TCP - Transmission Control Protocol

What is TCP?

TCP stands for Transmission Control Protocol. It’s one of the core protocols of Internet Protocol(IP) suite. TCP ensures reliable, ordered, and error-checked delivery of data between applications running on devices connected to a network.

Why is TCP useful?

Without TCP, data sent over the internet could arrive out of order, get lost, or be duplicated. So TCP solves this by * Breaking data into Packets. * Ensuring all packets arrive correctly and in order. * Retransmitting lost packets. * This makes it ideal for applications where accuracy and reliability are critical like web browsing, email.

How it works?

  • Connection setup (3-way handshake) – The client and server exchange messages to establish a connection.

  • Data transfer – Data is broken into packets, sent, acknowledged, and reassembled in the correct order.

  • Error checking – TCP checks for errors and requests retransmission if needed.

  • Connection termination – Once data transfer is complete, the connection is closed gracefully.

Where is TCP used?

  • Web browsing – HTTP and HTTPS use TCP to load websites reliably.

  • Email – Protocols like SMTP, IMAP, and POP3 rely on TCP.

  • File transfers – FTP and other file-sharing tools use TCP for accurate data delivery.

  • Remote access – SSH and Telnet use TCP to ensure secure and reliable sessions.

Why OSI Layer: Transport Layer (Layer 4)?

  • It manages end-to-end communication between devices.

  • It handles segmentation, flow control, error correction, and retransmission.

  • It provides a reliable communication channel for applications running at the Application Layer (Layer 7).

What are the key features?

  • Connection-oriented.

  • Reliable data transfer.

  • Error detection and recovery.

What is a three-way handshake in TCP?

A process used to establish a TCP connection: * SYN – Client sends a synchronize request. * SYN-ACK – Server acknowledges and synchronizes. * ACK – Client acknowledges server’s response.

What is a TCP segment?

A TCP segment is the unit of data sent over a TCP connection. It consists of: * TCP header (including sequence number, acknowledgment number, flags, etc.). * Data payload.

How is retransmission handled in TCP?

If an ACK is not received within the timeout interval, TCP retransmits the segment. Modern TCP uses: * Timeout-based retransmission. * Fast retransmit (based on duplicate ACKs).

What are some applications that use TCP?

  • HTTP/HTTPS (Web).

  • FTP (File Transfer).

  • SMTP (Email).

  • Telnet / SSH (Remote Access).

What is the working flow of TCP?

  • Connection Establishment (3-Way Handshake): * Client sends SYN → Says I want to start a connection. * Server replies with SYN-ACK → Says Okay, I got it. Let’s sync. * Client sends ACK → Says Confirmed. Let’s communicate.

  • Data Transfer: * Data is broken into segments with sequence numbers. * Each segment is sent from sender to receiver. * Receiver sends ACKs back to confirm receipt. * TCP uses a sliding window to send multiple packets before waiting. * If data is lost or corrupted, TCP retransmits it. * Flow control makes sure the sender doesn’t overload the receiver. * Congestion control adjusts speed based on network traffic.

  • Connection Termination (4-Way Handshake): * Sender sends FIN → Says I’m done sending data. * Receiver sends ACK → Says Okay, I got it. * Receiver sends FIN → Says I’m done too. * Sender sends ACK → Says Goodbye.

What is the difference between a port and an IP address?

  • IP address identifies a host in the network.

  • Port number identifies a specific process/service on that host. Together (IP + Port) = Socket.

Topics in this section,

  • In this section, you are going to learn

  • Terminology

  • Version Info

TCP Version

RFC

Year

Core Ideas/Contribution

TCP v1

RFC 675

1974

First specification of TCP,combined with IP; intial handshake

concept introduced .

TCP v2

RFC 761

1980

Defined TCP separately from IP and updated segment formats.

TCP v3(stable TCP)

RFC 793

1981

The foundational TCP specifications; still forms the backbone

of modern TCP implementations.

TCP Performance Extensions

RFC 1323

1992

Intoduction of Window scaling,timestamps and PAWS.

TCP SACK

RFC 2018

1996

Selective retransmission based on ACKs of received blocks.

TCP Congestion Control

Algorithm

RFC2581

1999

Defined algorithms like slow start, fast retransmit , congestion

avoidance.

TCP ECN Support

RFC3168

2001

Enabled routers to signal congestion without packet loss.

Modern Congestion Control

update

RFC5681

2009

superseded RFC2581 with more robust congestion recovery

mechanisms.

TCP Fast open (TFO)

RFC7413

2014

Reduced Web Latency; improved Performance for short,

repeated TCP connections

Updated TCP specification

RFC9293 (RFC 793bis)

2021

clarified outdated parts of RFC 793; incorporated decades of

extensions and best practices.

  • setup

  • setup

TCP SYN Packet

S.No

Protocol Packets

Description

Size(bytes)

1

SYN Packet

Used to initiate a TCP connection. It is part of the

18

three-way handshake process.

Source Port

port number of the sender(Client)

2

Destination Port

port number of the receiver

2

Sequence Number

Initial sequence number

4

Acknowledgment Number

0(not set)

0

Data Offset

4-bit field that specifies size of TCP header in

1

32-bit words.

Reserved

3 bits reserved for future purpose

1

Flags

SYN flag set

2

Window Size

Size of the sender’s receive window.

2

Checksum

Error-checking field.

2

Urgent Pointer

points to the urgent data if the URG flag is set.

2

Options

May include options like Maximum Segment Size,Window scaling,Timestamps.

variable

Data(Payload)

does not carry any payload data.

0

TCP SYN-ACK Packet

s.no

Packet details

descrption

size(Bytes)

2

SYN-ACK Packet

Sent by the receiver to acknowledge the SYN

22

packet and establish the connection.

Source Port

port number of the receiver

2

Destination Port

port number of the sender(Client)

2

Sequence Number

server’s ISN(Initial sequence number)

4

Acknowledgment Number

Client’s ISN+1

4

Data Offset

4-bit field that specifies size of TCP header in

1

32-bit words.

Reserved

3 bits reserved for future purpose

1

Flags

SYN=1, ACK=1

2

Window Size

Size of the receiver’s receive window.

2

Checksum

Error-checking field.

2

Urgent Pointer

points to the urgent data if the URG flag is set.

2

Options

May include options like Maximum Segment Size,Window scaling,Timestamps.

variable

(MSS).

Data(Payload)

does not carry any payload data.

0

TCP ACK Packet

s.no

Packet details

descrption

size(bytes)

3

ACK Packet

Used to acknowledge received data or control packets.

22

Source Port

port number of the sender(Client)

2

Destination Port

port number of the receiver

2

Sequence Number

next byte the sender expects to receive.

4

Acknowledgment Number

Sequence number of the last received byte + 1.

4

Data Offset

4-bit field that specifies size of TCP header in

1

32-bit words.

Reserved

3 bits reserved for future purpose

1

Flags

ACK=1

2

Window Size

Size of the Sender’s receive window.

2

Checksum

Error-checking field.

2

Urgent Pointer

points to the urgent data if the URG flag is set.

2

Options

May include options like Maximum Segment Size,Window scaling,Timestamps.

variable

(MSS).

Data(Payload)

can carry payload data if it is piggybacking on data being sent.

variable

TCP SUBSCRIBE Packet

TCP PSH-ACK Packet

s.no

packet details

descrption

size(bytes)

4

PSH-ACK Packet

Used to push data immediately to the receiving

22

application and acknowledge the receipt of data.

Source Port

port number of the sender(Client)

2

Destination Port

port number of the receiver

2

Sequence Number

sequence number of the first byte of data in the current

4

message.

Acknowledgment Number

sequence number of the next byte that the sender is

4

expecting to receive.

Data Offset

4-bit field that specifies size of TCP header in

1

32-bit words.

Reserved

3 bits reserved for future purpose

1

Flags

PSH=1, ACK=1

2

Window Size

Size of the Sender’s receive window.

2

Checksum

Error-checking field.

2

Urgent Pointer

points to the urgent data if the URG flag is set.

2

Options

May include options like Maximum Segment Size,Window scaling,Timestamps.

variable

(MSS).

Data(Payload)

carries payload data that needs to be processed immediately by the receiving application

variable(1460)

TCP FIN Packet

s.no

packet details

descrption

size(Bytes)

5

FIN Packet

Used to terminate a TCP connection.

22

Source Port

port number of the sender(Client)

2

Destination Port

port number of the receiver

2

Sequence Number

Sequence number of the last byte sent.

4

Acknowledgment Number

Acknowledges received data

4

Data Offset

4-bit field that specifies size of TCP header in

1

32-bit words.

Reserved

3 bits reserved for future purpose

1

Flags

FIN=1

2

Window Size

Size of the Sender’s receive window.

2

Checksum

Error-checking field.

2

Urgent Pointer

points to the urgent data if the URG flag is set.

2

Options

May include options like Maximum Segment Size,Window scaling,Timestamps.

variable

(MSS).

Data(Payload)

does not carry any payload data.

0

TCP FIN-ACK Packet

S.No

Packets details

Description

Size(bytes)

6

FIN-ACK Packet

Sent to acknowledge the receipt of a FIN packet and

22

signal the end of data transmission.

Source Port

port number of the sender(Client)

2

Destination Port

port number of the receiver

2

Sequence Number

Sequence number of the last byte sent.

4

Acknowledgment Number

sequence number of the next byte that the sender is expecting to receive

4

Data Offset

4-bit field that specifies size of TCP header in

1

32-bit words.

Reserved

3 bits reserved for future purpose

1

Flags

FIN=1,ACK=1

2

Window Size

Size of the Sender’s receive window.

2

Checksum

Error-checking field.

2

Urgent Pointer

points to the urgent data if the URG flag is set.

2

Options

May include options like Maximum Segment Size,Window scaling,Timestamps.

variable

(MSS).

Data(Payload)

does not carry any payload data.

0

TCP RST Packet

s.no

packet details

descrption

size(bytes)

7

RST Packet

Used to reset a TCP connection.

