Goal of Today s Lecture EE 122: Designing IP Ion Stoica TAs: Junda Liu, DK Moon, David Zats http://inst.eecs.berkeley.edu/~ee122/ (Materials with thanks to Vern Paxson, Jennifer Rexford, and colleagues at UC Berkeley) Work through process of designing IP, the Internet s (sole) network-layer protocol 1 2 Our Story So Far (Context) The Internet Hourglass The Internet uses packet-switching rather than circuitswitching in order to Achieve higher levels of utilization (we can use statistical multiplexing to more aggressively share network links) Avoid state inside the network (robust fail-over) Make interconnection between different parties easy (minimal service promises) The Internet architecture uses layering to partition functionality into modules The internetworking layer (or just network layer) forms the waist of the layering hourglass 3 SMTP HTTP TCP IP DNS UDP NTP Ethernet SONET 82.11 Copper Fiber Waist Radio Applications Transport Data Link Physical The Hourglass Model There is just one network-layer protocol, IP. The narrow waist facilitates interoperability. Our Story So Far (Context), Con t The End-to-End Principle guides us in where to place functionality If hosts can implement functionality correctly, implement it in a lower layer only as a performance enhancement But do so only if it does not impose burden on applications that do not require that functionality The principle of Fate Sharing guides us to keep state with the elements that rely on it, when possible IP Service: Best-Effort Packet Delivery source Packet switching Divide messages into a sequence of packets Each packet (datagram) is dealt with individually Best-effort delivery Packets may be lost Packets may be corrupted Packets may be delivered out of order IP network destination 1
IP Service Model: Why Best-Effort? IP means never having to say you re sorry Don t need to reserve bandwidth and memory Don t need to do error detection & correction Don t need to remember from one packet to next Easier to survive failures Transient disruptions are okay during failover but, applications do want efficient, accurate transfer of data in order, in a timely fashion IP Service: Best Effort Suffices No error detection or correction Higher-level protocol can provide error checking Successive packets may not follow the same path Not a problem as long as packets reach the destination Packets can be delivered out-of-order Receiver can put packets back in order (if necessary) Packets may be lost or arbitrarily delayed Sender can send the packets again (if desired) No network congestion control (beyond drop ) Sender can slow down in response to loss or delay 7 8 Let s Design IP What does it mean to design a protocol? Answer: specify the syntax of its messages and their meaning (semantics). Syntax = elements in packet header, their types & layout representation Semantics = interpretation of elements information For IP, what fields do we need & why? Information to Capture in IP Addresses: datagram destination & source Framing: datagram length Priority: any special forwarding? Extensibility: what if we need to tweak/change IP? Dealing with problems: Integrity: is the header what it s supposed to be? Loop avoidance: make sure packets don t endlessly circulate Fragmentation: what if the datagram is too large? 9 1 Minute Break Questions Before We Proceed? IP 1-bit Total (Bytes) 1-bit Checksum 11 2
1-bit Total (Bytes) 1-bit Checksum IP Packet Fields number ( bits) Indicates the version of the IP protocol Necessary to know what other fields to expect Typically (for IPv), and sometimes (for IPv) length ( bits) Number of 32-bit words in the header Typically (for a 2-byte IPv header) Can be more when IP options are used Type-of-Service (8 bits) Allow packets to be treated differently based on needs E.g., low delay for audio, high bandwidth for bulk transfer 1 1-bit Total (Bytes) 1-bit Checksum IP Packet Fields (Continued) Total length (1 bits) Number of bytes in the packet Maximum size is,3 bytes (2 1-1) though underlying links may impose smaller limits Fragmentation: when forwarding a packet, an Internet router can split it into multiple pieces ( fragments ) if too big for next hop link Fragmentation information (32 bits) Packet identifier, flags, and fragment offset 1 Where does Reassemble Happen? Where does Reassemble Happen? Host A MTU=B R1 R2 Host B Host A MTU=B R1 R2 Host B MTU (Maximum Transfer Unit) = Maximum packet size handled by network MTU (Maximum Transfer Unit) = Maximum packet size handled by network A1: router R2 A2: end-host B (receiver) 17 18 3
