in forwarding table large table Solution: Share one network number between several networks. another level is added to address/routing hierarchy : subnet 35

36 Network Mask (according to classes) The network mask is generally used to find the network and host bits of an IP address and thus determine the network address. Network mask allows the two computers on the network to understand that they are on the same network or not. class IP addr. Net Host Net Host Network No No bit # bit # Mask A w x.y.z B w.x y.z C w.x.y z ( ( (

37 Network Mask, example IP address: Network address: Host address: Network mask: Broadcast address: IP address: Network address: Network mask: Broadcast address: IP address: Network address: Network mask: Broadcast address:

38 Subnetting Makes most sense for large corporations or campuses Corporation networks share 1 network number Number of other networks within the corporation, using subnet masks E.g., a class B address, is shared among 8 networks (2 3 ), by using a 19-bit subnet mask ( = ) i.e., subnet addresses are defined by first 19 bits of the IP address. Host part now has a subnet part in it. Class B network address continues to be advertised to the rest of the Internet, subnetting only used within campus (i.e. Subnets visible only within site). 38

39 Subnetting Another example: a class B address, is shared among 256 networks (2 8 ), by using a 24-bit subnet mask : ( = ) 39

40 Subnet example Forwarding table at R1: 40

41 Forwarding algorithm with subnetting D = destination IP address for each entry < SubnetNum, SubnetMask, NextHop> D1 = SubnetMask & D if D1 = SubnetNum if NextHop is an interface deliver datagram directly to destination else deliver datagram to NextHop (a router) Q: Apply the algorithm when H1 is sending to H2 and H3. Notes: Would use a default router if nothing matches Not necessary for all ones in subnet mask to be contiguous Can put multiple subnets on one physical network Subnets not visible from the rest of the Internet 41

42 Subnetting Example Where does the router forward packets addressed to: > Interface > R > R > R > R4 Subnet Number Subnet Mask Next Hop Interface Interface R R3 <default> R4 42

43 Supernetting (CIDR) (1) Classful addressing: inefficient use of address space, address space exhaustion subnetting does not get around the fact that a network with more than 255 hosts requires a class B address class B net allocate addresses for 65K hosts, even if only 2K hosts in that network CIDR: Classless InterDomain Routing network portion of address of arbitrary length address format: a.b.c.d/x, where x is # bits in network portion of address network part /23 host part 43

44 Supernetting (CIDR) (2) Assign block of contiguous network numbers to near-by networks helps to aggregate routers less entries in the forwarding tables Restrict block sizes to powers of 2 E.g. Consider an Autonomous system(as) with 16 class C network numbers ( to ) the top 20 bits are same ( ) 20-bit network number (a single network prefix) can be used in forwarding tables. Something between class B and class C classless! CIDR tries to balance the desire to minimize the number of routes that a router needs to know against the need to hand out addresses efficiently. 44

45 Classless Addressing Requires to hand out blocks of class C addresses that share a common prefix The convention is to place a /X after the prefix where X is the prefix length in bits For example, the 20-bit prefix for all the networks through is represented as /20 By contrast, if we wanted to represent a single class C network number, which is 24 bits long, we would write it /24 45

46 Route aggregation with CIDR To achieve scalability, you need to reduce the amount of information stored in routers Uses a single entry in the forwarding table to tell the router how to reach a lot of different networks Breaks the rigid boundaries between address classes Hierarchical aggregation: IP has a two-level hierarcy (network+host) 46

47 A simple example Suppose hosts had arbitrary addresses Then every router would need a lot of information to know how to direct packets toward the host host host... host host host... host LAN 1 router router router WAN WAN LAN forwarding table 47

48 Example continued.. Number related hosts from a common subnet /24 on the left LAN /24 on the right LAN host host... host host host... host LAN 1 router router router WAN WAN LAN / /24 forwarding table 48

49 Example continued.. Easy to add a new host No need to update the routers E.g., adding a new host on the right Doesn t require adding a new forwarding entry host host... host host host... host LAN / /24 router router router WAN WAN LAN 2 host forwarding table 49