22

Source Port

port number of the sender(Client)

2

Destination Port

port number of the receiver

2

Sequence Number

Sequence number of the last byte sent.

4

Acknowledgment Number

Typically set to 0, as the connection is being reset.

4

Data Offset

4-bit field that specifies size of TCP header in

1

32-bit words.

Reserved

3 bits reserved for future purpose

1

Flags

RST=1

2

Window Size

Size of the Sender’s receive window.

2

Checksum

Error-checking field.

2

Urgent Pointer

points to the urgent data if the URG flag is set.

2

Options

May include options like Maximum Segment Size,Window scaling,Timestamps.

variable

(MSS).

Data(Payload)

does not carry any payload data.

0

TCP-Use Cases

S.No

Use Case

Descrption

1

Web Browsing(HTTP/HTTPS)

Ensures reliable delivery of web pages & resources over protocols.

2

Email(SMTP, IMAP,POP3)

Guarantees ordered and complete message delivery for sending

and retrieving emails.

3

File Transfer(FTP,SFTP)

Reliable and ordered transmission of large files over networks.

4

Data Base Access

Maintains consistent & reliable connections for updating data base.

5

VoIP with TCP fallback

Some VoIP appilications use TCP for control/data fall back.

6

IOT Device Control

Requiring reliable communication between IOT devices .

S.No

Feature

Descrption

1

Connection-oriented

Enables a reliable communication b/w sender & receiver.

2

Reliable Data Transfer

Ensures data is delivered accurately and in order using ACks and

retransmissions.

3

Error Detection

Uses checksum to detect corrupted data segments in transmission.

4

Flow Control

Prevents overwhelming receiver using the sliding Window protocol.

5

Port Addressing

Uses port numbers to identify sending and receiving appilications on

hosts.

6

Graceful Connection Termination

Closes connections cleanly using a 4-step termination process.

7

Congestion Control

Adjusts transmission rate based on network congestion using

algorithms like slow start and congestion avoidance.

8

Stream-Oriented

Treats data as a continous byte stream rather than discrete packets.

9

Ordered Delivery

Reassembles segments in the correct order before passing data to

the appilication layer.

Connection Oriented - Testcases

S.No

Test Case

Description

Expected Result

1

TCP 3-Way Handshake

Initiate connection

SYN, SYN-ACK, ACK exchanged

2

TCP Connection Establishment

Client connects to server

Connection established

3

TCP Connection Termination

Close connection

FIN, ACK exchanged

4

TCP Connection Reset

Send RST

Connection reset immediately

5

Duplicate SYN Handling

Send duplicate SYN

Handshake continues correctly

6

Delayed ACK Handling

Delay ACK response

Connection remains stable

7

Retransmission on Timeout

Drop ACK

Packet retransmitted

8

Sequence Number Validation

Check sequence numbers

Numbers increment correctly

9

Acknowledgment Number Validation

Check ACK numbers

ACKs match expected values

10

Window Size Negotiation

Exchange window size

Flow control established

11

Congestion Control Start

Start slow start

Congestion window increases exponentially

12

Congestion Avoidance

Enter avoidance phase

Window increases linearly

13

Fast Retransmit

Trigger triple duplicate ACKs

Packet retransmitted

14

Fast Recovery

Recover after fast retransmit

Congestion window adjusted

15

TCP Keepalive

Enable keepalive

Idle connection maintained

16

TCP Timeout

Idle beyond timeout

Connection closed

17

Half-Open Connection

Client closes, server doesnt

Server detects and closes

18

Simultaneous Open

Both sides send SYN

Connection established

19

Simultaneous Close

Both sides send FIN

Connection closed cleanly

20

TCP Port Unreachable

Connect to closed port

RST received

21

TCP with NAT

Use NAT device

Connection maintained

22

TCP with Firewall

Block port

Connection fails

23

TCP with VPN

Use VPN tunnel

Connection established

24

TCP with Packet Loss

Drop packets

Retransmission occurs

25

TCP with Packet Reordering

Reorder packets

Data reassembled correctly

26

TCP with Duplicate Packets

Send duplicates

Duplicates discarded

27

TCP with Fragmentation

Send large data

Data reassembled correctly

28

TCP with MSS Negotiation

Negotiate MSS

Data sent within limits

29

TCP with Window Scaling

Enable scaling

Large window supported

30

TCP with Selective ACK

Enable SACK

Only missing segments retransmitted

31

TCP with Timestamps

Enable timestamps

RTT measured accurately

32

TCP with Delayed ACK

Enable delayed ACK

ACKs sent after delay

33

TCP with Nagles Algorithm

Enable Nagle

Small packets coalesced

34

TCP with Push Flag

Set PSH flag

Data pushed immediately

35

TCP with Urgent Pointer

Use URG flag

Urgent data processed

36

TCP with Zero Window

Set window to 0

Sender pauses transmission

37

TCP Window Update

Increase window size

Sender resumes transmission

38

TCP with High Latency

Simulate delay

Connection remains stable

39

TCP with High Bandwidth

Send large data

Connection handles throughput

40

TCP with Multiple Streams

Open multiple connections

All connections maintained

41

TCP with TLS

Wrap in TLS

Secure connection established

42

TCP with IPv6

Use IPv6

Connection established

43

TCP with IPv4

Use IPv4

Connection established

44

TCP with SYN Flood

Send many SYNs

Server handles or drops

45

TCP with FIN Flood

Send many FINs

Server handles or drops

46

TCP with RST Flood

Send many RSTs

Server handles or drops

47

TCP with Checksum Error

Corrupt checksum

Packet discarded

48

TCP with Invalid Flags

Send invalid flags

Packet dropped or ignored

49

TCP with Connection Migration

Change IP mid-session

Connection fails (TCP not mobile)

50

TCP with Application Layer

Use HTTP over TCP

Data transmitted reliably

Reliable Data Transfer - Testcases

S.No

Test Case

Description

Expected Result

1

Data Delivery Confirmation

Send data from client to server

ACK received for each segment

2

Ordered Data Delivery

Send multiple segments

Data received in correct order

3

Duplicate Data Handling

Send duplicate segments

Duplicates discarded

4

Lost Segment Recovery

Drop a segment

Retransmission occurs

5

Timeout-Based Retransmission

Simulate timeout

Segment retransmitted

6

Sequence Number Tracking

Monitor sequence numbers

Numbers increment correctly

7

Acknowledgment Number Tracking

Monitor ACK numbers

ACKs match received data

8

Flow Control with Window Size

Adjust window size

Sender adapts transmission rate

9

Congestion Control Activation

Simulate congestion

TCP reduces transmission rate

10

Fast Retransmit

Trigger triple duplicate ACKs

Lost segment retransmitted quickly

11

Fast Recovery

Recover after fast retransmit

Congestion window adjusted

12

Selective Acknowledgment (SACK)

