Internet Networking Spring 2004

1. Internet Networking Spring 2004 Tutorial 1 Subnetting and CIDR Proxy ARP

2. Administrative Information • Course site: webcourse.technion.ac.il/236341 • Assistants: Rami Cohen, Taub 524, (829)4899 email: ramic@cs.tecnion.ac.il Galperin Vadim, Taub 441, (829)3941 email: gvadim@cs.tecnion.ac.il

3. Reminding – IP Addressing (Original Classful Scheme) • IP Address – 32-bit integer globally unique address • Dotted Notation: 132.68.37.54 • IP Classes – dividing an address to net id and host id • The prefix (net id) identifies a network. • The suffix (host id) identifies a host on this network.

4. Reminding – IP Addressing (Original Classful Scheme) • Class A – 7 bits to net id, 24 bits to host id 1.0.0.0 – 126.0.0.0 • Class B – 14 bits to net id, 16 bits to host id 128.1.0.0 – 191.255.0.0 • Class C – 21 bits to net id, 8 bits to host id 192.0.1.0 – 223.255.255.0 • Class D – for multicasting • Class E – reserved for future use (used for private addresses)

5. Weaknesses of Classful Scheme • Growth of routing tables in routers • Tens of thousands small (class C) networks. • Each network must be advertised. • Inflexible • Lack of a network classes for mid-sized organization (between class B and C). • Address space will be eventually exhausted

6. Classless Inter-Domain Routing (CIDR) • The concept of network "class" is deprecated. • Routing destinations are represented by network and mask pairs. • Routing is performed on a longest-match basis • Advantages: • Flexibility • Enabling network aggregation; thereby reducing the size of routing table. • Defined in RFC 1519.

7. Subnet Addressing • A site has a single IP network address assigned to it, but has two or more physical networks. • Different technologies. • Limits of technologies. • Network congestion. • Security consideration. • VLAN – separate one physical network into a few logical networks. • Administration (e.g. deferent departments in academic institute).

8. Subnet Addressing • From outside it looks like a single network • Only local routers know about multiple physical networks inside and how to route traffic among them • Host ID is divided into a subnet ID and host ID • Accepted as a standard at 1985 (RFC 950).

9. Subnet Routing • When a router gets a packet, it isolates by Net mask the packet net id address. • Each routing entry contain a net mask. • Routing is done on a longest-match basis. • If the packet is destined to other network then the router sends it to another router. • Otherwise the router sends the packet to the appropriate host on its attached networks.

10. Subnetting - Example Network 128.10.1.0/24 • A site with two physical networks. • Using subnetting, R advertise these networks as a single network (thus, R accepts all traffic for net 128.10.0.0) • Internal routing is done according to subnet id (i.e. the third octet of the address). 128.10.1.1 128.10.1.2 H1 H2 Rest of the Internet R . Network 128.10.2.0/24 All traffic to 128.10.0.0/16 128.10.2.1 128.10.2.2 H3 H4

11. Variable-Length Subnetting • Motivation: Consider the case when an organization has a few network of different sizes. • When we choose the subnet partitioning, we actually define constant number of possible physical subnetworks with maximum number of hosts on them. • Difficult to keep small (waist of subnet numbers) and big (the host id needs more bits) sub networks and there could be unnecessary spending of address space. • Solution: Variable-Length Subnetting. A subnet partition is selected on a per-network basis.

12. Example – Configuring a Network withVariable-Length Subnetting • We have a network with IP 202.128.236.0/24 • We need to support next sub networks: • 6 networks with 26 hosts • 3 networks with 10 hosts • 4 networks with 2 hosts

13. Example – Configuring a Network withVariable-Length Subnetting • If we take subnet mask of /27 bits then we can get 8 sub networks of 30 hosts (all 0’s and all 1’s of host addresses are reserved). • 11111111.11111111.11111111.11100000 • We need only 6 such sub networks.

14. Example – Configuring a Network withVariable-Length Subnetting • The rest 2 sub networks we will partition by subnet mask of /28 bits. • We will get 4 sub networks of 14 hosts in each • We need only 3 such sub networks.

15. Example – Configuring a Network withVariable-Length Subnetting • The rest we will partition by subnet mask of /30 bits. • We will get 4 sub networks of 2 hosts in each.

16. Example – Configuring a Network withVariable-Length Subnetting • Subnet mask #1 = 202.128.236.0 /27 • 11001010.10000000.11101100.11100000 • Subnet mask #2 = 202.128.236.0 /28 • 11001010.10000000.11101100.11110000 • Subnet mask #3 = 202.128.236.0 /30 • 11001010.10000000.11101100.11111100

17. Reminding - ARP • Two machines on a given network can communicate only if they know each other’s physical network address • ARP (Address Resolution Protocol) serves for mapping from high-level IP address into low level MAC address.

18. 10.4.10.91 00:11:22:33:44:55:bc Reminding – ARP ARP Response (00:11:22:33:44:55:12) ARP Request (10.4.10.90) 10.4.10.90 00:11:22:33:44:55:12 10.4.10.86 00:11:22:33:44:55:2e 10.4.10.20 00:11:22:33:44:55:6b 10.4.10.23 00:11:22:33:44:55:40

19. Proxy ARP • Proxy ARP (also called promiscuous ARP or ARP hack) is a technique used to map a single IP network prefix into two physical addresses. • Using the same network address space for more than one physical address • Assume that there are 2 networks A and B connected by router R that runs Proxy ARP • Using Proxy ARP, R can use the same net id for both networks.

20. 10.10.1.5 10.10.1.3 10.10.1.4 Proxy ARP Network A: 10.10.1.x/24 R PPP . Network B: 10.10.1.x/24

21. Proxy ARP • Router R replies to ARP requests that are generated by hosts on the PPP connection (Network B), in which the target IP is on network A, namely it sends its MAC address. • R know which hosts are connected through the PPP. • These host assume that the destination host are on the same physical network. • In their ARP table the router MAC address is associated with the destination IP address.