Chapter 4 Network layer

Transcription

1 Departamento de Tecnología Electrónica Computer Networking Chapter 4 Network layer The most of this slides is given as material with copyright from: Computer Networking: A Top Down Approach, 5 th edition. Jim Kurose, Keith Ross Addison Wesley, April 2009.

2 Chapter 3: Transport Layer Our goals: understand principles behind Network Layer services: Network Layer service models forwarding versus routing inside router Example: implementation in the Internet Network Layer 4 2

3 Chapter 4 outline 4. 1 Introduction 4.2 Virtual circuit and datagram networks 4.3 Router in datagram networks 4.4 IP: Internet Protocol Datagram format IPv4 addressing Basic ICMP How it works Network Layer 4 3

4 Chapter 4 outline 4. 1 Introduction 4.2 Virtual circuit and datagram networks 4.3 Router in datagram networks 4.4 IP: Internet Protocol Datagram format IPv4 addressing Basic ICMP How it works Network Layer 4 4

5 Network Layer transport segment (T_PDUs) from sending to receiving host on sending side encapsulates T_PDUs into datagrams (N_PDU) on receiving side, delivers T_PDU to transport layer Network Layer protocols in every host, router router examines header fields (N_PCI) in all IP datagrams (N_PDU) passing through it application transport network dt data link physical network data link physical network data link physical network data link physical network data link physical network data link physical network data link physical network data link physical network data link physical network data link physical network data link physical network data link physical application transport network data link physical Network Layer 4 5

6 Network Layer All layers from application to network are implemented in software. Data Link and Physical are implemented in hardware, known as network interface card or NIC. Each interface implements a particular Data Link and Physical protocol, known as link technology, network technology or just technology. Each interface has associated a link address, known as physical address or MAC address with 48 bits that identifies it. For example, 00:BF:3C:23:45:30 Application Transport Network Data Link Physical NIC Software Hardware More about physical addresses in the data link layer Network Layer 4 6

7 Network Layer In general, end systems usually use only one network interface, although mayhave several (e.g., ethernet and Wi Fi). Routers have several network interfaces. Each one connects to other routers or end systems. network link link link physical physical physical application transportt network link physical Network Layer 4 7

8 Network Layer Data Link Layer, through network interface, offers to Network Layer a service of N_PDUs deliver between routers or end systems connected by physical media and devices that implements up to Data Link Layer. Routers and end systems connected in this way are in the same broadcast domain. network link link link physical physical physical link link link physical physical physical application transportt network link physical application Transport network Link physical Network Layer 4 8

9 Two Key Network Layer Functions forwarding: move packets (N_PDUs) from router s input to appropriate router output routing: determine route taken by packets (N_PDUs) fromsource to dest. analogy: routing: process of planning trip from source to dest forwarding: process of getting through single interchange routing algorithms Network Layer 4 9

10 Interplay between routing and forwarding routing algorithm routing table _ p N PCI output link More about routing table soon value in arriving packet s header (N _ PCI) C) 0111 Q: Which interface will be forwarded by? Network Layer 4 10

11 Connection setup 3 rd important function in some network architectures: ATM, frame relay, X.25 before interchanging N_PDUs, two end hosts and intervening routers establish virtual connection, known as virtual circuit (VC) routers get involved (allocating resources) network vs transport layer connection service: network: between two transport entities and network layer of hosts and routers in the path are involved transport: between two application processes Network Layer 4 11

12 Network service model Q: What service model for channel transporting T_PDUs from sender to receiver? Example: services for individual T_PDUs: guaranteed delivery dli guaranteed delivery with less than 40 msec delay Example: services for a flow of T_PDUs: in order TPDUdelivery T_PDU guaranteed minimum bandwidth to flow restrictions on changes in inter T T_PDU PDUspacing Network Layer 4 12