Enable SACK

Only missing segments retransmitted

13

Data Integrity Check

Corrupt data

TCP checksum detects error

14

Data Reassembly

Send fragmented data

Receiver reassembles correctly

15

Zero Window Advertisement

Set window to 0

Sender pauses transmission

16

Window Update

Increase window size

Sender resumes transmission

17

Retransmission Limit

Exceed retry limit

Connection reset

18

Keepalive Mechanism

Enable keepalive

Idle connection maintained

19

Connection Termination

Send FIN

All data acknowledged before close

20

Connection Reset

Send RST

Connection terminated immediately

21

Data Transfer Over VPN

Use VPN tunnel

Data reliably transferred

22

Data Transfer Over NAT

Use NAT device

Data reliably transferred

23

Data Transfer Over IPv6

Use IPv6

Data reliably transferred

24

Data Transfer Over IPv4

Use IPv4

Data reliably transferred

25

Data Transfer with Packet Loss

Drop packets

TCP recovers lost data

26

Data Transfer with Packet Reordering

Reorder packets

TCP reorders correctly

27

Data Transfer with Duplicate Packets

Send duplicates

TCP discards duplicates

28

Data Transfer with Fragmentation

Fragment large data

TCP reassembles correctly

29

Data Transfer with High Latency

Simulate delay

TCP maintains reliability

30

Data Transfer with High Bandwidth

Send large data

TCP handles throughput

31

Data Transfer with TLS

Use TLS over TCP

Secure and reliable transfer

32

Data Transfer with Application Layer

Use HTTP over TCP

Data reliably delivered

33

Data Transfer with Nagles Algorithm

Enable Nagle

Small packets coalesced

34

Data Transfer with Push Flag

Set PSH flag

Data pushed immediately

35

Data Transfer with Urgent Pointer

Use URG flag

Urgent data processed

36

Data Transfer with Checksum Error

Corrupt checksum

Segment discarded

37

Data Transfer with Invalid Flags

Send invalid flags

Segment dropped

38

Data Transfer with SYN Flood

Simulate attack

Server handles or drops

39

Data Transfer with FIN Flood

Simulate attack

Server handles or drops

40

Data Transfer with RST Flood

Simulate attack

Server handles or drops

41

Data Transfer with Multiple Streams

Open multiple TCP connections

All streams reliable

42

Data Transfer with Connection Migration

Change IP mid-session

TCP connection fails

43

Data Transfer with Firewall

Block port

Connection fails

44

Data Transfer with Proxy

Use proxy server

Data reliably transferred

45

Data Transfer with Load Balancer

Use load balancer

Data reliably routed

46

Data Transfer with Monitoring Tool

Monitor TCP stream

Data integrity confirmed

47

Data Transfer with Automation Script

Automate TCP test

Reliable delivery verified

48

Data Transfer with Compliance Audit

Run audit

TCP meets reliability standards

49

Data Transfer with Simulation

Simulate TCP session

Data reliably transferred

50

Data Transfer with Packet Sniffer

Capture TCP packets

All segments verified

Error Detection - Testcases

S.No

Test Case

Description

Expected Result

1

Null Input

Submit empty input

Error detected: Null input

2

Invalid Format

Input in wrong format

Error detected: Format mismatch

3

Missing Field

Required field missing

Error detected: Missing field

4

Extra Field

Extra field in input

Error detected: Unexpected field

5

SQL Injection

Input contains SQL code

Error detected: Malicious input

6

XSS Attack

Input contains script tags

Error detected: XSS attempt

7

Special Characters

Input has unsupported characters

Error detected: Invalid characters

8

Long Input

Input exceeds max length

Error detected: Input too long

9

Short Input

Input below min length

Error detected: Input too short

10

Incorrect Data Type

Input type mismatch

Error detected: Type error

11

Duplicate Entry

Input already exists

Error detected: Duplicate data

12

Invalid Email

Email format incorrect

Error detected: Invalid email

13

Invalid Date

Date format incorrect

Error detected: Invalid date

14

Future Date

Date is in the future

Error detected: Invalid date

15

Past Date

Date is too old

Error detected: Invalid date

16

Invalid Number

Non-numeric input in numeric field

Error detected: Not a number

17

Negative Number

Negative value in positive-only field

Error detected: Invalid value

18

Zero Value

Zero in non-zero field

Error detected: Invalid value

19

Boundary Value

Value on boundary

Error detected: Boundary violation

20

HTML Tags

Input contains HTML

Error detected: HTML not allowed

21

JSON Malformed

JSON structure broken

Error detected: Invalid JSON

22

XML Malformed

XML structure broken

Error detected: Invalid XML

23

File Upload Error

Corrupt file uploaded

Error detected: File error

24

Unsupported File Type

Wrong file format

Error detected: Unsupported type

25

Large File Upload

File exceeds size limit

Error detected: File too large

26

Timeout Error

Input processing timeout

Error detected: Timeout

27

Network Error

Network failure during input

Error detected: Network issue

28

Unauthorized Access

Input from unauthorized user

Error detected: Access denied

29

Session Expired

Input after session timeout

Error detected: Session expired

30

Invalid Token

Wrong auth token

Error detected: Invalid token

31

Incorrect Password

Wrong password entered

Error detected: Authentication failed

32

Invalid Username

Username not found

Error detected: User not found

33

Blank Spaces

Input only has spaces

Error detected: Empty input

34

Tab Characters

Input contains tabs

Error detected: Invalid characters

35

Unicode Characters

Input has unsupported Unicode

Error detected: Encoding error

36

Case Sensitivity

Wrong case used

Error detected: Case mismatch

37

Logic Error

Input causes logic failure

Error detected: Logic error

38

API Error

API returns error

Error detected: API failure

39

Dependency Failure

External service fails

Error detected: Dependency error

40

Retry Limit Exceeded

Too many retries

Error detected: Retry limit reached

41

Invalid URL

URL format incorrect

Error detected: Invalid URL

42

Broken Link

URL leads to 404

Error detected: Link not found

43

Invalid Phone Number

Wrong phone format

Error detected: Invalid phone

44

Invalid Country Code

Wrong country code

Error detected: Invalid code

45

Invalid Currency

Unsupported currency format

Error detected: Invalid currency

46

Invalid Timezone

Timezone not recognized

Error detected: Invalid timezone

47

Invalid Language Code

Language code not supported

Error detected: Invalid language

48

Invalid Boolean

Non-boolean in boolean field

Error detected: Type mismatch

49

Invalid Enum

Value not in enum list

Error detected: Enum mismatch

50

System Crash

Input causes crash

Error detected: System failure

Flow Control - Testcases

S.No

Test Case

Description

Expected Result

1

Valid Flow

Execute normal flow

Flow proceeds without error

2

Invalid Transition

Trigger invalid state change

Error detected: Invalid transition

3

Missing Step

Skip a required step

Error detected: Step missing

4

Repeated Step

Repeat a step unnecessarily

Error detected: Redundant step

5

Unauthorized Flow

User tries restricted flow

Error detected: Access denied

6

Timeout in Flow

Step takes too long

Error detected: Timeout

7

Interrupted Flow

Flow interrupted mid-process

Error detected: Flow interrupted

8

Loop Detection

Detect infinite loop

Error detected: Loop detected

9

Conditional Branching

Test conditional path

Correct branch executed

10

Parallel Execution

Run steps in parallel

All steps complete successfully

11

Sequential Execution

Run steps in sequence

Steps execute in correct order

12

Retry Mechanism

Retry failed step

Step retried successfully

13

Max Retry Exceeded

Retry limit reached

Error detected: Retry limit exceeded

14

Rollback on Failure

Rollback after error

Previous state restored

15

Resume Flow

Resume after pause

Flow resumes correctly

16

Pause Flow

Pause mid-process

Flow paused successfully

17

Cancel Flow

Cancel ongoing flow

Flow cancelled successfully

18

Invalid Input in Flow

Provide wrong input

Error detected: Invalid input

19

Flow Completion

Complete all steps

Flow ends successfully

20

Flow Abandonment

User abandons flow

Flow marked as incomplete

21

Flow Restart

Restart from beginning

Flow restarted successfully

22

Flow Status Check

Query flow status

Correct status returned

23

Flow History

View past flow steps

History displayed correctly

24

Flow Logging

Log each step

Logs created successfully

25

Flow Metrics

Measure flow performance

Metrics recorded accurately

26

Flow Validation

Validate flow configuration

Flow validated successfully

27

Flow Configuration Error

Misconfigured flow

Error detected: Configuration issue

28

Flow Trigger

Trigger flow manually

Flow triggered successfully

29

Auto Trigger Flow

Trigger flow automatically

Flow starts as expected

30

Flow Dependency Check

Check dependent flows

Dependencies validated

31

Flow Versioning

Use different flow versions

Correct version executed

32

Flow Upgrade

Upgrade flow version

Upgrade successful

33

Flow Downgrade

Downgrade flow version

Downgrade successful

34

Flow Access Control

Restrict flow access

Access control enforced

35

Flow Audit Trail

Track flow actions

Audit trail generated

36

Flow Notification

Notify on flow events

Notifications sent

37

Flow Alert

Alert on flow failure

Alert triggered

38

Flow Escalation

Escalate on error

Escalation initiated

39

Flow Integration

Integrate with external system

Integration successful

40

Flow API Call

Trigger flow via API

API call successful

41

Flow UI Trigger

Trigger flow via UI

UI trigger successful

42

Flow Data Mapping

Map data between steps

Data mapped correctly

43

Flow Data Loss

Data lost during flow

Error detected: Data loss

44

Flow Data Validation

Validate data at each step

Data validated successfully

45

Flow Step Skipping

Skip optional step

Flow continues correctly

46

Flow Step Reordering

Change step order

Flow adapts correctly

47

Flow Step Failure

Step fails unexpectedly

Error handled gracefully

48

Flow Step Success

Step completes successfully

Step marked as complete

49

Flow Step Timeout

Step exceeds time limit

Error detected: Step timeout

50

Flow End Condition

Check end condition

Flow ends as expected

Port Addressing - Testcases

S.No

Test Case

Description

Expected Result

1

Valid Port Number

Use a valid port (e.g., 80)

Port accepted

2

Port Number Too Low

Use port number < 0

Error: Invalid port

3

Port Number Too High

Use port number > 65535

Error: Invalid port

4

Reserved Port

Use reserved port (e.g., 0)

Error: Reserved port

5

Duplicate Port

Assign same port to two services

Error: Port conflict

6

Port Already in Use

Try to bind to used port

Error: Port in use

7

Dynamic Port Range

Use port in dynamic range (4915265535)

