Computer Networks Lecture 24: The TCP and UDP Headers

Summary

In this lecture, Professor Jeff Messier delves into the intricacies of TCP and UDP headers, emphasizing the importance of understanding network protocols. He explains the structure of the TCP header, including fields like port numbers, sequence numbers, and acknowledgement numbers, and discusses the role of flags in session management. Messier also contrasts TCP's reliability with UDP's simplicity and speed, highlighting UDP's preference in real-time applications like gaming. This module serves as an introduction to a detailed TCP session analysis in the subsequent lecture.

Takeaways

• 🖥️ The lecture is part of a series on computer networks, focusing on TCP and UDP headers.
• 📜 Standards documents are detailed and essential for network professionals, though this series focuses on broader concepts.
• ⚙️ The lecture dives into TCP headers, explaining the function of each field and its significance.
• 🔢 The source and destination port numbers in the TCP header identify the communication ports on the sender and receiver machines.
• 🔄 Sequence numbers track the byte index of the first byte in the TCP packet's payload, with the acknowledgment number indicating the next expected byte.
• 📏 TCP headers can vary in size due to optional control information, making the header length field essential.
• 🚩 The TCP header contains nine flag bits, each with specific roles like urgent (URG), acknowledgment (ACK), push (PSH), reset (RST), and others related to connection management.
• 📊 The TCP window size field specifies how much data can be sent before requiring acknowledgment, influencing flow control.
• 💡 UDP is a simplified version of TCP, offering minimal overhead and is ideal for applications needing fast, real-time communication like gaming.
• 🎮 UDP lacks many reliability features of TCP, such as acknowledgments and sequence numbers, making it suitable for scenarios where speed is more critical than data integrity.

Q & A

What is the primary focus of the lecture series mentioned in the transcript?

The lecture series focuses on the fundamental problems and challenges that network design must address, regardless of the specific protocol standard. The goal is to provide an enduring understanding of how networks behave and operate.

Why does the professor avoid going into extreme detail about any one protocol?

The professor avoids deep dives into specific protocols to emphasize the broader understanding of network challenges and behaviors. However, there are times when diving into the details, such as in this module on TCP and UDP headers, is necessary.

What is the purpose of the TCP header's sequence number field?

The sequence number in the TCP header represents the byte index of the first byte carried in the payload of the TCP packet. It is initialized to the initial sequence number (ISN) and is used to keep track of data bytes in the transmission.

What role does the acknowledgement number play in a TCP header?

The acknowledgement number indicates the index of the next byte that the receiver expects to receive. It confirms that all bytes up to and including the one before this number have been received successfully.

Why does the TCP header include a header length field?

The header length field is included because the TCP header supports optional control information, which can result in a variable-sized header. This field specifies the length of the header.

What is the purpose of the reserved bits in the TCP header, and why can they be a security concern?

The reserved bits in the TCP header are currently unused but reserved for future use. They can be a security concern because malware can exploit these unused fields to store and exchange covert information.

How does UDP differ from TCP in terms of functionality?

UDP is a simplified, stripped-down version of TCP that lacks features like acknowledgements and sequence numbers, making it less reliable but faster. UDP is often used in applications where real-time performance is crucial, such as gaming or time-sensitive communication.

Why might applications choose UDP over TCP for network communication?

Applications may choose UDP over TCP because UDP provides faster communication with less overhead, which is beneficial for real-time applications where delays, such as in gaming or live streaming, are unacceptable.

What is the function of the 'push' (PSH) flag in a TCP header?

The 'push' (PSH) flag in a TCP header indicates that the data packet should be immediately pushed to the application layer, rather than being buffered. This ensures timely delivery of the data.

What are some of the optional fields that can be included in a TCP header?

Optional fields in a TCP header can include the maximum segment size, window scale, selective acknowledgement (SACK) options, and timestamp values, among others. These options are used to enhance TCP's functionality and performance.

Outlines

00:00

🎓 Introduction to TCP and UDP Headers

In this introductory segment, Professor Jeff Messier from the Schulich School of Engineering outlines the focus of Module 24 in his computer networks lecture series. He explains that while the course typically avoids diving deeply into specific protocol details, this module will cover the intricacies of TCP and UDP headers. The aim is to provide a fundamental understanding of how network design operates across different protocols, which is essential for professionals in network operation or design. The professor also notes that the details discussed will be crucial for understanding the control packets exchanged during a TCP session, which will be covered in the next module.

05:01

🔍 Detailed Examination of TCP Header Fields

This paragraph delves into the specifics of the TCP header fields, starting with the source and destination port numbers, which are crucial for establishing a connection between devices. Professor Messier explains the significance of the 32-bit sequence number, which tracks the byte index in the TCP packet payload. The paragraph also touches on the importance of understanding the initial sequence number (ISN) and how TCP manages sequence numbers through wrapping. The discussion highlights that while this module focuses on these technical details, they are essential for grasping the flow of a TCP session.