13 Network Layer service models Guarantees? Network Service Congestion Architecture Model Bandwidth Loss Order Timing feedback Internet t ATM ATM best effort CBR ABR none constant rate guaranteed Minimum no no no yes yes yes No Yes No no (inferred via loss) no congestion yes CBR: Constant bit rate ABR: Available bit rate Network Layer 4 13

14 Chapter 4 outline 4. 1 Introduction 4.2 Virtual circuit and datagram networks 4.3 Router in datagram networks 4.4 IP: Internet Protocol Datagram format IPv4 addressing Basic ICMP How it works Network Layer 4 14

15 Network Layer connection and connection less service datagram network (packet switching) provides network layer connectionless service VC network (circuit switching) provides networklayer connection service analogous to the transport layer services, but: service: host to host to no choice: network provides only one type of service implementation: in hosts and in network core Network Layer 4 15

16 Virtual circuits source to dest path behaves much liketelephone circuit performance wise network actions along source to dest path There are three phases: call setup N_PDUs flow teardown each N_PDU carries VC identifier inside N_PCI (not destination host address) every router on source dest path maintains state for each passing connection link, router resources (bandwidth, buffers) may be allocated to VC (dedicated resources = predictable service) Network Layer 4 16

17 VC implementation a VC consists of: 1. path from source to destination 2. VC numbers, one number for each link along path 3. entries in forwarding tables in routers along path that indicates the path and the VC number to use in every case. N_PDUs belonging to VC carries VC number (rather than dest address) VC number can be changed on each link. VC number interface number Network Layer 4 17

18 Virtual circuits: signaling protocols used to setup, maintain and teardown VC used in ATM, frame relay, X.25 not used in today s Internet t application transport network data link physical ysca 5. Data flow begins 6. Receive data 4. Call connected 3. Accept call 1. Initiate call 2. incoming call application transport network data link physical Network Layer 4 18

19 Datagram networks no call setup at Network Layer routers: no state about end to end connections no network level concept of connection NPDUsforwarded N_PDUs using destination host address N_PDUs between same source dest pair may take different paths application transport network data link physical 1. Send N_PDU 2. Receive N_PDU application transport network data link physical Network Layer 4 19

20 Datagram or VC network: why? Internet (datagram) data exchange among computers elastic service, no strict timing req. smart end systems (computers) can adapt, p,p perform control, error recovery simple inside network, complexity at edge many link types different characteristics uniform service difficult ATM (VC) evolved from telephony human conversation: strict timing, reliability requirements need for guaranteed service dumb end systems telephones complexity inside network Network Layer 4 20

21 Chapter 4 outline 4. 1 Introduction 4.2 Virtual circuit and datagram networks 4.3 Router in datagram networks 4.4 IP: Internet Protocol Datagram format IPv4 addressing Basic ICMP How it works Network Layer 4 21

22 Router interfaces Technology used in each interface in the same router is independent. Por example, the edge router in a domestic network usually has an Ethernet/WI FI interface and an ADSL interface. Router interfaces are identified by A letter ltt depending di on the technology and E.g.: E:Ethernet 10 Mbps, Fa:Fast Ethernet, Gi:Gigabit Ethernet, To: Token Ring, Se:Serial A number to distinguish interfaces with the same technology E.g.: E0, Fa0, F0F1S0 Fa1, Se0.. E0 Se0 E1 Network Layer 4 22

23 Router and Logic Networks (I) In each interface of a router, using the appropiate transmission media, end systems or other routers can be connected. It is possible to use other device that implements up to Data Link Layer. For example, switch, access point or hub. All of them are in the same broadcast tdomain, that t is, they process all frames (E_PDUs) that has a destination physical address or multicast inside the DL_PCI. In general, all belong to the same logic network. They share the same network identifier in the IP address. notes 1. A broadcast physical address is for identifing a group ofnetwork interfaces. 2. IP addresses are hierarchical. The bits that make up them have two parts, one identifies the logic network (network identifier) and the other identifies to the end system or router. (host identifier) In general we can say that IP address =NetworkX.HostY. More soon Network Layer 4 23