Port accepted

8

Well-Known Port

Use port in 01023 range

Port accepted with admin rights

9

Registered Port

Use port in 102449151 range

Port accepted

10

Port Format String

Input port as string

Error: Invalid format

11

Port Format Float

Input port as float

Error: Invalid format

12

Port Format Hex

Input port in hex

Error or convert to decimal

13

Port Format Negative

Input negative port

Error: Invalid port

14

Port Format Null

No port provided

Error: Port required

15

Port Format Empty

Empty string as port

Error: Port required

16

Port Format Special Char

Port with special characters

Error: Invalid characters

17

Port Format Alphanumeric

Port with letters

Error: Invalid format

18

Port Binding Success

Bind to available port

Binding successful

19

Port Binding Failure

Bind to unavailable port

Error: Binding failed

20

Port Release

Release port after use

Port released successfully

21

Port Reuse

Reuse released port

Port reused successfully

22

Port Timeout

Port not released after timeout

Error: Port still in use

23

Port Scan Detection

Detect port scan attempt

Alert triggered

24

Port Forwarding

Forward traffic to another port

Forwarding successful

25

Port Blocking

Block specific port

Port blocked successfully

26

Port Filtering

Allow only specific ports

Filtering applied

27

Port Range Validation

Validate port range input

Range accepted or rejected

28

Port Conflict Resolution

Resolve port conflict

Conflict resolved

29

Port Logging

Log port usage

Logs generated

30

Port Monitoring

Monitor port activity

Activity tracked

31

Port Access Control

Restrict access to port

Access denied

32

Port Encryption

Secure port communication

Encryption enabled

33

Port Authentication

Require auth for port access

Auth enforced

34

Port Mapping

Map external to internal port

Mapping successful

35

Port NAT Traversal

Handle NAT traversal

Traversal successful

36

Port Load Balancing

Distribute traffic across ports

Load balanced

37

Port Failover

Switch to backup port

Failover successful

38

Port Health Check

Check port availability

Port is healthy

39

Port Throttling

Limit traffic on port

Throttling applied

40

Port Flood Protection

Detect port flooding

Protection triggered

41

Port Isolation

Isolate port from network

Isolation successful

42

Port Redirection

Redirect traffic to another port

Redirection successful

43

Port Protocol Match

Match port with correct protocol

Protocol matched

44

Port Protocol Mismatch

Use wrong protocol on port

Error: Protocol mismatch

45

Port Logging Disabled

Disable port logging

Logging disabled

46

Port Access Log

View access logs

Logs displayed

47

Port Usage Stats

View usage statistics

Stats generated

48

Port Configuration Error

Misconfigured port

Error: Configuration issue

49

Port Security Breach

Simulate breach attempt

Breach detected

50

Port Recovery

Recover from port failure

Recovery successful

Graceful Connection Termination - Testcases

S.No

Test Case

Description

Expected Result

1

Normal Termination

Close connection after data transfer

Connection closed gracefully

2

FIN-ACK Exchange

Verify proper FIN and ACK sequence

FIN and ACK exchanged correctly

3

Double FIN

Both sides send FIN

Connection terminated properly

4

FIN Timeout

FIN not acknowledged

Error: Timeout detected

5

Abrupt Close

Close without FIN

Error: Unclean termination

6

Half-Closed Connection

One side sends FIN, other continues

Half-close handled correctly

7

Data After FIN

Send data after FIN

Error: Data rejected

8

FIN Retransmission

FIN packet lost

FIN retransmitted successfully

9

ACK Lost

ACK for FIN lost

ACK retransmitted

10

FIN Delay

Delay in sending FIN

Connection remains open until FIN

11

FIN Collision

Both sides send FIN simultaneously

Connection closed correctly

12

FIN with Data

FIN sent with last data packet

Data and FIN processed correctly

13

FIN on Idle Connection

FIN sent after idle period

Connection closed gracefully

14

FIN on Busy Connection

FIN sent during data transfer

Data completed before termination

15

FIN with RST

FIN followed by RST

Error: Abrupt termination

16

FIN Retry Limit

Retry FIN multiple times

FIN sent within retry limit

17

FIN on Secure Channel

FIN sent over TLS

Secure termination completed

18

FIN on Encrypted Data

FIN after encrypted data

Data decrypted and connection closed

19

FIN on Compressed Data

FIN after compressed data

Data decompressed and closed

20

FIN on Fragmented Data

FIN after fragmented packets

All fragments reassembled before close

21

FIN with Keep-Alive

FIN sent after keep-alive

Keep-alive stopped, connection closed

22

FIN with Heartbeat

FIN sent after heartbeat

Heartbeat stopped, connection closed

23

FIN with Authentication

FIN after auth session

Session terminated securely

24

FIN with Authorization

FIN after access control

Access revoked, connection closed

25

FIN with Logging

Log termination event

Termination logged

26

FIN with Audit Trail

Record termination in audit

Audit trail updated

27

FIN with Notification

Notify user on termination

Notification sent

28

FIN with Alert

Alert on termination

Alert triggered

29

FIN with Metrics

Record termination metrics

Metrics updated

30

FIN with Monitoring

Monitor termination process

Monitoring confirms graceful close

31

FIN with Load Balancer

Terminate connection via LB

LB handles termination correctly

32

FIN with Proxy

Terminate via proxy server

Proxy forwards FIN correctly

33

FIN with NAT

Terminate behind NAT

NAT handles FIN correctly

34

FIN with Firewall

Terminate through firewall

Firewall allows FIN packets

35

FIN with VPN

Terminate over VPN

VPN tunnel closed gracefully

36

FIN with QoS

Terminate with QoS rules

QoS maintained during termination

37

FIN with Throttling

Throttled connection termination

Throttling ends, connection closed

38

FIN with Rate Limiting

Rate-limited termination

Rate limits respected

39

FIN with Session Expiry

Terminate on session expiry

Session expired, connection closed

40

FIN with Idle Timeout

Terminate after idle timeout

Timeout triggered, connection closed

41

FIN with Manual Trigger

User manually closes connection

Manual termination successful

42

FIN with Scripted Trigger

Script initiates termination

Script executes FIN correctly

43

FIN with API Call

API triggers termination

API call closes connection

44

FIN with UI Action

UI button closes connection

UI action triggers graceful close

45

FIN with Error Recovery

Recover from error before FIN

Recovery successful, then FIN

46

FIN with Retry Logic

Retry failed termination

Retry successful

47

FIN with Backup Channel

Switch to backup before FIN

Backup used, then closed

48

FIN with Redundancy

Redundant connection closed

Redundancy handled correctly

49

FIN with Failover

Failover before termination

Failover complete, then FIN

50

FIN with Resource Cleanup

Cleanup resources on termination

Resources released successfully

Congestion Control - Testcases

S.No

Test Case

Description

Expected Result

1

Normal Traffic

Send data under normal conditions

No congestion detected

2

High Traffic Volume

Send large data quickly

Congestion detected

3

Sudden Traffic Spike

Abrupt increase in traffic

Congestion control triggered

4

Packet Loss Detection

Simulate packet loss

Congestion control activated

5

Timeout Detection

Simulate timeout

Congestion window reduced

6

Duplicate ACKs

Send duplicate ACKs

Fast retransmit triggered

7

Triple Duplicate ACKs

Send 3 duplicate ACKs

Fast recovery initiated

8

Slow Start Phase

Begin connection

Exponential window growth

9

Congestion Avoidance Phase

After slow start

Linear window growth

10

Fast Retransmit

Detect loss via ACKs

Retransmit lost packet

11

Fast Recovery

Recover from loss

Resume transmission smoothly

12

Window Size Increase

Increase congestion window

Throughput improves

13

Window Size Decrease

Decrease window after loss

Throughput reduced

14

Zero Window Size

Receiver sets window to 0

Sender pauses transmission

15

Window Probe

Probe zero window

Receiver responds with window update

16

RTT Measurement

Measure round-trip time

RTT calculated accurately

17

Bandwidth Estimation

Estimate available bandwidth

Bandwidth used efficiently

18

Buffer Overflow

Simulate buffer overflow

Congestion detected

19

Queue Delay

Increase queue delay

Congestion signaled

20

ECN Enabled

Use Explicit Congestion Notification

ECN flags processed

21

ECN Disabled

Disable ECN

No ECN flags used

22

Random Early Detection

Drop packets early

Congestion avoided

23

TCP Reno Behavior

Use Reno algorithm

Reno phases executed correctly

24

TCP Tahoe Behavior

Use Tahoe algorithm

Tahoe phases executed correctly

25

TCP Cubic Behavior

Use Cubic algorithm

Cubic growth observed

26

TCP BBR Behavior

Use BBR algorithm

Bandwidth-based control applied

27

Congestion Window Saturation

Max out window size

No further growth allowed

28

Congestion Window Reset

Reset window after loss

Window set to initial value

29

ACK Delay

Delay ACKs artificially

RTT increases, window adjusted

30

ACK Burst

Send burst of ACKs

Window grows rapidly

31

Retransmission Timeout

Trigger RTO

Congestion window reset

32

Multiple Loss Events

Simulate multiple losses

Multiple recovery phases triggered

33

No Congestion

Send small data

No control triggered

34

Idle Connection

No data sent for long

Window remains unchanged

35

Reconnection

Reconnect after termination

Slow start initiated again

36

Flow Control Interaction

Combine with flow control

Both mechanisms work together

37

Application-Limited Flow

App sends limited data

Window growth paused

38

Delayed ACKs

ACKs delayed intentionally

RTT increases, window adjusted

39

ACK Loss

ACKs lost in transit

Retransmission triggered

40

Mixed Traffic

Mix of large and small packets

Adaptive control applied

41

Network Jitter

Simulate jitter

RTT variation handled

42

Network Congestion

Simulate real congestion

Control mechanisms activated

43

Link Failure

Simulate link drop

Connection reset

44

Recovery After Congestion

Resume normal traffic

Window grows again

45

Congestion in VPN

Simulate congestion over VPN

VPN handles congestion

46

Congestion in NAT

Simulate congestion behind NAT

NAT handles control correctly

47

Congestion in Proxy

Simulate congestion via proxy

Proxy forwards control signals

48

Congestion in Cloud

Simulate cloud network congestion

Cloud handles control efficiently

49

Congestion in Mobile Network

Simulate mobile congestion

Mobile TCP adapts

50

Congestion Logging

Log congestion events

Logs generated successfully

Stream Oriented - Testcases

S.No

Test Case

Description

Expected Result

1

Continuous Data Flow

Send continuous stream

Data received in order

2

Byte Order Preservation

Check byte sequence

Bytes received in same order

3

No Message Boundaries

Send multiple messages

Messages merged as stream

4

Fragmented Transmission

Send data in fragments

Reassembled correctly

5

Large Stream

Send large data stream

Entire stream received

6

Small Stream

Send small data stream

Data received correctly

7

Stream with Delay

Introduce delay in stream

Stream resumes correctly

8

Stream with Pause

Pause and resume stream

No data loss

9

Stream with Retransmission

Simulate packet loss

Retransmission successful

10

Stream with ACK

Verify ACK for stream

