Notes

Outline

ICS 156: Networking Lab Magda El Zarki Professor, ICS UC, Irvine

Course Outline Ch 1: Introduction Ch 2:Bridges Ch 3:Routers Ch 4: Transport Protocols

Ch. 1 Introduction (1/2) Layered Architectures The TCP/IP (or Internet) Architecture A Networking Example IP Addressing Packet Encapsulation Port Numbers Internet Standards 802.3/Ethernet IP

Ch 1. Introduction. (2/2) Address Resolution Protocol Internet Control Message Protocol Ping Traceroute

1.1 Layered Architectures Networking protocols are developed in layers so as to simplify the task of 1) developing new services and 2) upgrading and maintaining the system. Each layer performs a well defined function. The internals of a layer can be changed without affecting the neighboring layers. Layers exchange information via clearly specified interfaces. Layers communicate on a peer to peer basis.

A Layered Architecture

The 5 Basic Layers Layer 1: Physical Layer - mechanical and electrical network interface definitions Layer 2:  Data Link Layer - framing, data transparency, error control over one link/trunk Layer 3: Network Layer - routing, addressing, call set-up and clearing across the data network, flow control (internal and external) Layer 4: Transport Layer - end-to-end  message transfer,  connection management, error control Layer  5: Application Layer – user processes

Ch. 1 Introduction Layered Architectures The TCP/IP (or Internet) Architecture A Networking Example IP Addressing Packet Encapsulation Port Numbers Internet Standards 802.3/Ethernet IP

1.2 The TCP/IP Suite

Ch. 1 Introduction Layered Architectures The TCP/IP (or Internet) Architecture A Networking Example IP Addressing Packet Encapsulation Port Numbers Internet Standards 802.3/Ethernet IP

1.3 A Networking Example

Ch. 1 Introduction Layered Architectures The TCP/IP (or Internet) Architecture A Networking Example IP Addressing Packet Encapsulation Port Numbers Internet Standards 802.3/Ethernet IP

1.4 IP Addressing Every device on an internet must have a unique address (just like everyone on the phone network has a unique telephone number!). Each network element and attached host has two addresses: Physical Address and Network Address

1.4.1 Address Allocation

1.4.2 Addresses and Names The physical address is generally a flat address whereas the network address is generally hierarchical. The network address is referred to as an IP address, and is the address of the network interface. In addition to network addresses, hosts are often given a name too. The name makes it easier for people to remember. They are generally hierarchical too, identifying subnet and net where host is located.

Addressing Example

Multiple addresses per Device A network device that connects to several networks, will have a network address for each network it is connected to and it will also have a physical address per network interface. A Router is a perfect example of a device that has more than one network address.

A Multiple Interface Device

1.4.3 Address Mapping Tables map names to network addresses. Domain Name Servers are used to help find adresses. Tables also map network addresses to physical addresses. If a physical address is not found in the table, a protocol is used to locate the host and obtain its physical address (ARP and RARP).

1.4.4 Structure of IP Addresses IPv4 uses 32 bits for addressing, whereas IPv6 has assigned 128 bits. Both IPv4 and IPv6 have a particular structure for their address space: it is partially hierarchical, it has an internet routing part and a local routing part. Subnet addressing was introduced to facilitate the routing and administration within organizations. If forms part of the local part and does not affect the internet routing. IPv4 has 5 address classes. An organization is assigned an address space in a class based upon its size.

1.4.5 IPv4 Address Classes

1.4.6 Addressing Formats IP addresses are represented in dotted decimal format. Each 8 bits (byte) is represented by a decimal number. The 4 bytes are separated by dots: 00001010 00000000 00000000 00000000 = 10.0.0.0 = Class A, net ID: 10 10000000 00000011 00000010 00000011 = 128.3.2.3 = Class B, net ID: 128.3, host ID: 2.3 11000000 00000000 00000001 11111111 = 192.0.1.255 = Class C, net ID: 192.0.1, host ID: all hosts (broadcast)

Ch. 1 Introduction Layered Architectures The TCP/IP (or Internet) Architecture A Networking Example IP Addressing Packet Encapsulation Port Numbers Internet Standards 802.3/Ethernet IP

1.5 Encapsulation User data -> application header/trailer -> TCP (or UDP) header/trailer -> IP header/trailer -> data link layer header (e.g. Ethernet header)/trailer. When the TCP/UDP header is attached the packet is called the TCP/UDP segment. When the IP header is attached it is referred to as the IP datagram. When the data link layer header is attached, it is called the link frame (e.g., Ethernet frame)