10:01

📊 Acknowledgement Process and TCP Header Fields

This section covers how TCP handles acknowledgments within its standard packet headers, allowing for efficient communication between nodes. Professor Messier explains the purpose of the acknowledgment number and its role in ensuring reliable data transfer. He also introduces the header length field, which accommodates optional control information, and reserved bits, which are typically unused but can pose security risks. The paragraph further details the various flag bits, such as the urgent (URG), acknowledgment (ACK), push (PSH), reset (RST), synchronization (SYN), and finish (FIN) flags, which control different aspects of TCP connections. The window size, checksum, and urgent pointer fields are also discussed for their roles in managing TCP sessions.

🔐 Security Concerns and Optional TCP Header Fields

Professor Messier highlights security concerns related to unused TCP header fields, which can be exploited by malware. He discusses optional TCP header fields that can extend up to 352 bits, serving various purposes such as defining the maximum segment size and enabling selective acknowledgment (ARQ). The paragraph also covers the window scale option, which allows for larger TCP window sizes, essential for high-throughput applications. Additionally, timestamp values are mentioned, which are occasionally used in TCP communications.

⚡ Introduction to UDP: A Lightweight Alternative to TCP

This final paragraph introduces the User Datagram Protocol (UDP), a simplified version of TCP that sacrifices reliability for speed. Professor Messier explains that while TCP provides various mechanisms for reliable communication, these can introduce delays, which are undesirable in certain applications like real-time gaming or time-sensitive data transmission. UDP, with its minimalistic header containing only port numbers, packet length, and a checksum, is preferred in scenarios where speed is more critical than reliability. The paragraph concludes by emphasizing UDP's relevance in specific applications, despite its lack of complex session management features.

Keywords

💡TCP Header

A TCP header is a part of the Transmission Control Protocol that includes various fields used to manage the transmission of data between computers. In the video, the TCP header is explained in detail, with focus on its fields like port numbers, sequence numbers, and flags, which are essential for ensuring reliable communication in network protocols.

💡UDP

UDP (User Datagram Protocol) is a simpler communication protocol compared to TCP, primarily used for applications that require fast, efficient transmission without the need for reliability, like online gaming or streaming. The video contrasts UDP with TCP, highlighting that UDP lacks many of TCP's features, such as acknowledgments and sequence numbers, making it faster but less reliable.

💡Port Numbers

Port numbers are used in both TCP and UDP to identify specific processes or network services on a device. The video describes how port numbers are integral to both protocols, with source ports identifying the sender's application and destination ports identifying the receiving service, like a web server using port 80.

💡Sequence Number

The sequence number in a TCP header indicates the byte index of the first byte in the packet's data segment, helping to ensure data is reassembled correctly. The video discusses how sequence numbers are crucial for tracking the order of packets, particularly in ensuring the reliable transmission of data over TCP.

💡Acknowledgment Number

The acknowledgment number is part of the TCP header and indicates the next byte that the receiver expects to receive, confirming the successful reception of data. The video emphasizes its role in TCP's reliable delivery mechanism, where it allows the sender to know which data has been successfully received.

💡Checksum

A checksum is a field in both TCP and UDP headers used for error-checking. It helps ensure that data has not been corrupted during transmission. The video mentions how the checksum is calculated and embedded in the packet header, allowing the receiving device to verify data integrity.

💡Sliding Window

The sliding window is a flow control mechanism in TCP that determines how much data can be sent before receiving an acknowledgment. The video explains how the window size, which is a field in the TCP header, adjusts dynamically to control the flow of data and prevent network congestion.

💡Connection Setup and Termination

Connection setup and termination in TCP involve a three-way handshake (using SYN and ACK flags) and a four-step connection teardown (using FIN and ACK flags). The video introduces these concepts as part of understanding how TCP manages the beginning and end of a communication session between two nodes.

💡Flags

Flags in the TCP header are binary indicators used to manage various aspects of the TCP connection, like SYN for initiating a connection and FIN for terminating it. The video provides an overview of the different flags and their purposes, explaining how they control the flow of data and manage connections.

💡Standards Documents

Standards documents refer to detailed technical specifications that define protocols like TCP and UDP. The video mentions these documents as essential reading for professionals in network design or operation, as they provide in-depth information on packet formats, header fields, and other protocol specifics.

Highlights

Introduction to TCP and UDP headers by Professor Jeff Messier.

Discussion on the importance of understanding network protocols for network operation or design.

Emphasis on the fundamental problems and challenges in network design rather than deep-diving into specific protocols.