ACKs received correctly

11

Stream with Windowing

Use TCP windowing

Flow controlled properly

12

Stream with Congestion

Simulate congestion

Stream slows down

13

Stream with Flow Control

Receiver limits flow

Sender adjusts rate

14

Stream with Buffer Overflow

Overflow receiver buffer

Error or flow control triggered

15

Stream with Buffer Underflow

Receiver reads too fast

Stream handled correctly

16

Stream with Timeout

Timeout during stream

Retransmission triggered

17

Stream with Keep-Alive

Maintain idle stream

Connection stays alive

18

Stream with Termination

Gracefully close stream

FIN-ACK exchanged

19

Stream with RST

Abruptly close stream

Connection reset

20

Stream with Multiple Clients

Multiple streams to server

Streams handled independently

21

Stream with Multiple Servers

Client connects to multiple servers

Streams managed separately

22

Stream with Encryption

Encrypt stream data

Data decrypted correctly

23

Stream with Compression

Compress stream data

Data decompressed correctly

24

Stream with Authentication

Authenticate before stream

Stream starts after auth

25

Stream with NAT

Stream through NAT

Data flows correctly

26

Stream with Proxy

Stream through proxy

Proxy forwards stream correctly

27

Stream with VPN

Stream over VPN

Secure stream maintained

28

Stream with Packet Reordering

Reorder packets

Stream reassembled correctly

29

Stream with Packet Duplication

Duplicate packets

Duplicates discarded

30

Stream with Packet Loss

Drop packets

Retransmission successful

31

Stream with Jitter

Vary packet timing

Stream integrity maintained

32

Stream with Latency

High latency environment

Stream adapts

33

Stream with Bandwidth Limit

Limit bandwidth

Stream throttled

34

Stream with QoS

Apply QoS rules

Priority maintained

35

Stream with Logging

Log stream activity

Logs generated

36

Stream with Monitoring

Monitor stream health

Metrics collected

37

Stream with Metrics

Measure throughput

Metrics accurate

38

Stream with Alerts

Alert on stream failure

Alert triggered

39

Stream with Recovery

Recover from failure

Stream resumes

40

Stream with Retry

Retry failed stream

Retry successful

41

Stream with Redundancy

Use redundant paths

Stream continues on backup

42

Stream with Load Balancer

Distribute stream load

Balanced correctly

43

Stream with Firewall

Stream through firewall

Allowed or blocked as configured

44

Stream with IDS

Intrusion detection on stream

Malicious stream flagged

45

Stream with Logging Disabled

Disable logging

No logs generated

46

Stream with Session Expiry

Expire session mid-stream

Stream terminated

47

Stream with Reconnection

Reconnect after drop

Stream resumes

48

Stream with Data Integrity

Verify checksum

Data integrity confirmed

49

Stream with Header Manipulation

Modify TCP headers

Stream still valid

50

Stream with Port Change

Change port mid-stream

Connection reset or re-established

Ordered Delivery - Testcases

S.No

Test Case

Description

Expected Result

1

Sequential Packets

Send packets in order

Packets received in same order

2

Out-of-Order Packets

Simulate out-of-order delivery

TCP reorders packets correctly

3

Packet Loss

Drop a packet mid-sequence

Retransmission ensures correct order

4

Duplicate Packets

Send duplicate packets

Duplicates discarded, order preserved

5

Delayed Packet

Delay one packet

TCP waits and reorders correctly

6

Fragmented Data

Send fragmented data

Reassembled in correct order

7

Large Data Stream

Send large stream

Data received in correct sequence

8

Small Data Stream

Send small stream

Order maintained

9

Mixed Packet Sizes

Vary packet sizes

Order preserved

10

ACK Delay

Delay acknowledgments

TCP maintains order

11

ACK Loss

Lose ACKs

Retransmission preserves order

12

Retransmission

Retransmit lost packet

Order restored

13

Congestion

Simulate congestion

TCP slows down but maintains order

14

Flow Control

Receiver limits flow

Order maintained

15

Window Size Change

Change TCP window size

Order unaffected

16

Connection Pause

Pause and resume connection

Order preserved

17

Connection Reset

Reset mid-transfer

Order lost, connection restarted

18

Multiple Connections

Multiple TCP streams

Each stream maintains its own order

19

NAT Traversal

Send through NAT

Order preserved

20

Proxy Routing

Send through proxy

Order preserved

21

VPN Routing

Send through VPN

Order preserved

22

Packet Reordering by Network

Network reorders packets

TCP reorders them correctly

23

Packet Duplication by Network

Network duplicates packets

TCP discards duplicates

24

Packet Delay by Network

Network delays packets

TCP buffers and reorders

25

Packet Loss by Network

Network drops packets

TCP retransmits and reorders

26

TCP Reno Behavior

Use Reno algorithm

Order maintained

27

TCP Cubic Behavior

Use Cubic algorithm

Order maintained

28

TCP BBR Behavior

Use BBR algorithm

Order maintained

29

TCP Tahoe Behavior

Use Tahoe algorithm

Order maintained

30

Application-Level Check

App verifies order

Data matches expected sequence

31

Header Manipulation

Modify TCP headers

Order preserved if headers valid

32

Sequence Number Wraparound

Sequence number exceeds max

TCP handles wraparound correctly

33

Sequence Number Collision

Simulate collision

TCP resolves and maintains order

34

Stream Termination

End stream with FIN

Final data received in order

35

Stream Restart

Restart stream

New stream starts fresh order

36

Keep-Alive Enabled

Maintain idle connection

Order preserved after idle

37

Keep-Alive Disabled

Idle connection closed

Order lost after termination

38

TLS Encrypted Stream

Encrypt data

Decrypted data in correct order

39

Compressed Stream

Compress data

Decompressed in correct order

40

Multithreaded Sender

Send from multiple threads

TCP serializes and preserves order

41

Multithreaded Receiver

Receive on multiple threads

TCP delivers in order

42

Mobile Network

Send over mobile data

Order preserved despite jitter

43

Satellite Network

Send over satellite

Order preserved despite latency

44

IoT Device Communication

Send from IoT device

Order maintained

45

Cloud Server Communication

Send to cloud server

Order preserved

46

Firewall Traversal

Send through firewall

Order preserved

47

IDS Monitoring

Monitor for order violations

No violations detected

48

Logging Enabled

Log packet sequence

Logs show correct order

49

Logging Disabled

No logs

Order verified manually

50

Stress Test

High volume, high speed

TCP maintains ordered delivery

S.No

TCP Header

Description

Size(Bytes)

Position(bits)

1

Source Port

Port number of a sending application

2

0-15 (16- bit)

2

Destination Port

Port number of a receiving application

2

16-31 (16-bit)

3

Sequence Number

Byte offset of the first data byte in this segment

4

32-63 (32-bit)

4

Acknowledgment Number

If ACK flag is set,this is the next expected byte.

4

64-95 (32-bit)

5

Data Offset(Header length)

Number of 32-bit words in TCP header,used to find start of playload.

4 bit

96-99 (4-bit)

6

Reserved

Reserved for future use, must be zero.

3 bit

100-102 (3 bits)

7

Flags(Control bits)

Includes flags: NS, CWR, ECE, URG, ACK, PSH, RST, SYN, FIN

9 bit

103-111 (9 bits)

8

Window Size

Number of bytes the sender is willing to receive.

2

112-127 (16-bit)

9

Check Sum

Error-checking value for header + data (with pseudo-header)

2

128-143 (16-bit)

10

Urgent Pointer

If URG flag is set,points to the last urgent byte in the segment.

2

144-159 (16-bit)

11

Data (Pay Load)

Actual appilication data being transported

variable

after header

12

Padding

Extra bits to ensure header ends on a 32-bit boundary

variable

End of options

13

Options(optional)

optional settings like MSS, Window Scaling, Timestamps,SACK, etc.

variable

160+

Kind

Option

Description

Length(Bytes)

0

End Of Option List(EOL)

Marks the end of option list and also useful for padding

1

1

No-Operation(NOP)

Used for padding between options for alignment

1

2

Maximum Segment Size(MSS)

specifies maximum amount of data a host is willing to receive in one

4

segment.

3

Window Scale

Used to increase the TCP receive window size beyond 65,535 bytes

3

4

SACK Permitted

Indicates the sender supports SACK

2

5

SACK

One or more blocks of received data to improve retransmission.

Variable(10-34)

6

Echo(obsolote)

Used for RTT measurment ( now replaced by timestamps).

6

7

Echo Reply(obsolote)

pairedd with Echo option;also obsolote

6

8

TimeStamps(Tsopt)

Used for RTT measurment and PAWS

10

9

Partial Order Connection Permitted

Rarely used;experimental;allows out-of-order-delivery

2

10

partial Order Service Profile

Works with POCP to describe how partial order is handled.

3

11

CC(Connection Count)

Part of TCP Control Block Interdependence( for TCP research)

6

12

CC.NEW

Used in expermental congestion control

6

13

CC.ECHO

Echoes CC.NEW value

6

14

TCP Alternate checksum Request

Request alternate checksum algorithm

3

15

TCP Alternate checksum Data

Contains alternate checksum data

Variable.

28

User Timeout Option(UTO)

Allows peers to advertise their connection timeout values

4

29

TCP Authentication Option

Provides authentication for TCP segments,replacing TCP MD5 option

Variable.

253

Experimental(RFC6994)

Reserved for experimental TCP options.

Variable.

254

Experimental(RFC6994)

Reserved for experimental TCP options.

Variable.

255

Reserved

Reserved it is not used

S.No

Protocol/Application

Descrption

1

HTTP/HTTPS

  1. Used for web browsing and secure web communications. Relies on TCP for reliable delivery

2

FTP(File Transfer Protocol)

Transfers files between client and server with control and data channels over TCP

3

SMTP

Sends emails between mail servers using reliable TCP connections.

4

DNS(Zone Transfer Only)

Zone Transfers(AXFR) use TCP for reliability.

5

IMAP/POP3

Retrieves emails from mail servers while ensuring message integrity.

6

SSH

Provides secure remote login and command execution over encrypted TCP sessions.

7

Telnet

Allows remote terminal access over TCP.

8

SFTP(SSH File Transfer Protocol)

Securly transfer files over SSH,which runs over TCP.

9

LDAP over TCP

Directory services often run over TCP to support reliable data access.

S.No

Common TCP-Level Options

Explanation

Details

1

TCP_NODELAY

It is a socket option used with TCP sockets to

Header file : <netinet/tcp.h>

the use of Nagle’s algorithm.

Level : IPPROTO_TCP.

Purpose

To reduce the no of small TCP packets sent over

Disable’s Nagle’s algorithm to reduce latency for

the network,to improve bandwidth effeciency.

small packets.

Type

It is type of boolean ,because it rooted in how socket implemented in the os stack.

Boolean(int).

Values

These are commonly called as flags where the value 1 disables algorthim to send as small packets fastly,reduce latency; 0 enables it to combine packets, reducing network traffic.

0 - Diaabled.

1 - Enabled.

Default value

Default it is enabled to combine packets and

Disabled.(TCP_NODELAY=0) but Nagle’s algo is

reduce netwrok traffic at any cost of latency.

enabled may delay small packets.

Behavior

  1. When Nagle’s algorithm is enabled

In wireshark - fewer packets, delayed in PSH flag,

combined send() data.

  1. When Nagle’s algorithm is disabled(sent immediate)

In wireshark- more packets,small TCP segments,

PSH flag set,each send() data is one packet.

