designing ip

1

Designing IP

EE 122: Intro to Communication Networks

Fall 2010 (MW 4-5:30 in 101 Barker)

Scott Shenker

TAs: Sameer Agarwal, Sara Alspaugh, Igor Ganichev, Prayag Narula

http://inst.eecs.berkeley.edu/~ee122/

Materials with thanks to Jennifer Rexford, Ion Stoica, Vern Paxsonand other colleagues at Princeton and UC Berkeley

2

Goals of Today’s Lecture

Transitioning from high-level principles to low-level design

• Review basic design decisions

• Discuss the design of IP– The “what” and “why” of the IP header

• Compare with IPv6

Review of Basic Design Decisions

3

Five Basic Design Decisions (so far)

1. Packet-switching

2. Best-effort service model

3. Layering

4. A single internetworking layer

5. The end-to-end principle (and fate-sharing)

4

1. What is Packet Switching?

• Divide messages into a sequence of packets

• Network deals with each packet individually– What is an alternative to this?

• Means that each packet must contain all relevant network information in its header– Design of protocol almost synonymous with header

5

Why Packet-Switching?

• Achieve higher levels of utilization– Statistical multiplexing– Why is this more important for the Internet than for the

phone network?

• Avoid per-flow state inside the network– Plenty of routing state, but no per-flow state– Follows from notion of fate-sharing– Enables robust fail-over

• Why not virtual circuits?– The notion of “soft-state” is midway between DG and VC6

7

2. What is “Best Effort”?

• Network makes no service guarantees– Just gives its “best effort

• Service can be imperfect–Packets may be lost–Packets may be corrupted–Packets may be delivered out of order–Packet may be significantly delayed

source destination

IP network

8

Why Best-Effort (BE)?

• BE means never having to say you’re sorry…– No need to reserve bandwidth and memory– No need to do error detection & correction– No need to remember from one packet to next– No need to make packets follow same path

• Easier to survive failures– Transient disruptions are okay during failover

• Simplifies interconnection between networks– Minimal service promises

But What About Applications?

• Some applications want more:– No losses or corruption, with packets delivered in order

o They just want the data

– Minimal and predictable delayso And they want it now!

• Is Best Effort good enough for these applications?

• Perhaps most important question Internet got right

9

10

Use Higher Layers to Compensate

• 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)– Receiver can buffer packets for smooth playout

• No network congestion control (beyond “drop”)– Sender can slow down in response to loss or delay

What Can’t Higher Layers Do?

• Higher layers cannot make delay smaller

• If applications needs guarantee of low-delay, then need to ensure adequate bandwidth– Will keep queueing delay low– No way to help with speed-of-light latency

• What applications need guaranteed low-delay?

• Can the Internet support phone calls?11

3. What is and Why Layering?

• Modularity partitions functionality into modules

• Laying is a particularly simple form of modularity

• Modules only deal with layers above and below– Simplifies interactions between modules– Simplifies introduction of new protocols

12

4. Why one networking layer protocol?

• Unifies the architecture

• As long as applications can run over IP-based protocols, they can run on any network

• As long as networks support IP, they can run any application

13

14

The Internet Hourglass

Data Link

Physical

Applications

The Hourglass Model

Waist

There is just one network-layer protocol, IP.The “narrow waist” facilitates interoperability.

SMTP HTTP NTPDNS

TCP UDP

IP

Ethernet SONET 802.11

Transport

FiberCopper Radio

Alternatives to universal IP?

• What would happen if we had more than one network layer protocol?

• Are there disadvantages to having only one network layer protocol?– Some loss of flexibility, but the gain in interoperability

more than makes up for this– Because IP is embedded in applications and in

interdomain routing, it is very hard to change– Having IP be universal made this mistake easier to

make, but it didn’t cause this problem

15

16

5. What is the End-to-End Principle?

• E2E Principle guides placement of 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

• Fate Sharing guides placement of state– Keep state on elements that rely on it, when possible

The Design of IP

17

18

Designing 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

o representation

– Semantics = interpretation of elementso information

• What semantics do we need in the IP header?

What Info Should Header Provide?

• Getting the packet there:– Where is packet going?– Which protocol will process packet on host?

• Network handling of packet:– How should the packet be forwarded (e.g., priority)– Where does header and packet end?

• Dealing with problems:– Has header been corrupted? [Why not payload?]– Has packet been fragmented?

o If so, provide information needed to reconstruct

– Is packet caught in a loop?o If so, drop packet

19

Header Info (continued)

• Extensibility: (how can we let IP change?)– Which IP version and options are expected?

• Miscellaneous:– Where did packet come from? [Why is this needed?]

20

From Semantics to Syntax

• The past two slides discussed the kinds of information the header must provide

• Will now show the syntax (layout) of the header, and discuss the semantics in more detail

21

22

5 Minute Break

Questions Before We Proceed?

IP Packet Structure

4-bitVersion

4-bitHeaderLength

8-bitType of Service

(TOS)16-bit Total Length (Bytes)

16-bit Identification3-bitFlags 13-bit Fragment Offset

8-bit Time to Live (TTL) 8-bit Protocol 16-bit Header Checksum

32-bit Source IP Address

32-bit Destination IP Address

Options (if any)

Payload

25

IP Packet Header Fields

• Version number (4 bits)– Indicates the version of the IP protocol– Necessary to know what other fields to expect– Typically “4” (for IPv4), and sometimes “6” (for IPv6)

• Header length (4 bits)– Number of 32-bit words in the header– Typically “5” (for a 20-byte IPv4 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

IP Packet Structure

4-bitVersion

4-bitHeaderLength







Options (if any)

Payload

27

IP Packet Header Fields (Continued)

• Total length (16 bits)– Number of bytes in the packet– Maximum size is 65,535 bytes (216 -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

• Must reassemble to recover original packet– Need fragmentation information (32 bits)– Packet identifier, flags, and fragment offset

28

Where Should Reassembly Happen?