24 Router and Logic Networks (II) Every interface of a router belongs to a one different logic network. It will have a different IP address. The part that identifies the network will be different, e.g., Net1, Net2, Every interface of a router is a broadcast domain. A router is not a transparent device, althoughdoesn t modify the datagrams (N_PDUs) that cross through it. End systems must know its IP address. E.g., in soho networks it is necessary to know which IP address has the Ethernet or WI_FI interface of the edge router. Routers must know IP address of routers to forward NPDUs N_PDUs (routers connected directly). Network Layer 4 24

25 Example Q. How many broadcast domain have? Q. How many logic networks have? What if we assign Net1.Host3 and Net1.Host4 the network identifier Net3? Q. If Net1.Host1 wants to send data to Net2.Host2, who has the network layer to deliver the N_PDU that encapsulates this data to? Which source and destination address will appear in that N_PDU? Q. Which source IP address will have the N_PDU that end system Net2.Host2 receives in the N_PCI? Net1.Host1 SUP. NET. NETWORK SUP. NET. Net2.Host2 Link Physical Link Physical Link Physical Link Physical Net1.Host4 Net 1 Net1.Host2 P0 Net2.Host1 P1 Net2 Router Net1.Host3 Net1.Host5 Net1.HostN Net2.Host3 Net2.HostN Network Layer 4 25

26 Routing table (I) The two key functions of network layer use routing table (RT). Routing can modify its content. Forwarding to know which is the target interface to forward a N_PDU to get the destination. End systems and routers have a routing table. In forwarding process RT entries are for knowing next hop in the path. RT entries are for knowing the path to follow in forwarding process and at least the following appears: Network Next hop Interface Network identifier 3-layer address of the next hop router Output network interface Network Layer 4 26

27 Routing table (II) Q. How RT entries are filled? A1. Automatically When assigning IPaddressto network layerdevice an entryisadded forthe logic network where this device belongs. A2. Manually By using management commands it is possible to fill researchable networks in the RT. A3. Dinamically By routing protocols that implement some algorithms that set which is the best path to a particular logic network. It is specific for routers. Typical Internet routing protocolos: RIP OSPF BGP Also, in the RT a special entry is usually included, known as default route, that will be used in case there isn t any specific entry for a particular logic network. A reserved network identifier is used. Network Layer 4 27

28 Routing table use In every data request (Data.request) Network Layer receives a N_IDU that has a N_ICI that is the destination IP address. The Network Layer with the destination IP address looks for a matching entry in the RT to check if there is an entry that indicates how reaching the destination network. In case there is an entry in RT, the router delivers the N_ PDU (without modifing source and destination IP in the N_PCI) to the next hop through the indicated interface. If destination is in thesamenetwork, t known as connected tddirectly, router delivers the N_PDU to it through the indicated interface. In not, router will discard the N_PDU. That network wouldn t be reachable by that Network Layer. Network Layer 4 28

29 Routing table example RT Host Net1.Host1 Network Next hop Interface RT Router 1 Network Next hop Interface Net2.Host2 Net1.Host1 Net1.Host2 Net2.Host1 NET 1 E0 Router 1 E1 NET 2 Net2.Host3 Net2.HostN Net1.Host3 Net1.HostN Network Layer 4 29

30 Routing table example RT Host Net1.Host1 Network Next hop Interface Net1 E Net2 Net1.Host2 E Net1.Host1 RT Router 1 Network Next hop Interface Net1 E0 Net2 E1 Net2.Host2 NET 1 Net1.Host2 Net2.Host1? E2 E0 E0 E1? E2 Router 1 NET 2 Router 2 E1 Net3.HostN Net1.Host3 Net1.HostN NET 3 Net2.Host3 Net2.HostN Net3.Host1 Net3.HostN 1 Network Layer 4 30