Ch. 1 Introduction Layered Architectures The TCP/IP (or Internet) Architecture A Networking Example IP Addressing Packet Encapsulation Port Numbers Internet Standards 802.3/Ethernet IP

1.6 Port Numbers TCP and UDP use port numbers to identify the application. The port numbers are 16 bits. Servers are assigned well known port numbers, e.g., FTP is assigned TCP port number 21, telnet is assigned TCP port number 23. These port numbers are assigned by the Internet Assigned Numbers Authority (IANA). Ports 1 to 1023 are all well known (listed in /etc/services). Clients are allocated ephemeral port numbers between 1024 and 5000. They are short lived.

Ch. 1 Introduction Layered Architectures The TCP/IP (or Internet) Architecture A Networking Example IP Addressing Packet Encapsulation Port Numbers Internet Standards 802.3/Ethernet IP

1.7 Internet Standards The Internet is run by several standards organizations. The Internet Engineering Task Force (IETF) is responsible for near term planning and develops the specifications that become the Internet standards. The Internet Research Task Force (IRTF) is responsible for long term planning. They do not produce any standards. The Internet Architecture Board (IAB) oversees the work of the IETF and IRTF. It makes sure that all the standards are coherent and correct before final approval. The Internet Society (ISOC) is responsible for the overall growth and global reach of the Internet. It oversees the IAB.

1.7.1 Request for Comments (RFCs) All Internet related standards are published as RFCs. An RFC index indicates the latest version of an RFC. An RFC, called the IAB Official Protocol Standards, specifies the status of all the RFCs, i.e., whether or not it is a standard, and what its recommendation is.

The meaning of RFCs Each protocol has a standardization state: standard, draft standard, proposed standard, experimental, informational, or historic. Each protocol also has a requirement level: required, recommended, elective, limited use or not recommended.

Ch. 1 Introduction Layered Architectures The TCP/IP (or Internet) Architecture A Networking Example IP Addressing Packet Encapsulation Port Numbers Internet Standards 802.3/Ethernet IP

1.8 Ethernet/802.3 LAN Architecture Topology CSMA/CD Media Access Control (MAC) The Frame: Format and Structure and Uses.

1.8.1 The 802.3 and Ethernet Architectures Ethernet only consists of one single layer, the Media Access Control (MAC) layer IEEE 802.3 has two sublayers: The Media Access Control sub-layer (MAC): responsible for gaining access to the shared medium. It is topology dependent. The Logical Link Control sub-layer (LLC): responsible for reliable communications over a LAN and provides a unified interface to the network layer, i.e., hides the underlying LAN from the upper layers.

The 802.3 Layers

Slight Differences in Format Ehternet was developed by Xerox in the late 70’s and was so popular that it developed into the IEEE 802.3 standard. Ethernet and 802.3 are very similar except for the LLC sublayer and one field in the header. The length field in 802.3 is used for packet type in Ethernet. The type field in 802.3 is contained in the 8 extra bytes of header of the LLC. The IEEE 802.3 standard refers to a whole family of CSMA/CD systems that range from 1 to 10Mbps on various media.

1.8.2 Ethernet/802.3 Topologies

1.8.3 The MAC: Based on CSMA/CD A broadcast bus - stations transmit in both directions Stations sense the channel before they transmit When channel is sensed idle they transmit Stations listen while transmitting Two stations or more could start transmission at approx. the same time -> a collision results

CSMA/CD When a collision is detected, stations abort their transmission. Note only transmitting stations can detect a collision (compare signal they detect on the bus to what they are sending)! Stations involved in the collision wait a random period of time and repeat the process starting with 2 above.

Operation of CSMA/CD 1/2

Operation of CSMA/CD2/2 The vulnerable period on the bus is defined as the time from when a station transmits its first bit to the time when the furthest station on the bus has sensed the transmission (the worst case is one propagation delay on the bus, related to the two stations attached to the end points of the bus). A contention period consists of: one propagation delay for vulnerable period plus another propagation delay for the transmitting station to detect the collision (worst case!).

1.8.4 Types of CSMA/CD When channel is sensed busy, stations wait till it becomes idle and then schedule their transmission. If length of busy period is long, several stations could become active during that period and a collision becomes inevitable when the channel becomes idle again. This is called persistent CSMA/CD.  (IEEE 802.3 and Ethernet) Non persistent CSMA/CD has stations wait a random period whenever the channel is sensed busy, i.e, not grab it immediately it becomes available. Non persistent gives better throughput under heavy loading, persistent gives better throughput under light loads.