2

TCP_CORK

It is a Linux-specific TCP socket option that controls how data is sent over a TCP connection

Header file : <netinet/tcp.h>(linux only)

Level : IPPROTO_TCP.

Purpose

A linux-specific TCP option to control how TCP

Mainly used to optimize packetization.

segments are transmitted- mainly used to delay

& group small writes into full TCP segments.

Type

It is a boolean inerperted as int

Boolean(int).

Values

1- it hold back data until uncorked and

1-Enable

0-it sends out any buffer data.

0-Disable

Default value

it is off,TCP sends data normaly

0-Diasable(TCP_CORK=0)

Behavior

  1. when the TCP_cork is disable ,where the data sent as

In wireshark- small packets and each write()

written,may result in small packets

may be seperate TCP segment.

2.when the tcp_cork is enable ,buffers all the data until

In wireshark -fewer packets and each write()

explicitly flushed or full.

on large TCP segment.

3

TCP_KEEPALIVE

To check if an idle connection is still valid and to

Header file: <netinet/tcp.h>

prevent it from being drooped due to inactivity.

Level : SOL_SOCKET

Purpose

A mechanism that sends keepalive probed to check if the other end of TCP connections is still

Detect dead TCP connections.

avaliable.

Type

It is boolean interperted as int

Boolean(int).

values

For 0 the keep-alive is truned off & 1 it turned on.

0- Disable, 1-Enable.

Default value

Determines the time connection needs to remain

7200 seconds(2 hours).

idle before TCP starts sending keepalive probes

Appiles when

The connections remain active by sending keepalive

enabled on the socket.

probes after the connection is idle duration specified

by TCP_KEEPIDLE.

Behavior

Mechanism to check if peer is alive without data

In Wireshark shows small ACK packets with no payload

4

TCP_KEEPIDLE

set the idle time (in secs) before the first keep-alive

Header file: <netinet/tcp.h>

probe is sent.

Level : IPPROTO_TCP

Purpose

long-lived TCP connections where detecting dead

Detects dead Tcp connections early

peers quickly is important.

Type

It is interperted as int

socket option is at TCP level

Default value

if no data sent or recieve for 2 hrs the kernal will sent

7200 seconds (2 hours) on linux varies by OS.

the first keep alive if remote peer is still reachable

option

Requries keepalive to be enable first

SO_KEEPALIVE

Behavior

time of inactivity before first keepalive probe is sent.

Wireshark shows the time before keepalive seen in trace

5

TCP_KEEPINTVL

The time interval b/w individual keepalive probes sent

Header file : <netinet/tcp.h>

to remote peer when no ACK is received for the prevoius keep-alive probes.

Level : IPPROTO_TCP.

purpose

Control the frequency of probes when no ACK is recevied.

Detects dead Tcp connections quickly.

Type

It is interperted as int

Socket option is at TCP level.

Default value

Balance between network traffic & detection speed

75 seconds in linux.

option

Requries keepalive to be enable first

SO_KEEPALIVE

Behavior

Interval between the keepalive probes after the keepidle.

In Wireshark if no ACK is send to the respective packet

then only the we can see the intervals for given seconds.

6

TCP_KEEPCNT

Number of Keep-alive probes sent before declaring the

Header file : <netinet/tcp.h>

connection is dead.

Level : IPPROTO_TCP.

purpose

Controls the grace period for network issues

Premature connection termination.

Type

It is interperted as int

Socket option is at TCP level.

Default value

If there is no response ,the OS sends upto 9 more probes,

9 on linux,varies by OS.

each spaced by the interval defined in keep-interval.

Option

Requries keepalive to be enable first

SO_KEEPALIVE

Behavior

Number of unachknowledged probes before the

In wireshark it shows the multiple keepalives before the

connection dead.

RST/Close.

7

TCP_MAXSEG

The maximum amount of TCP payload data that host is

Header file : <netinet/tcp.h>

willing to receive in a single TCP segment.

Level : IPPROTO_TCP.

Purpose

Max no of bytes a TCP peer is accepting in single segment.

It doesnot include TCP headers.

Default value

Maximum size of TCP payload not the entire packet and

Typically 1460 bytes(when MTU =1500).

the value determined based on MTU interface.

Relation to MTU

It is directly derived from MTU,where its maximum size of

TCP_MAXSEG=MTU - IP Header - TCP Header.

TCP payload that fits within a single packet without causing fragmentation.

TCP option kind

Identifies the MSS option in TCP header

Kind=2(0x02).

Behavior

Indicates the largest TCP payload ize the sender can

In Wireshark - highlights MSS option in handshake packets

receive without fragmentation.

#NAME?

#NAME?

8

TCP_INFO

It is a linux socket structure that provides real-time

Header file : <netinet/tcp.h>

statistics about active TCP connection

Level : IPPROTO_TCP.

Purpose

Monitor TCP connection parameters such as RTT,

Can retrieve using getsockopt() system call.

retransmissions,congestion,window size and more….

Type

It is a linux specific , low overhead tool to observe and

Kernal-level TCP socket diagnostic interface.

tune TCP behavior without needing to sniff packets

Access method

It is accessed using getsockopt() syscall with IPPROTO_TCP

Via getsockopt() syscall with IPPROTO_TCP and TCP_INFO

level and TCP_INFO to retrieve real-time TCP connections.

option.

Behavior

TCP_INFO gives you kernel-level metrics: RTT, congestion

Wireshark gives you packet-level visibility: flags, options,

window, retransmissions, etc.

sequence numbers, timestamps.Together, they provide a complete picture of TCP behavior.

9

TCP_CONGESTION

It is a mechanism used in the TCP to manage how much

Header file : <netinet/tcp.h>

data is sent over a network to avoid congestion.

Level : IPPROTO_TCP.

Purpose

It is to ensure the efficient,fair, and stable data connection

Preventing congestion,maximum throughput,ensuring fairness

Working

TCP uses variable called Congestion Window (cwnd) to

Three categories: 1) Slow start 2) Congestion Avoidance

control how much data can be sent before waiting for ACK.

  1. Continously Congestion Detection Phase

Kernal Parameters

Currently active TCP congestion control it sets the default

/proc/sys/net/ipv4/tcp_congestion _control

algorithms for all new TCP connections

List and avaliable algorithms compiled on the system

/proc/sys/net/ipv4/tcp_avaliable_congestion _control

This list can be restrcted by admins to allow only specific

/proc/sys/net/ipv4/tcp_allowed_congestion _control

algorithms even if others are compiled into the kernal

Behaviour

Shows congestion behaviour over time(growth,loss,delay)

In Wireshark- time sequence graph,RTT graph,Throughput graph

Identify loss,recovery,congestion window chnages

In Wireshark we can see difference in window size, RT, Time

Congestion Algorithms

RENO

It is to detect and react to congestion via packet loss and

Version: Classic. Introduced : around 1990 by RFC 2001.

duplicate ACKs.uses for Fast Retransmission and Fast

Type : Loss Based AIMD

Recovery.simple and robust but not optimized for high

Growth pattern: Linear.

BDP networks.

Default: Before Kernal 2.6.8, default in Windows 2000/XP

Behavior

Watch for Fast Retransmits, Dup ACKs, and window size

In Wireshark on graph it shows classic sawtooth pattern of congestion window and after loss, window drops to half.

drops.

In Packets we can see frequent duplicate ACKs,retransmission.

CUBIC

It is defined as cubic function based congestion window

Version: Ongoing tuning. Introduced in : 2005 by linux devs.

growth algorithm .It modifies the way the TCP CWND grows

Type: Loss Baesd with cubic function growth.

after congestion,using a cubic function of time instead of a

Growth Pattern: Cubic(non-linear).

linear one.

Default: In modern linux.

Behavior

  1. slow start: TCP tries to send more data qucikly to test the

In Wireshark - sequence numbers increase rapidly ;

network capability,where rapid growth in CWND.

steep slope in the time sequence graph.

  1. cubic growth: uses a curve to find right speed without

In Wireshark we can see a smooth,curved upward growth in

overwhelming the network.

time sequence graph not a straight line like reno.

3)Congestion Avoidance: It avoids the repating congestion.

In Wireshark we can see flattering curve in the graph.

  1. Packet loss: It indicates that congestion occurs and TCP

In Wireshark there is a sudden drop in the curve possibly

reduces sending data.

followed by retransmissions and duplicate ACK’s.

In Packets it is similar to reno, but with smoother growth.

BBR

It is Bottleneck Bandwidth and Round trip Propagation time

Version: BBR V2(2021)

is a modern based congestion control algorithm created by

Introduced : 2016 by Google.

google. instead of relying on packet loss like reno or cubic

Type : Modern based (Bandwidth * RTT).

BBR measures the avaliable bandwidth and minimum RTT

Growth Pattern: Smooth,Steady

to send data at the most efficient rate.

Default: Requires Kernal >= 5.4.

Behavior

BBR doesnot slow down on the minor packets and there it

In Wireshark we can observe there would be less or no re-

will not trigger loss-based recovery.

transmissions and very rare duplicate ACK’s.

BBR sends data steadily without reacting to loss.

In Wireshark - we can see in time sequence graph smooth and

straight line no sawtooth pattern.

BBR keeps delay low by avoiding queue bulidup

In Wireshark we can see very flat and stable RTT line graph.

10

TCP_DEFER_ACCEPT

It is a linux-specific socket that tells the kernal to delay passing an incoming TCP connection to the appilication via accept() until data has been received on socket or timeout.

Header file : <netinet/tcp.h>

Level : IPPROTO_TCP.

purpose

TCP_DEFER_ACCEPT prevents the server from waking up

Reduce Wakeups for idle TCP connections.

for clients that connect but stay silent,saving resources.

Time out vaue

How long the Kernal waits for data before delivering the

integer(seconds).

connection to accept()

Behavior

It happens when TCP_defer_accept delays the server’s

In Wireshark we can see TCP 3-way handshake completed

accept() call until client sends data. the server remains

sucessfully,but there will be a delay before the server sends

silent until data arrives,visible pause between handshake

any response packet back.

and the server’s first reply.