31 Routing table example RT Host Red1.Host1 Network Next hop Interface Net1 E Net2 Net1.Host2 E Net3 Net1.Host2 E Net4 Net1.Host2 E Net1.Host1 RT Router 1 Network Next hop Interface Net1 E0 Net2 E1 Net4 E2 Net3 Net4.Host1 E2 Net2.Host2 Net1.Host2 Net2.Host1 NET 1 Router 2 Net4.Host2 E0 E0 E2 Net4.Host1 Router 1 E1 NET 2 E1 Net3.HostN Net1.Host3 Net1.HostN NET 3 Net2.Host3 Net2.HostN Net3.Host1 Net3.HostN 1 Network Layer 4 31

32 Routing table example RT Host Red1.Host1 Network Next hop Interface Net1 E Default Red1.Host2 E RT Router 1 Network Next hop Interface Net1 E0 Net2.Host2 Net2 E1 Net1.Host1 Net4 E2 Net3 Net4.Host1 E2 Net1.Host2 Net2.Host1 NET 1 Router 2 Net4.Host2 E0 E0 E2 Net4.Host1 Router 1 E1 NET 2 E1 Net3.HostN Net1.Host3 Net1.HostN NET 3 Net2.Host3 Net2.HostN Net3.Host1 Net3.HostN 1 Network Layer 4 32

33 Routing table principles Eachdevice with ihnetwork layer does its own decisions, i based on information that mantains in its routing table. Not all network kdevices have the same information in their routing tables. Routing information about a route from a network kto other doesn t provide routing information about the return route. Network Layer 4 33

34 Example RT Router 1 RT Router 2 Net. Next hop Interface Net. Next hop Interface Net1 E0 Net4 E0 Net2.Host2 Net3 E2 Net2 E2 Net4 E1 Net3 Net4.Host2 E1 Net2 Net4.Host1 E1 Net1.Host1 Net1.Host2 Router 1 Net4.Host2 Router 2 NET 1 E0 Net3.HostN E2 E1 E0 Net4.Host1 E1 Net2.Host1 NET 2 NET 3 Net1.Host3 Net1.HostN Net3.Host1 Net3.HostN 1 Net2.Host3 Net2.HostN Q1. Is it possible that Net1.Host3 sends N_PDUs to Net3.Host1? and viceversa? Q2. Is it possible that Net1.Host3 sends NPDUs _ to Net2.Host2? and viceversa? Network Layer 4 34

35 Buffering It allows routers to store arriving N_PDUs through interfaces before processing them. It allows routers to store N_PDUs before being transmitting by an interface. When buffering? Heuristic rule (RFC 3439): average amount of space in buffer should be equal to RTT times interface bandwidth (R) (typical value for RTT is 250 msec) e.g., assuming RTT=250msec and interface with R = 10 Gpbs, buffer needed in the interface is 10 9 x0,25 = 2,5 Gbit The current recommendation, dti supposing N TCP flows going through the interface is: RTT. R Buffer = N Network Layer 4 35

36 Chapter 4 outline 4. 1 Introduction 4.2 Virtual circuit and datagram networks 4.3 Router in datagram networks 4.4 IP: Internet Protocol Datagram format IPv4 addressing Basic ICMP How it works Network Layer 4 36

37 The Internet Network Layer There are several Nt Network protocols thatt are in host and routers. Not all routing protocols works in Network Layer. They are only in routers. Transport layer: TCP, UDP Routing protocols path selection RIP, OSPF, BGP Network layer ICMP protocol error reporting router signaling l IP protocol (RFC 791) ARP protocol addressing conventions Matching physical addr Datagram (N_PDU) format and IP addr Forwarding N_PDU Link layer physical layer routing table Network Layer 4 37

38 Chapter 4 outline 4. 1 Introduction 4.2 Virtual circuit and datagram networks 4.3 Router in datagram networks 4.4 IP: Internet Protocol Datagram format IPv4 addressing Basic ICMP How it works Network Layer 4 38