Explanation of the value of occasionally diving into the details of protocols, such as TCP session control packets.

Introduction of the TCP header fields, starting with the source and destination port numbers.

Description of the 32-bit sequence number and its role in TCP packet payload.

Clarification of the acknowledgement number in TCP and its significance in data packet order.

Explanation of the header length field due to the variable size of the TCP header.

Mention of reserved bits in TCP header and their potential security implications.

Description of the flag bits in TCP for controlling the flow and state of a connection.

Function of the window size field in TCP for managing data flow.

Role of the checksum field in TCP for error detection.

Brief introduction to the User Datagram Protocol (UDP) and its differences from TCP.

UDP's simplicity and use cases, such as real-time applications and gaming.

UDP header's structure, including source and destination port numbers, total length, and checksum.

Comparison of TCP's reliability and overhead with UDP's speed and simplicity.

Highlight of the practical applications of TCP and UDP in network communication.

Transcripts

hello my name is jeff messier i'm a

professor in electrical and computer

engineering in the schulich school of

engineering

and this is module 24 in my computer

networks lecture series

where i talk about tcp and udp

headers now you'll have noticed and i

have discussed

um in the past during this lecture

series that

you know i don't tend to go into any one

protocol in a huge amount of detail so

if you do find yourself making a living

in

network operation or network design at

some point you will

have to become very sort of intimately

familiar with a particular communication

standard maybe

you're provisioning ip networks and you

need to sink yourself into what ip is

all about

maybe you have a job with a cellular

provider and you need to learn a

particular cellular standard

regardless of what it is at some point

you're going to have to you're going to

find yourselves

being

having the opportunity to read a

standards document and standards

documents tend to be very long

very detailed and very full of

things like header specifications

packet formats all that kind of stuff

and

i try not to get into that level of

detail too much

in this lecture series because instead i

want to focus on kind of the

fundamental problems and challenges

that network design has to contend with

regardless of whether we're looking at

our specific wireless standard or

particular wired

standard and so my hope is that what

you're going to get out of this lecture

series

is sort of kind of the enduring

understanding of how

networks behave and operate regardless

of

the particular protocol standard that

you're working with

that being said every once in a while

there is value to diving down into the

details and that's what i'm going to do

here

so this lecture module is kind of a

little bit of a

intro or companion module to module 25

which is where we look at a

tcp session in detail so we're going to

look

at all of the control packets that get

sent back and forth when

setting up and tearing down a tcp

session between two nodes

and in order to understand all of the

detail that we're going to be looking at

we do

need to examine specifically

the different fields that you find in a

tcp

header and what they all mean a lot of

it is going to be

kind of just showing you

where certain things are implemented

that you already understand so things

like port numbers for example

we've talked about what port numbers are

you'll just see where they fit into the

the tcp header

and tacked on to the end of this

discussion i'm also going to introduce

very briefly the udp standard

talk about how it differs from tcp and

what its header looks like as well

okay so let's dive right in so to start

off

the first four

fields in the tcp header the fir the

first are given in the slide the first

two

are the port numbers so we have a source

port number

and a destination port number

port numbers are 16 bits as we've

already discussed

and the difference between these two

values is

the source port is the port number on

the machine where the packet is being

generated

the destination port is the port number

we want to connect to on our destination

machine so for example if our computer

was connecting to a web server the

destination port

would be port 80 and the source port

would be whatever randomly generated

port number

our operating system decided to assign

to this particular tcp connection

the next field we have is the 32-bit

sequence number

and as we've talked about the sequence

number is basically the

byte index of the first byte carried in

the payload of

our tcp packet and

we initialize the sequence number to the

initial sequence number or isn't

the isn can be basically anything

but quite often we

[Music]