11

TCP_QUICKACK

A TCP Socket option that disable’s ACKs,causing the kernal

Header file : <netinet/tcp.h>

to send ACK immediately for the next received packet.

Level : IPPROTO_TCP.

Purpose

To reduce latency by sending ACKs without delay

Useful for interactive or real-time applications.

default value

Linux normally uses delayed ACKs for reducing TCP

0 to restore default behavior.

overhead.

Type

TCP_QUICKACK uses an boolean value as input.

integer.

Value

It set to 1 makes TCP send ACK immediately for next packet,

1- Enable QUICKACK

and set 0 restores normal delayed ACK behavior.

0- Disable QUICKACK / Restore default

Behavior

The immediate ACK happens because TCP_QUICKACK

In Wireshark ACK is sent fastly after receiving data,with no delay

temporarily disables delayed ACKs,allowing you to see the

unlike delayed ACK mode where a small pause is observed

ACK packet appear right after the data.

before the ACK is transmitted.

S.no

General Socket-Level Options

Explanation

Details

1

SO_REUSEADDR

It is a Socket option used in network programming to allow a

Header file : <sys/socket.h>

socket to bind to a local address that is in TIME_WAIT state

Level : SOL_SOCKET.

or was recently used.

Purpose

It is to allow quick rebinding to a recently used port,even if it

Reuse local address for binding

hasn’t fully timed out yet

when to call

It changes how the OS ingterpets address/port avaliability at

Before Bind()

the moment you bind the socket.

Socket Types

It is used for both TCP and UDP sockets.

TCP(SOCK_STREAM), UDP(SOCK_DGRAM).

Value Type

Always use int for the value,even though the option affects

1(non-zero)- Enable SO_REUSEADDR.

low-level behavior

0- Disable SO_REUSEADDR.

Behavior

We cant see any behavior directly in wireshark but it reflect

In wireshark we can observe fast TCP port reuse and also TCP

indirectly effect of the option in captured network traffic.

port in TIME_WAIT reused.

2

SO_REUSEPORT

It is a socket option that allows multiple sockets to bind to

Header file : <sys/socket.h>

the same IP address and port comination.

Level : SOL_SOCKET.

Purpose

The core purpose of SO_REUSEPORT is to improve

Enables Load balancing, Parallel processing, Seamless server

performance and scalability.

restarts.

Value Type

Allows multiple processes or threads to bind to the same IP

Enable - 1 and 0- disable (default value).

combines with

They are used together to ensure flexibility and correctness.

Often used with so_reuseaddr.

Socket Types

It is used for both TCP and UDP sockets.

TCP(SOCK_STREAM), UDP(SOCK_DGRAM).

Behavior

Here the behavior in Wireshark is cant be seen directly

In Wireshark we can see multiple processes bound to the same

because so_reuseport is local socket option .

port and also fast reconnect after server restart.

3

SO_RCVBUF / SO_SNDBUF

These sockets are used with setsockopt() and getsockopt()

Header file : <sys/socket.h>

to control the buffer size used by the OS kernal for sending

Level : SOL_SOCKET.

and receiving data on a socket.

Purpose

Larger buffer can improve throughput, especially on

By recv() or read() and send() or write().

high-latency networks.

Value Type

Both expect an int value which the size in bytes of the buffer

Int (Buffer size in bytes).

you want the kernal to allocate.

When to call

These options should be set before calling connect or bind

Before Connect() ( client sockets )

to ensure the buffer size are properly appiled by the OS

Bind() ( server sockets).

Units

It is always taken in the form of Bytes

Common values: 65536(64 KB) , 1048576(1 MB) etc…

How to pass

It is passed via pointer type to integer

&buffersize via setsockopt() or getsockopt().

default value

While using these both the socket option it could be system

usually 128 KB to 256 KB, if needed can check with getsockopt()

dependent

it is a system dependent.

Max size Limits

Setting above max may be silently capped and limited by

we can configure through

Kernal.

/proc/sys/net/core/rmem_max.

Behavior

In SO_RCVBUF It controls size of receive queue in kernal,

In Wireshark we can see Large advertised TCP window size in

If its larger buffer size then more buffer data we get.

packets and the flow is smoother

If its smaller buffer size then it drops or flow control.

In Wireshark we get smaller TCP window size,Packet loss signals

and also some times re transmission happens.

IN SO_SNDBUF It controls the size of send queue in kernal,

In Wireshark we get Larger TCP segments and also possible

If its larger buffer size then its smoother in sending the data.

delayed sending if buffer is full.

If its smaller buffer size then the sometimes the send may

In Wireshark we get smaller TCP segments and also data flow

blocks or slows.

would be slow.

4

SO_RCVLOWAT / SO_SNDLOWAT

These both options are used in minimum no of byes that must

Header file : <sys/socket.h>

be avaliable in the sockets receive or send buffer before

Level : SOL_SOCKET.

blocking recv() call will return or send() call will proceed.

Purpose

The core purpose is to define minimum data threshold for

prevent returning partial messages and avoid sending tiny

recv() or send().

fragments.

Socket Direction

Directional Control: receive vs send.

SO_RCVLOWAT- Receive buffer , SO_SNDLOWAT - Send buffer.

Value Type

must be passed as pointer to int.

Integer.

Units

Threshold measured in bytes.

Bytes.

default value

recv() returns even if 1 byte is avaliable unless overridden

Typically 1 byte.

System Support

On Linux, SO_SNDLOWAT is ignored in most cases and

SO_RCVLOWAT- supports

also SO_RCVLOWAT is supported

SO_SNDLOWAT - sometimes ignored (Linux).

Behavior(blocking)

Only appiles to blocking sockets.

recv() or send() blocks until threshold is satisfied.

Behavior(unblocking)

Ignored in the non-blocking mode.

Both of them has no effect on 0_NONBLOCK sockets.

Behavior

Because of app waits to read/send large chunks and it waits

In Wireshark we cant see these output directly but indirectly we

for watermark thershold.

can see TCP segments carry larger payloads and delays between

request and response.

5

SO_RCVTIMEO / SO_SNDTIMEO

These sets the maximum amount of time to recv() or read() &

Header file : <sys/socket.h> / <sys/time.h>

send() or write() call on a blocking socket will wait for data or

Level : SOL_SOCKET.

will wait to send data before time out.

Purpose

Prevents call from a blocking forever if no data is received or

set a timeout duration for blocking recv() or send() calls.

send buffer is full

Value Type

It is a structure which must include both seconds & micro sec.

struct timeval

default value

There is no value until we explicitly set it

No timeout(i.e blocks indefinitely).

units

These were passed as struct timeval.

Time( seconds + microseconds).

Appiles To

Works only on blocking sockets.

SO_RCVTIMEO - recv() / read()

SO_SNDTIMEO - send() / write().

Behavior

In SO_RCVTIMEO we can see it waits for data,but timeout

In Wireshark we can observe no response or delayed from

occured and also it may ignore socket and broken message

application and also gaps in appilication data flow .

handling.

In SO_SNDTIMEO we can see send() may blocks due to full

In Wireshark we can observe like application stop sending

buffer,timed out and also send() may hit timeout if peer

suddenly and also TCP buffer full(zero Window size from peer) and also outgoing data ends with RST or FIN.

doesn’t open Window.

6

SO_LINGER

This socket option that specifies whether a close() should

Header file : <sys/socket.h>

block to allow unsent data to be transmitted before closing a

Level : SOL_SOCKET.

TCP socket

purpose

Controls whether close() waits to deliver unsent data or

Manage proper shutdowns,avoid RSTs,or force quick

aborts the connection immediately.

disconnects.

Value Type

It is a structure type which is struct linger

It contains l_onoff and l_linger fields.

Appiles To

It is mainly for TCP and no menaing for UDP sockets

TCP sockets(SOCK_STREAM).

default value

close() returns immediately,TCP tries to flush remaining data

l_onoff=0 (linger disabled).

in background.

Behavior

when l_onoff=0(disabled)

In Wireshark we can see FIN from the closing side,followed by

FIN-ACK fromm peer-graceful 4-way termination.

when l_onoff=1, where l_linger>0

In Wireshark it send pending data;if timeout hits we can see RST

instead of FIN- indicates the force close.

when l_onoff=1, where l_linger=0

In Wireshark we can see immediately TCP RST instead of a FIN-

where connection is reset ungracefully.

S.No

Parameter Name

File Path

Description

Cat Outputs

1

tcp_abort_on_overflow

/proc/sys/net/ipv4/tcp_abort_on_overflow

Drop overflowed receive buffers instead of dropping incoming

0

overflow.

segments.

2

tcp_allowed_congestion_control

/proc/sys/net/ipv4/tcp_allowed_congestion_control

Congestion control algorithm allowed by the kernal.

Reno, cubic

3

tcp_adv_win_scale

/proc/sys/net/ipv4/tcp_adv_win_scale

Controls how much of the TCP receive is advertised to the remote

1

peer.

4

tcp_app_win

/proc/sys/net/ipv4/tcp_app_win

Flow control window left to the application

31

5

tcp_autocorking

/proc/sys/net/ipv4/tcp_autocorking

Controls whether the Linux TCP stack automatically corks (delay

1

sending) small packets

6

tcp_avaliable_congestion_control

/proc/sys/net/ipv4/tcp_avaliable_congestion_control

currently avaliable congestion control modules.

reno,cubic.

7

tcp_avaliable_ulp

/proc/sys/net/ipv4/tcp_avaliable_ulp

Displays a list of all avaliable ULP modules that can be used with TCP.

tls

8

tcp_base_mss

/proc/sys/net/ipv4/tcp_base_mss

Minimum segment size used by TCP.

1024

9

tcp_challenge_ack_limit

/proc/sys/net/ipv4/tcp_challenge_ack_limit

maximum challenge Acks per second.

2147483647

10

tcp_child_ehash_entries

/proc/sys/net/ipv4/tcp_child_ehash_entries

specifies a limit on the number of child entries in the TCP established

0

hash table.