39 IP datagram format IPprotocol version number IP Header Length (IP_PCI) in 32 bits words (4 bytes) type of N_UD max number remaining hops (decremented at each router) (1 byte) Multiplexion/ Demultiplexion how much overhead (PCI) with TCP? 20 bytes of TCP 20 bytes of IP = 40 bt bytes + app layer overhead (A_PCI) ver IHL type of service 32 bits 16 bit identifier time to protocol live flgs length 32 bit source IP address fragment offset header checksum 32 bit destination IP address Options (if any) data (variable length, typically a TCP or UDP segment) total datagram (IP_PDU) length (bytes) for fragmentation/ reassembly IP_PCI E.g. timestamp, record route taken, specify list of routers to visit. IP_UD Network Layer 4 39

40 IPv4 Fragmentation & Reassembly (I) network links have MTU (Maximum Transfer Unit) largest possible link level frame (E_SDU and R_PDU). different link technology types, different MTUs If IP_PDU size > transmitting interface MTU, large IP_PDU divided ( fragmented ) within net one IP_PDU PDU becomes several IP_PDUs with appropriate size (< origin IP_PDU) reassembled only at final destination (destination Network Layer) IP header bits used to identify, order related fragments reassembly fragmentation: in: one large IP_PDU out: 3 smaller IP_DPUs Network Layer 4 40

41 IPv4 Fragmentation & Reassembly (II) IP_ PCI used for fragmentation, identifier, e flag and offset of the fragment IP_PCI flag is made up by three bits 0D f M f, where: D f (Don t fragment), if set to 1 indicates not fragment. M f (More fragments), if set to 1 indicates there are more fragments. There isn t any more fragment is set to 0. Identifier IP_PCI is for labeling a IP_PDU and distinguish from the rest. All fragments of a IP_ PDU have the same identifier. Segment offset IP_PCI is for knowing the place that IP_UD bytes carried by it has, inside the total number of IP_SDU bytes that was carried by the origin IP_PDU. It is given in multiple of 8. Network Layer 4 41

42 How fragmentation works When a Network Layer fragments: All fragments that it generates will have the same IP_PCI of the origin IP_PDU but bit Mf, TTL and fragment offset. Number of bytes of IP_PCI of the fragments matches the origin I_PDU one. checks bit Mf: Ifi itwasset to 0, set itto1 in allfragments but the last one, that it has this bit set to 0. otherwise, all fragments will have bit MF set to 1. Fragment offset will indicate the relative position of the IP_UD of the fragment in 8 bytes blocks and it s zero only for the first fragment. The reason to use 8 bytes blocks is because the field is 13 bits width and IP_ PDU is up to 2 16 bytes length. The number of bytes of IP_UD for each fragment but the last must be multiple of 8. The maximum number of bytes of IP_UD will be the MTU lengthin bytes of the IP_PCI. If MTU length in bytes of the IP_PCI is not multiple of 8, bytes would be wasted. Network Layer 4 42

43 How reassembly works If destination Network Layer receives a IP_PDU with bit MF set to 1, knows that IP_UD that contains is not complete, has received a fragment and has to wait for receiving all IP_PDUs with the same identifier and bit MF set to 1 or to 0. Network Layer will know that it s finished when there isn t any gap between segment offset IP_PCI of the received IP_PDUs (fragments). Reassembly consists of ordering IP_PDUs by segment offset IP_PCI. IP_UD of each fragment is taken, put in order (indicated by fragment offset IP_ PCI) and delivered in a IP_ SDU to the upper layer when it is reassembled. Network Layer 4 43

44 Example IDENTIFIER= 111 IHL= 5 LENGTH= 2020 FLAG = 0, DF=0, MF=0 Offset = 0 TTL= 5 Protocol = 1 IDENTIFIER= 111 IHL= 5 LENGTH= 1500 FLAG = 0, DF=0, MF=1 Offset = 0 TTL= 4 Protocol = 1 MTU 3000 bytes MTU 1500 bytes IDENTIFIER= 111 IHL= 5 LENGTH= 540 FLAG = 0, DF=0, MF=0 Offset = 185 TTL= 4 Protocol = 1 Network Layer 4 44