will see zero used as the isn

and the sequence number wraps around so

it wraps to zero after it reaches

um two to the 32 minus one

and that's no problem because we use

windows

and so we're never going to have more

packets outstanding than 2 to the 32 and

so the wrap around doesn't cause us any

problems

the next field is the acknowledgement

number and the way tcp works

is we

don't have like a special dedicated

header

or packet type for acknowledgements

instead the acknowledgement information

is just embedded in the standard tcp

packet header and that's so to allow

us to kind of piggyback acknowledgement

information on a data frame that's going

in the opposite direction so typically

data frames are being sent back and

forth between the two nodes

this just allows us to piggyback

acknowledgments

sometimes however frames will be sent

with

no data payload and with just

acknowledgement information

in the frame as we're gonna see and of

course

as we've seen in our our previous

modules the

um the value of the act number is

basically the index

of the byte that the receiver is

expecting to

receive next and so that means

everything

up to and including one minus the

acknowledgement value has been

received successfully in order by the

receiver

the next field we have is header length

the reason why we have a header length

field is because the tcp header does

support

optional sort of control information

that might be present in the header or

it might not and as a result we can have

a variable sized header

so

following the header length we have

six reserved bits that are used for

nothing

and you know it's it's fine to have

reserved bits i i guess

on on some levels but as we're going to

see this is this can be kind of a

dangerous thing from a security point of

view because when we

when we talk about security we'll see

that you know sometimes

malware traffic that that's trying to um

you know covertly or secretly exchange

information

will sometimes intentionally set things

like these reserved bits to particular

values

so that other malware programs can

recognize

malware traffic following the reserve

bits we have

nine flag bits the first

three non-cdwr and ecn echo we're not

really going to talk about very much

they're used for congestion control

the next

flag is the urg or urgent

flag when this is set the urgent pointer

is

valid and we'll talk about the urgent

pointer in a second this is

basically meant to

[Music]

sort of expedite or prioritize

this packet being delivered to

the the higher layers of of the protocol

stack but it

it's not really used anymore either

the ack flag is definitely used so when

the act flag is set that means

that the header contains valid ack

information

if the act flag is zero then the header

still contains an

ack number but that act number is not

not considered valid there's psh

or the push flag the push flag if this

is set

then it means that the data packet

should be immediately

pushed up to the application layer

sometimes different tcp implementations

will buffer several packets together

before sending them up to the

application layer but the

the push flag forces that applicator or

forces the

the um the frame to get pushed up to the

application

the reset flag

is if this is set um we abort the

connection due to some sort of abnormal

conditions that may have occurred

and the sin and the fin flags as we're

going to see

are used for connection setup and

connection termination and we're going

to see those featured very prominently

in the next module where we go through

our tcp session example

finally we have our window size

and we've talked about that already as

well this is the number of bytes

the center of the package is willing to

accept before

a an acknowledgment needs to be sent so

this is the sliding window act field

the checksum field

holds the internet checksum that we've

studied when back when we were looking

at error detection

and then following the checksum we have

the urgent pointer

and if the urgent or urg flag is set

then this urgent pointer value was meant

to be used to send the

the data in this tcp packet quickly up

to the application

using kind of a side channel provided by

the operating system

rather than just going through the

regular sort of protocol stack

socket interface and

this isn't really used anymore because

it's not really super compatible across

multiple operating systems

and as a result this is another bit of a

security concern in tcp

so whenever you have a field that isn't

used again as i was saying malware can

use this use these unused fields to

store

to store values

there are going to be a few of the

optional

tcp header fields that we're gonna see

options can be contained

or the the options field can stretch

from zero all the way up to 352 bits

and there's a whole bunch of different

purposes but the more common ones

will be is related first of all

to the maximum segment size

or packet size that the sender will

accept

there is um a field

that indicates whether or not selective

arq

is allowed this window scale

is a very common field so if we go back

we can see that the window

size is only a 16 bit number

and that's often too small sometimes

especially for like really sort of

high data rate high throughput

applications

an application will want a larger tcp

window than what can be contained in 16

bits

and so this windows scale field scales

up the window by a factor of two

raised to the exponent of whatever is in

the windows scale

we then also have some timestamp

values that are used

as well that sometimes show up

the user datagram protocol or udp

you can think of as a super super

stripped down version of tcp

so tcp provides all this functionality

the

the acknowledgements the

sequence numbers

all kinds of things to make sure that

the connection

is very reliable

and looking from the user perspective

however

all of this overhead and this

calculation does tend to slow the

connection down a little bit and

sometimes

applications just want very simple

very raw communication and

when that's the case they tend to opt

for udp

udp tends to be used for

things that send very simple

communication packets and

things that tend to not want

very out of date information so

one thing about you know the selective

repeat

arq is that you know packets can get

buffered

and they are delivered reliably but at

the cost of increased delay so for an

application like a

an internet time server for example you

just want to get that packet out there

and if it gets sort of um

caught up in a on a congested link you

don't really want

re-transmission of old-time information

you would just prefer that

packet gets lost and then you know have

a more up-to-date packet filter through

again

at some point many games many network

games will also use

udp because you want that real-time game

game information

you know if you um

you know if a packet is delivered

reliably but late

then you start to get some lag in your

gameplay and that's

generally seen as unacceptable for a lot

of game players and so udp

is basically just a super super simple

thing that gets bolted on top

of an ip packet that essentially

provides port numbers and not much

else so if we look at the udp

header it's super simple we've got our

source port number

our destination port number

the total length of the udp

packet and the internet checksum is

included as well

and that's it so again no

acknowledgments nothing like that

so it's important to be aware of udp

because many of you if you develop

application software will be

using udp again particularly for sort of

gaming and kind of real time

type applications however from a

protocol perspective

and a session management perspective

there's really not much to it