11

tcp_comp_sack_delay_ns

/proc/sys/net/ipv4/tcp_comp_sack_delay_ns

sets the delay(in nano seconds) before sending a compressed SACK.

1000000

12

tcp_congestion_control

/proc/sys/net/ipv4/tcp_congestion_control.

default congestion control algorithm

cubic

13

tcp_comp_sack_nr

/proc/sys/net/ipv4/tcp_comp_sack_nr

number of threshold for delaying SACK

44

14

tcp_dsack

/proc/sys/net/ipv4/tcp_dsack

Enable duplicate Selective SACK

1

15

tcp_early_retrans

/proc/sys/net/ipv4/tcp_early_retrans

Number of dup ACKs to trigger early retransmission.

3

16

tcp_ecn

/proc/sys/net/ipv4/tcp_ecn

Enable Explicit congestion control

2

17

tcp_ecn_fallback

/proc/sys/net/ipv4/tcp_ecn_fallback

Fallbacks to regular TCP if ECN fails

1

18

tcp_fack

/proc/sys/net/ipv4/tcp_fack

Enable forward ACK

0

19

tcp_fastopen

/proc/sys/net/ipv4/tcp_fastopen

Enable tcp fastopen.

1

20

tcp_fastopen_blackhole_timeout_sec

/proc/sys/net/ipv4/tcp_fastopen_blackhole_timeout_sec

Timeout for blackhole detection.

0

21

tcp_fastopen_key

/proc/sys/net/ipv4/tcp_fastopen_key

Session tickey key for secure fast open

22

tcp_fin_timeout

/proc/sys/net/ipv4/tcp_fin_timeout

FIN-WAIT timeout before the socket is closed.

60

23

tcp_frto

/proc/sys/net/ipv4/tcp_frto

Enable forward RTO recovery

2

24

tcp_recovery

/proc/sys/net/ipv4/tcp_recovery

Configures TCP recovery methods.

1

25

tcp_reordering

/proc/sys/net/ipv4/tcp_reordering

Max out-of-order packets before triggering fast transmit

3

26

tcp_retrans_collapse

/proc/sys/net/ipv4/tcp_retrans_collapse

collapse multiple retransmissions into one

1

27

tcp_rmem

/proc/sys/net/ipv4/tcp_rmem

receive buffer sizes[min,default,max]

4096 131072 6291456

28

tcp_rto_min_us

/proc/sys/net/ipv4/tcp_rto_min_us

Minimum retransmission timeout (micro seconds)

200000

29

tcp_sack

/proc/sys/net/ipv4/tcp_sack

Enable selective ACK

1

30

tcp_slow_start_after_idle

/proc/sys/net/ipv4/tcp_slow_start_after_idle

Slow start after idle period

1

31

tcp_stdurg

/proc/sys/net/ipv4/tcp_stdrug

Accept out-of-band urgent data

0

32

tcp_syn_retries

/proc/sys/net/ipv4/tcp_syn_retries

SYN retries for outgoing connections.

6

33

tcp_synack_retries

/proc/sys/net/ipv4/tcp_synack_retries

SYN-ACK retries for incoming connections

5

34

tcp_syncookies

/proc/sys/net/ipv4/tcp_syncookies

Enable SYN cookies

1

35

tcp_syn_linear_timeouts

/proc/sys/net/ipv4/tcp_syn_linear_timeouts

Linear timeouts for this streams

4

36

tcp_timestamps

/proc/sys/net/ipv4/tcp_timestamps

Enable TCP timestamps

1

37

tcp_thin_dupack

/proc/sys/net/ipv4/tcp_thin_dupack

Dedup ACK reduction for thin streams

0

38

tcp_thin_linear_timeouts

/proc/sys/net/ipv4/tcp_thin_linear_timeouts

Enables linear timeouts for thin streams

1

39

tcp_tw_reuse

/proc/sys/net/ipv4/tcp_tw_reuse

Allows reuse of TIME-WAIT sockets for new connections

1

40

tcp_tw_rfc1337

/proc/sys/net/ipv4/tcp_tw_rfc1337

Enables RFC1337 fix for TIME-WAIT assassination

1

41

tcp_max_syn_backlog

/proc/sys/net/ipv4/tcp_max_syn_backlog

Max number of remembered connection requests in SYN_RECV state

2048

42

tcp_low_latency

/proc/sys/net/ipv4/tcp_low_latency

Enables low latency mode (disables TCP performance features)

0

43

tcp_limit_output_bytes

/proc/sys/net/ipv4/tcp_limit_output_bytes

Limits number of bytes in output queue

131072

44

tcp_no_metrics_save

/proc/sys/net/ipv4/tcp_no_metrics_save

Prevents saving TCP metrics on socket close

0

45

tcp_mtu_probing

/proc/sys/net/ipv4/tcp_mtu_probing

Enables MTU probing to discover path MTU

1

46

tcp_keepalive_time

/proc/sys/net/ipv4/tcp_keepalive_time

Time before sending keepalive probes

7200

47

tcp_keepalive_intvl

/proc/sys/net/ipv4/tcp_keepalive_intvl

Interval between keepalive probes

75

48

tcp_keepalive_probes

/proc/sys/net/ipv4/tcp_keepalive_probes

Number of keepalive probes before dropping connection

9

49

tcp_max_orphans

/proc/sys/net/ipv4/tcp_max_orphans

Max number of TCP orphaned connections

4096

50

tcp_max_tw_buckets

/proc/sys/net/ipv4/tcp_max_tw_buckets

Max number of TIME-WAIT sockets

180000

51

tcp_wmem

/proc/sys/net/ipv4/tcp_wmem

Send buffer sizes [min, default, max]

4096 16384 4194304

52

tcp_mem

/proc/sys/net/ipv4/tcp_mem

Memory thresholds for TCP: [low, pressure, high] in pages

181419 241895 362838

53

tcp_min_tso_segs

/proc/sys/net/ipv4/tcp_min_tso_segs

Minimum number of segments for TSO (TCP Segmentation Offload)

2

54

tcp_min_rtt_wlen

/proc/sys/net/ipv4/tcp_min_rtt_wlen

Window length for minimum RTT estimation

300

55

tcp_probe_interval

/proc/sys/net/ipv4/tcp_probe_interval

Interval to probe TCP sockets for memory usage

600

56

tcp_probe_threshold

/proc/sys/net/ipv4/tcp_probe_threshold

Threshold for probing TCP sockets

8

57

tcp_tso_win_divisor

/proc/sys/net/ipv4/tcp_tso_win_divisor

Divides TCP window for TSO segment size

3

58

tcp_workaround_signed_windows

/proc/sys/net/ipv4/tcp_workaround_signed_windows

Enables workaround for broken TCP window scaling

0

59

tcp_window_scaling

/proc/sys/net/ipv4/tcp_window_scaling

Enables TCP window scaling (RFC 1323)

1

60

tcp_orphan_retries

/proc/sys/net/ipv4/tcp_orphan_retries

Max retries for orphaned sockets

0

61

tcp_base_mss

/proc/sys/net/ipv4/tcp_base_mss

Minimum TCP segment size

1024

62

tcp_cookie_size

/proc/sys/net/ipv4/tcp_cookie_size

Size of TCP cookie for TCP Cookie Transactions (TCPCT)

8

63

tcp_cts_protect

/proc/sys/net/ipv4/tcp_cts_protect

Protects against CTS (Clear to Send) attacks

0

64

tcp_dma_copybreak

/proc/sys/net/ipv4/tcp_dma_copybreak

Threshold for using DMA for TCP payload copying

4096

65

tcp_dsack_disable

/proc/sys/net/ipv4/tcp_dsack_disable

Disables DSACK (Duplicate SACK)

0

66

tcp_invalid_ratelimit

/proc/sys/net/ipv4/tcp_invalid_ratelimit

Rate limit for invalid TCP packets

500

67

tcp_l3mdev_accept

/proc/sys/net/ipv4/tcp_l3mdev_accept

Accept TCP connections on L3 master devices

0

68

tcp_moderate_rcvbuf

/proc/sys/net/ipv4/tcp_moderate_rcvbuf

Enables automatic tuning of TCP receive buffer sizes

1

69

tcp_notsent_lowat

/proc/sys/net/ipv4/tcp_notsent_lowat

Threshold for TCP_NOTSENT_LOWAT socket option

4294967295

70

tcp_reflect_tos

/proc/sys/net/ipv4/tcp_reflect_tos

Reflects Type of Service (ToS) from incoming to outgoing packets

0

71

tcp_autotune

/proc/sys/net/ipv4/tcp_autotune

Enables automatic tuning of TCP buffer sizes

1

72

tcp_challenge_ack_limit

/proc/sys/net/ipv4/tcp_challenge_ack_limit

Max challenge ACKs per second to mitigate spoofing

100

73

tcp_cookie_size

/proc/sys/net/ipv4/tcp_cookie_size

Size of TCP cookie for TCPCT (TCP Cookie Transactions)

8

74

tcp_limit_output_bytes

/proc/sys/net/ipv4/tcp_limit_output_bytes

Max bytes queued before TCP output is throttled

262144

75

tcp_min_snd_mss

/proc/sys/net/ipv4/tcp_min_snd_mss

Minimum send MSS

48

76

tcp_mtu_probing

/proc/sys/net/ipv4/tcp_mtu_probing

Enables path MTU discovery

1

77

tcp_mtu_probe_floor

/proc/sys/net/ipv4/tcp_mtu_probe_floor

Minimum floor for MTU probing

512

78

tcp_no_ssthresh_metrics_save

/proc/sys/net/ipv4/tcp_no_ssthresh_metrics_save

Avoid saving ssthresh in metrics cache

0

79

tcp_probe_interval

/proc/sys/net/ipv4/tcp_probe_interval

Interval for TCP memory pressure probing

600

80

tcp_probe_threshold

/proc/sys/net/ipv4/tcp_probe_threshold

Threshold for TCP memory pressure probing

8

  • Reference links