45 Chapter 4 outline 4. 1 Introduction 4.2 Virtual circuit and datagram networks 4.3 Router in datagram networks 4.4 IP: Internet Protocol Datagram format IPv4 addressing Basic ICMP How it works Network Layer 4 45

46 IPv4 addressing Known as IP address or IPv4 32 bits (4 bytes) with a hierarchical addressing scheme: Network ID (Red) Host Number of bits for network id and for host depends on the addressing scheme. Notationin o a IPv4 address: 32 bits address grouping bytes Each byte in decimal and separated by dot Manual or dynamic setup End systems has to set its IP address and the gateway IP address. Routers have an IP address for each interface. Network Layer 4-46

47 Types of IPv4 addresses Three types of IPv4 addresses: Unicast: To send IP_ PDUs to a single gedestination. All routers and end systems must have assigned at least one IP address of this type. Broadcast: To send IP_PDUs to all devices (hosts and routers) in the same logic network. Each logic network has an address of this type. Multicast: To send IP_PDUs PDUs to a group of devices (hosts and routers) in the same or different logic network. All devices must have setup the same address of this type. Network Layer 4 47

48 Special addresses Two meanings All 0s This host: Used as source IP address when this host doesn t have any (e.g. device without configuration). Identifier for any network (Default route in RT) Network address Network All 0s Identifier for the logic network (used by RT) Directed Network All 1s Identifies all devices in a network. Broadcast Limited All 1s Identifies all devices in the same network of source Loopback address 127 Any digit Used to check the network layer in a device. Network Layer 4 48

49 IPv4 addressing scheme To fix which part of the IPv4 address is for identify the network and which to identify the host is used: Classful l addressing Obsolete. Classless addressing Used currently. Network Layer 4 49

50 Classful addressing (I) It uses first byte of IP address to fix which part is for network and which is for host. There are 5 classes: 1stbyte 2nd byte 3rd byte 4th byte Class A Network Host Host Host xxxxxxx Class B Network Network Host Host Unicast xxxxxx Class C Network Network Network Host Class D Class E xxxxx xxxx xxxx Multicast Experimental Network Layer 4 50

51 Example of classful addressing Class C; last byte identifies the host, the rest of bytes identifies the network. Q. How many devices with network layer as maximum is possible to identify in each logic network? RT Router Network Next hop Interface E E E Network Network E E2 E IP_PDU to Limited broadcast RT Host Network Next hop Interface E E Network Network identifier Directed broadcast IP_PDU to Network Layer 4 51

52 Classful addressing summary Network Class A. Network 0 is not used is the address used if a device with network layer is not configured is for identifing to any network. It appears in theentryof theroutingtablethat h represent thedefault route. Example of default route RT Network Next hop Interface E0 Network 127 is not used. Any IP address of this network is for checking the network layer of a device networks with available addresses for devices Class B networks of hosts Class C networks with availabe addresses for devices Network Layer 4 52

53 Classless addressing (I) To fixwhich h part identifies host or network kin a IP address, a network prefix is used: 32 bits address followed by /x, where x indicates the number of bits more significant of the IPv4 address that identifies the network (the rest identifies the host). X can be 0 to 32. For example, the network prefix to identify a network with the same amount of network layer devices that a class B network would be /16. Broadcast address would be We can assign any address in the next host prefix range to Network layer : /16 to /16. Given a network prefix a.b.c.d/x, it s possibletoaddress (2 32 x 2) network layer devices. Network Layer 4 53

54 Classless addressing (II) Prefix notation is not usually used when configuring network layer devices. /x is replaced by a netmask or subnet mask. It uses the same notation that IPv4 address where First X bits are set to 1. Last 32 X bits are set to 0. For example, a device with host prefix /16 would be configured by: IP address: Netmask: /0 is the identifier for any network. Netmask notation Example default route RT Network Next hop Interface E0 Prefix notation Example default route RT Network Next hop Interface / E0 Network idenfier netmask Network Layer 4 54