1.8.5 Frame Format

Frame Fields (1/2) Preamble is 7 bytes of 10101010 and Start of Frame is 10101011 Addresses: first bit =1 -> group address - multicast, all address bits 1’s -> broadcast to whole system, second bit =1 -> global address - set by IEEE, second bit =0 local address set by network administrator.

Frame Fields 2/2 Length field used by 802.3 standard, packet type field used by Ethernet, indicates whether network layer is IP, IPX, AppleTalk, etc. In 802.3 the type field is included in the LLC header. Data field can be anywhere from 0 to 1500 bytes. PAD field used by MAC to make ensure that minimum frame length is 64 bytes. (when data =0bytes, pad =46 or 38 bytes)

Frame Requirements Why do we need a minimum length? Minimum length of 64 bytes guarantees that frame length will be 51.2 microsecs. This is the maximum roundtrip delay that can be incurred on an 802.3 LAN operating at 10Mbps, 2.5Km long and using 4 repeaters (introduce approx. a 20 bit delay) and includes a safety factor to makeup for node detection delays. Recall that collision detection takes at least two roundtrip delays for worst case scenario! If a station transmits a shorter frame, it will have terminated its transmission before discovering that it was involved in a collision. Checksum used to detect errors (discarded)

1.8.5 Collisions When a collision is detected by the transmitting station it aborts its transmission and transmits a JAM signal. This signal is 32-48 bits long. The NIC then computes a random wait period according to the following algorithm (binary exponential backoff): After n collisions, choose K at random from {0,1,2,3,...,2m-1}, where m = min{n,10}. The node waits K x 512 bit times before scheduling a transmission.

Ch. 1 Introduction Layered Architectures The TCP/IP (or Internet) Architecture A Networking Example IP Addressing Packet Encapsulation Port Numbers Internet Standards 802.3/Ethernet IP

1.9 The Internet Protocol - IP Provides connectionless service - Datagram It has evolved over time, currently version 6 is being deployed (IPv6). The IP layer is responsible for routing over subnets. It operates end to end over a subnet. Routers are devices that operate at layer 3 and are used to interconnect subnets. They are responsible for routing.

1.9.1 IP in Operation

IP and Routers

1.9.2 IP Packet Format

1.9.3 Packet Fields 1/5 Version No.: Identifies version of IP used Header Length: Minimum 5 32 bit words (without options) Type of Service: Precedence -  Priority 0-7 D: Low Delay T: High Throughput R: High reliability

IP Packet Fields 2/5 Total length of IP Datagram: Max 65,535 bytes Identification: When datagrams are fragmented, this field is used to uniquely identify the datagram fragments Flags -3bits: First is unused, 2nd is D bit indicating that IP packet/datagram should not be fragmented, 3rd is M bit indicating whether this is last fragment of a datagram or not (more bit!).

IP Packet Fields 3/5 Fragment Offset: 13 bits used to indicate position of this fragment in the original unfragmented packet. Measured in 8 byte units. Time to live: Set by source and decremented by routers as the packet passes though them. When is reaches 0 packet is discarded. This prevents eternal looping of packets. Protocal: Indicates the Transport Layer protocol being used. E.g., UDP 17, TCP 6, OSPF 89, etc.

IP Packet Fields 4/5 Header Checksum: Adds all the 16 bit fields of the header using 1’s complement and then obtains the inverse. Can be set to all 0’s if underlying subnet uses error detection. Source IP Address: 32 bits Destination IP Address: 32 bits Options: Can be used to enhance the capability of IP. Must be a multiple of 4 bytes,if not padding used. Allows the following:

IP Packet Fields 5/5 Strict source routing: identifies path to be followed Loose source routing: identifies some nodes to be passed Record route: records all nodes it passed through Time stamping:  Nodes append address and time  packet was processed Stream Identification: indicates type of data carried, e.g., voice Data: Transport layer information

1.9.4 Addresses, Subnetting and Subnet Masks When a router is introduced into a network, a new subnet is created. Recall that each network interface on a router has a unique network (IP) address. The addresses must be on a different (sub) network. Routers and hosts use subnet masks to identify which part of the address to use for routing.