• A1: router R2

1000500 500

MTU=1000BMTU=500B

MTU=1000BHost AHost B

R1

R2

MTU (Maximum Transfer Unit) = Maximum packet size handled by network

1000

29

Where does Reassemble Happen?

• A2: end-host B (receiver)

500 500

MTU=1000BMTU=500B


R1

R2


1000

30

Where does Reassemble Happen?

• A2: correct answer– Fragments can travel across different paths!

500

MTU=1000BMTU=500B


R1

R2


1000

R3

500

31

Fragmentation, con’t

• Identifier (16 bits): used to tell which fragments belong together

• Flags (3 bits):– Reserved (RF): unused bit– Don’t Fragment (DF): instruct routers to not fragment the

packet even if it won’t fito Instead, they drop the packet and send back a “Too Large” ICMP

control messageo 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

32

Example of Fragmentation

• Suppose we have a 4,000 byte datagram sent from host 1.2.3.4 to host 3.4.5.6 …

• … and it traverses a link that limits datagrams to 1,500 bytes

Version4

HeaderLength

5

Type of Service0

Total Length: 4000

Identification: 56273R/D/M0/0/0 Fragment Offset: 0

TTL127

Protocol6 Checksum: 44019

Source Address: 1.2.3.4

Destination Address: 3.4.5.6

(3980 more bytes here)

33

Example of Fragmentation (con’t)

2020

4000

39803980

2020 14801480

1500

2020 12001200

1220

2020 13001300

1320

• Datagram split into 3 pieces

• Example:

34

Example of Fragmentation, con’t

• Datagram split into 3 pieces. Possible first piece:

Version4

HeaderLength

5

Type of Service0

Total Length: 1500

Identification: 56273R/D/M0/0/1 Fragment Offset: 0

TTL127

Protocol6 Checksum: xxx



35


• Possible second piece:

Version4

HeaderLength

5

Type of Service0

Total Length: 1220

Identification: 56273R/D/M0/0/1

Fragment Offset: 185(185 * 8 = 1480)

TTL127

Protocol6 Checksum: yyy



36


• Possible third piece:

Version4

HeaderLength

5

Type of Service0

Total Length: 1320

Identification: 56273R/D/M0/0/0

Fragment Offset: 335(335 * 8 = 2680)

TTL127

Protocol6 Checksum: zzz



37

Offsets vs Numbering Fragments?

• 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

IP Packet Structure

4-bitVersion

4-bitHeaderLength







Options (if any)

Payload

39

Time-to-Live (TTL) Field (8 bits)

• Potentially lethal problem– Forwarding loops can cause packets to cycle forever– As these accumulate, eventually consume all capacity

• Time-to-live field in packet header– Decremented at each hop, packet discarded if reaches 0– …and “time exceeded” message is sent to the source

o Using “ICMP” control message; basis for traceroute

40


• Protocol (8 bits)–Identifies the higher-level protocol

o E.g., “6” for the Transmission Control Protocol (TCP)o E.g., “17” for the User Datagram Protocol (UDP)

–Important for demultiplexing at receiving hosto Indicates what kind of header to expect next

IP header IP header

TCP header UDP header

protocol=6 protocol=17

41


• Checksum (16 bits)–Complement of the ones-complement sum of all

16-bit words in the IP packet header

• If not correct, router discards packeto So it doesn’t act on bogus information

• Checksum recalculated at every router– Why?– Why include TTL?– Why only header?

IP Packet Structure

4-bitVersion

4-bitHeaderLength







Options (if any)

Payload

43

IP Packet Header (Continued)

• 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

IPv4 and IPv6 Header Comparison

Version IHL Type of Service Total Length

Identification FlagsFragment

Offset

Time to Live Protocol Header Checksum

Source Address

Destination Address

Options Padding

Version Traffic Class Flow Label

Payload LengthNext

HeaderHop Limit

Source Address

Destination Address

IPv4 IPv6

Field name kept from IPv4 to IPv6

Fields not kept in IPv6

Name & position changed in IPv6

New field in IPv6

Summary of Changes

• Eliminated fragmentation

• Eliminated header length

• Eliminated checksum

• New options mechanism (next header)

• Expanded addresses

• Added Flow Label

46

IPv4 and IPv6 Header Comparison

Version IHL Type of Service Total Length

Identification FlagsFragment

Offset

Time to Live Protocol Header Checksum

Source Address

Destination Address

Options Padding

Version Traffic Class Flow Label

Payload LengthNext

HeaderHop Limit

Source Address

Destination Address

IPv4 IPv6

Field name kept from IPv4 to IPv6

Fields not kept in IPv6

Name & position changed in IPv6

New field in IPv6

48

Next Lecture

• IP Addressing

• Read K&R: 4.4.2

• If you want to check out IPv6: K&R 4.4.4

designing ip

Documents

phone network

higher layers

delay smallerif applications

best effort good

besteffort service model

packet individuallywhat

best effortservice

relevant network information