50 Longest Prefix Match IP forwarding mechanism assumes that it can find the network number in a packet and then look up that number in the forwarding table With CIDR, prefixes may be of any length (from 2 to 32 bits) and may overlap (some addresses may match more than one prefix). For example, suppose we have the following in the forwarding table at a router: (a 16-bit prefix) (a 24-bit prefix) A packet with destination address matches both prefixes The rule is to choose the longest prefix(longest match) to forward to ( in this case) Consider also The longest match would be not Efficient algorithms are required for finding the longest match in a forwarding table 50

51 CIDR Example Suppose the router does the longest-prefix match. Where does the router forward packets addressed (in hex) to: C4.5E > B C4.5E > A C > E 5E > F C4.6D.31.2E > C C4.6B.31.2E > D Net / Mask Length Next Hop C / 12 C4.5E.10.0 / 20 C / 12 C / 14 A B C D / 1 E / 2 F / 2 G 51

52 Today s Internet Sites are connected arbitrarily rather than as a tree-like structure as was the case in 1990s Some large corporations connect with each other. This is called peering A simple multi-provider Internet 52

53 Problems How do we build a routing system that can handle hundreds of thousands of networks and billions of end nodes? How to handle address space exhaustion of IPV4? How to enhance the functionalities of Internet?

54 Routing in the Internet (1) The Global Internet consists of Autonomous Systems (AS) interconnected with each other. AS: corresponds to an administrative domain examples: University, company, backbone network Each AS is assigned a 16-bit number A network with two autonomous system 54

55 Routing in the Internet (2) We can classify ASs into three types: Stub AS: small corporation: a single connection to one other AS. only carry local traffic Multihomed AS: large corporation: multiple connections to other AS s. (no transit, i.e. refuses to carry others traffic) Transit AS: backbone provider, hooking many AS s together (an AS has connections to more than one other AS, and is designed to carry both transit and local traffic) Q: Which ASes are Stub, Multihomed and Transit in Slide 52? Two-level routing: Intra-AS: Routing within a single autonomous system (administrator responsible for choice of routing algorithm within network) Inter-AS: Routing between autonomous systems. 55 Unique standard for inter-as routing: BGP

56 Intra-AS, Inter-AS routing Most common Intra-AS routing protocols: - RIP: Routing Information Protocol - OSPF: Open Shortest Path First - IGRP: Interior Gateway Routing Protocol Most common Inter-AS routing protocol is Border Gateway Protocol (BGP) - Assumes that the Internet is an arbitrarily interconnected set of ASs - Leave optimality aside. Just find a loop-free path - Thus only bothers with reachability R5 AS1 (RIP intra - AS routing) R1 BGP R2 R3 BGP AS2 ( OSPF intra - AS routing) R4 AS3 ( OSPF intra - AS routing) 56

57 Major Features 128-bit addresses Multicast Real-time service IPv6 (1) Authentication and security Autoconfiguration End-to-end fragmentation Enhanced routing functionality, including support for mobile hosts Protocol extensions 57

58 IPv6 (2) IPv6 addresses Classless addressing/routing (similar to CIDR) Notation: x:x:x:x:x:x:x:x (x = 16-bit hex number) contiguous 0s are compressed: 47CD::A456:0124 IPv6 compatible IPv4 address: :: Address assignment provider-based geographic 58

59 40-byte base header Extension headers (fixed order, mostly fixed length) fragmentation source routing authentication and security other options IPv6 Header 59

60 Routing for Mobile Hosts Mobile IP home agent Router located on the home network of the mobile hosts home address The permanent IP address of the mobile host. Has a network number equal to that of the home network and thus of the home agent foreign agent Router located on a network to which the mobile node attaches itself when it is away from its home network

61 Routing for Mobile Hosts Problem of delivering a packet to the mobile node How does the home agent intercept a packet that is destined for the mobile node? Proxy ARP How does the home agent then deliver the packet to the foreign agent? IP tunnel Care-of-address How does the foreign agent deliver the packet to the mobile node?

62 Routing for Mobile Hosts Route optimization in Mobile IP The route from the sending node to mobile node can be significantly sub-optimal One extreme example The mobile node and the sending node are on the same network, but the home network for the mobile node is on the far side of the Internet Triangle Routing Problem Solution Let the sending node know the care-of-address of the mobile node. The sending node can create its own tunnel to the foreign agent Home agent sends binding update message The sending node creates an entry in the binding cache The binding cache may become out-of-date The mobile node moved to a different network Foreign agent sends a binding warning message