55 Classless addressing (III) Thisaddressing i scheme is known as CIDR (Classlessl InterDomain Routing) (RFC 4692). It allows to assign IPv4 address block depending on the actual needs, known as CIDR block. It doesn t waste IPv4 addresses. Example, if a company needs 2000 IPv4 addresses. With classfull addressing it needs a whole class B network addresses are wasting. With CIDR a network prefix X.X.X.X/21 would be enough. Network Layer 4 55

56 Example of classless addressing Q. What is the netmask to configure these devices? Q. How many network layer er devices as maximum m is it possible to identify in each logic network? / /22 Network / /22 RT Router Network Next hop Interface /22 E /22 E /22 E / /22 E0 E1 E2 Network / / / /22 Network / /22 note To know which is the right path to the destination IP addr, an and logic operation has to be done with the netmask of a particular RT entry and the destination IP addr. If the network identifier matches the one in the RT then this is the route to take / /22 RT Host /22 Network Next hop Interface /22 E / E IP_PDU to Network identifier Directed Broadcast Network Layer 4 56

57 Subnets It allows to address smaller logic networks from a CIDR block that fits to the amount of IP addresses needed. Each subnet can have different number of network layer devices. To create subnets some bits are borrowed from the bits that identify the host to identify the network. Given a network prefix with x bits to identify the network and 32 x bits to identify the host, if n bits are borrowed, with n<32 x 1, then: 2 n subnets with 2 32 n x 2 available IP addresses are created. Where x+n is the number of bits thatidentifies thenetworkt kinsideid each subnet. Network Layer 4 57

58 Subnets examples Let network prefix be /2323 0/23 Network part Host part If 1 bit is borrowed, two subnets are created: Network Host part part / /2423 0/24 If 1 bit is borrowed again from one of them, e.g /24, two new subnets would be created. Nt Network Host part part / /25 So from the CIDR block /23 three subnets have been created: /25 and /25 with available addresses and /24 with available addresses. Network Layer 4 58

59 Subnets advantages They y allow to agregate g routes in routing tables. When subnets are created from a CIDR block (the same network prefix), if all those subnets are reachable by the same interface, they can summarise by the origin network prefix. Company /23 Company /23 Company /23 Fly By Night ISP. E0. RT Internet Network Next hop Interface / E / /30 note Thesearchin the routing table carriesout by starting with entries beginning by the network prefix /32 and it ends with /0. The next hop will be the entry that has more bits in common with the destination IP addr from the IP_PDU. E1 Internet Network Layer 4 59

60 How CIDR blocks are assigned? Currently there isn t any available block. The last one was assigned in February Why have they run out? Mobile devices. Non efficient use of available address space. Internet user demography. ISPs distribute ib t among their clients the CIDR blocks that t have assigned. They don t usually assign a fixed IP anymore. If there isn t any available IPv4 addresses, how are devices identified? Private addressing and NAT IPv6 Progresive migration. Network Layer 4 60

61 Private addressing In 1996 a set of addresses were reserved called private addresses (RFC1918): They are a reserved IP address range to be used only in private networks (these IP addresses can t appear in network core of the Internet). For example, to address a non public Intranet, a laboratory, a domestic network... Reserved range: Class Address range CIDR prefix A /8 B /12 C /16 Network Layer 4 61

62 Private addressing and NAT Pi Private addresses and NAT (Network Address Translation, RFC 3022) are used to allow one network with private IP addresses can access to the Internet. NAT is usually implemented in routers. Internet Rest of the Internet Local network (e.g., domestic network) Network / All outgoing IP_PDUs have the same source IP address: IP_PDUs with source or destination inside this network have network addresses from network /24 as source or destination (as always) Network Layer 4 62

63 Chapter 4 outline 4. 1 Introduction 4.2 Virtual circuit and datagram networks 4.3 Router in datagram networks 4.4 IP: Internet Protocol Datagram format IPv4 addressing Basic ICMP How it works Network Layer 4 63