Where does Reassemble Happen? Host A R1 MTU=B R3 R2 Host B 1-bit Total (Bytes) RF/DF/ MF 1-bit Checksum MTU (Maximum Transfer Unit) = Maximum packet size handled by network A2: correct answer Fragments can travel across different paths! 19 Fragmentation (con t) Identifier (1 bits): used to tell which fragments belong together (3 bits): Reserved (RF): unused bit (why reserved?) Don t Fragment (DF): instruct routers to not fragment the packet even if it won t fit Instead, they drop the packet and send back a Too Large ICMP control message Forms the basis for Path MTU Discovery, covered later More (MF): this fragment is not the last one Offset (13 bits): what part of datagram this fragment covers in 8-byte units Thus, a fragment has either MF set or Offset > 21 Example of Fragmentation Suppose we have a, byte datagram sent from host 1.2.3. to host 3... Identification: 273 Source Address: 1.2.3. Total : // Fragment Offset: Destination Address: 3... (398 more bytes here) Checksum: 19 and it traverses a link that limits datagrams to 1, bytes 22 Example of Fragmentation (con t) Datagram split into 3 pieces Example: Example of Fragmentation (con t) Possible first piece: 2 398 Identification: 273 Total : 1 (2 + 18 = 1) //1 Fragment Offset: 2 18 2 12 2 13 Source Address: 1.2.3. Checksum: xxx 1 122 132 Destination Address: 3... 23 2
Example of Fragmentation (con t) Possible second piece: Example of Fragmentation (con t) Possible third piece: Identification: 273 //1 Total : 122 (2 + 12 = 122) Fragment Offset: 18 (18 * 8 = 18) Identification: 273 Total : 132 (2 + 388 28 = 132) // Fragment Offset: 33 (33 * 8 = 28) Checksum: yyy Checksum: zzz Source Address: 1.2.3. Destination Address: 3... Source Address: 1.2.3. Destination Address: 3... 2 2 Some Fragmentation Design Decisions Q: where are fragments reassembled? A: usually at the final destination. Why? Because different fragments can take different paths through the network - the whole collection might only be available at the receiver Q: why use a byte-offset for fragments rather than a numbering each fragment? Ans #1: with a byte offset, the receiver can lay down the bytes in memory when they arrive Ans #2 (more fundamental): allows further fragmentation of fragments 27 1-bit Total (Bytes) 1-bit Checksum Time-to- Field (8 bits) Potentially lethal problem Forwarding loops can cause packets to cycle forever As these accumulate, eventually consume all capacity IP Packet Fields (cont d) (8 bits) Identifies the higher-level protocol E.g., for the Transmission Control (TCP) E.g., 17 for the User Datagram (UDP) Important for demultiplexing at receiving host Indicates what kind of header to expect next Time-to-live field in packet header Decremented at each hop, packet discarded if reaches and time exceeded message is sent to the source Using ICMP control message; basis for traceroute 29 protocol= IP header TCP header protocol=17 IP header UDP header 3
IP Packet Fields (cont d) One s Complement Checksum (1 bits) Complement of the one s-complement sum of all 1-bit words in the IP packet header Each router computes ones-complement sum of entire header including checksum should get (or xffff) If not, router discards packet as corrupted So it doesn t act on bogus information 31 Assume numbers Numbers starting with are positive E.g., 1111 1; Numbers starting with 1 are negative; negative number is obtained by inverting all bits of the positive number E.g., 111-1 and 11111111 both represent Addition: add carry-on to result 1111 (1) + 1111111 (-8) 1 11 + 1 111 (7) 32 Checksum Example 11111 11 111 111 111 11 + 1 1 111 11 + 111 1 1 111 11 +1 111 111 + 11 11 1 11 +1 11 11 data (header) complement checksum 1111 33 1-bit Total (Bytes) 1-bit Checksum IP Packet (cont d) Two IP addresses Source IP address (32 bits) Destination IP address (32 bits) Destination address Unique identifier/locator for the receiving host Allows each node to make forwarding decisions Source address Unique identifier/locator for the sending host Recipient can decide whether to accept packet Enables recipient to send a reply back to source Next Lecture IP Addressing & Forwarding Read K&R:..2 If you want to check out IPv: K&R.. 3 3