Addresses and Routing

Subnetting and Routers

Subnets and Subnet Masks The previous example showed 4 network interfaces with the following addresses: 128.3.2.3, 128.3.4.1, 128.3.1.2, 128.3.3.1 This is a class B address, 128.3 indicates the internet routing part. The 2.3, 4.2, 1.2 and 3.1 digits at the end of the address form the host and subnet address. The subnet and the host address consist of 8 bits each in this example.

Subnets and Addresses 1/2 Can you think of an example where the subnet address consists of 4 bits and the host address of 12 bits? 16.1, 32.1, 17.1 Why? 16.1 = 00010000.00000001 (subnet 1, host 1) 32.1 = 00100000.00000001 (subnet 2, host 1) 17.1 = 00010001.00000001 (subnet 1, host 257) What is the subnet mask in this case? 255.255.240.0 -  (11111111.11111111.11110000.00000000)

Subnet Masks and Addresses 2/2 New Format for Subnet Masks: 130.91.66.100/24, or 130.91.66.100/22 If we use 4 bits for subnetting how many subnets can we have? > 24=16. Note that this is true for the new Internet routing protocols (RIP II, OSPF, etc.) called classless routing algorithms. The old classfull routing algorithms did not allow all 1’s or all 0’s. New routing algortihms send subnet mask along with IP address, old algorithms relied on IP address class recognition (A, B, C, etc.)

Subnet Masks If we have the following two addresses and corresponding subnet masks: 130.91.17.17/24 and 130.91.17.49/28 How would host “1” with its subnet mask of 24 interpret the address of host “2”? Would recognize it as a host on the same subnet (subnet ID=17) with host ID 49! How would host “2” with its subnet mask of 28 interpret the address of host “1”? Would recognize it as a host on a different subnet. Its subnet ID is 275 whilst other subnet ID is 273!

Ch 1. Introduction Address Resolution Protocol Internet Control Message Protocol Ping and Traceroute

1.10 ARP 1/3 When a network device wants to transmit a packet to another device it needs to know both the IP address and the physical address of the destination. IP address to physical address mapping is maintained in the ARP cache. The ARP cache is constantly updated. Network devices can identify local and remote addresses using the subnet mask.

1.10 ARP 2/3 For a remote address the IP address of the default router  is used for looking up a physical address. (IP routing table) The IP module will look up the IP address in the ARP cache. If it does not find an entry then it uses ARP to find the physical address. The ARP module will send a broadcast message asking for the destination physical address. As all network devices read broadcasts, the destination will read the request and respond to the source with its physical address.

1.10 ARP 3/3 Note that for a remote address the router will respond to the ARP message. Routers recognize remote addresses by using subnet masks. This is called a Proxy ARP. The router will assume responsibility for all packets addressed to the remote host and forward them accordingly. In other words, the router will pretend to be the host to the ARP request and send its physical address so that all packets will be directed to it.

Ch 1. Introduction Address Resolution Protocol Internet Control Message Protocol Ping and Traceroute

1.11 Internet Control Message Protocol (ICMP) 1/3 This protocol is used by the IP layer to carry out certain functions associated with network management/status. Functions of the ICMP: Error reporting: why a datagram that was not discarded due to errors was not delivered to the destination Reachability testing: sometimes certain destinations don’t respond, an echo message is sent to which the host must respond if it is up.

1.11 ICMP 2/3 Congestion control: when datagrams are dropped because of buffer overflow, a source quench message is sent. Route-change information: when a router realizes that a host should be using a different router to reach a destination it sends a message with the updated routing information. Performance measuring: a network manager can check the time it takes to send datagrams to particular locations. Subnet addressing: a host can request for the subnet mask of its local network from the router.

1.11 ICMP 3/3 ICMP messages are carried (encapsulated) within IP datagrams. The ICMP message has a common 4 byte header: 1 byte indicates message type 1 byte indicates the code 2 bytes are used for checksum When reporting error messages, its data field contains the IP header (incl. Options) and the 8 bytes that follow the IP header (indicating TCP, or UDP, port numbers, etc.) of the message in error.

Ch 1. Introduction Address Resolution Protocol Internet Control Message Protocol Ping and Traceroute

1.12 Ping and Traceroute Ping: Packet Internet Groper (PING) Used to test whether hosts are alive The host sends an ICMP echo message and expects one back from the remote end. Traceroute Used for diagnostic purposed to report on paths that datagrams take through the network for specified: source – destination pair(s), source(s) or destination(s).