64 ICMP: Internet Control Message Protocol (RFC 792) Used by end systems and routers to communicate network layer information Error report: unreachable host, or network, or port, or protocol Works over IP (network layer): ICMP_PDUs (known as ICMP messages) are encapsulated in IP_PDUs (IP datagrams). ICMP messages: Echo request/reply (used by ping command) Exceeded TTL (used by tracert command) ICMP IP A Ping Echo Request Eh Echo Reply Internet ICMP IP tracert Echo Request ICMP ICMP IP A TTL=1 Exceeded TTL IP.. Internet B B Network Layer 4 64

65 Chapter 4 outline 4. 1 Introduction 4.2 Virtual circuit and datagram networks 4.3 Router in datagram networks 4.4 IP: Internet Protocol Datagram format IPv4 addressing Basic ICMP How it works Network Layer 4 65

66 IPv4 operation (I) The Internet network layer needs that network layer devices are configured and have the routing table filled: There are two mechanisms to configure end systems: Manually using the operative system interface. At least IP address., netmask and default gateway must be configured (edge router). Addressof one or several DNS servers It is not necessary to know the IP addresses. Dinamically using some protocol, e.g. DHCP (Dinamic Host Configuration Protocol). Configuration is carried out automatically and during a period of time. It is possible to release and renew the configuration. In routers it is necessary to configure only IP and netmask of each interface. Network Layer 4 66

67 IPv4 operation (II) The Internet network layer needs that network layer devices are configured and have the routing table filled: End systems require as minimum two entries (included automatically): The one for the logic network the end system bl belongs to (doesn t need next hop). Default route whose next hopis the egde router. Routers requiere one entry for each reachable network: Introduced manually (static). E.g: default route. Learned dinamically by a routing protocol. As minimum, table includes network connected directly, that is, the network accessed directly by their interfaces (automatic). Network Layer 4 67

68 IPv4 operation (III) Before sending a IP_PDU, IP protocol checks if there is any entry in the RT for the destination network: If there is not any coincident entry, the IP_PDU will not be sent. Ifthere is a coincident entry, the IP_PDU PDUwillbe sent through the interface indicated in the RT entry using data link layer services to: the destination if it is connected directly. the device whose IP address matches the indicated one in the RT. In the end systems it will almost always be the interface of the edge router. note Before requesting the IP_PDU sending to the data link layer a mapping IP addr/mac addr will be needed, using ARP protocol. More in next chapter Network Layer 4 68

69 IPv4 operation (IV) When receiving a IP_PDU, network layer check if the destination network is the same, that is, if IP addr matches the one it has configured: If it matches, network layer will process the IP_PDU. If it does not match, then: if it is an end system, network layer will discard the IP_PDU. If it is a router, network layer will forward it if it knows how reaching the destination network, by this it does the following: Check and modify, if appropriate, the TTL value of the IP_PDU PDU header. If it is 1, it discards the IP_PDU (it doesn t forward it). otherwise, it decreases the TTL value by 1. Repeat the actions that network layer does to send a IP_PDU. Network Layer 4 69

70 Example of sending IP_PDUs Network RT Router 1 Next hop / / / / / / RT Host /22 Network Next hop 1. From to From to Q. What is the TTL value that is received by inside the IP_PDU? Network RT Router 2 Next hop / / / / / / / /24 E1 E0 E1 R /24 R / /22 E / /24 E / / / / / / /22 To: To: /24 E /24 E1 E1 E / /24 R / /24 R4 E2 E /22 RT Router 3 RT Router /22 Network Next hop Network Next hop / / / / / / / / / / / / / Note: Interfaces aren t shown in RTs / INTERNET Network Layer 4 70

71 Chapter 4: Summary Network Layer characteristics in datagram networks. It works same in hosts and routers Using routing tables. IP protocol: How it fragments and reassemblies. How devices are addressed. How it sends and receives IP_PDU. How it uses routing table. Next: Leaving network core and logic network (Network Layer) Incoming to physical network (broadcast domain) Network Layer 4 71