One place for hosting & domains

      Learning

      Open-Sourced Deep Learning With OpenVINO


      How to Join

      This Tech Talk is free and open to everyone. Register below to get a link to join the live event.

      Format Date RSVP
      Presentation and Q&A October 28, 12:00–1:00 p.m. ET

      If you can’t join us live, the video recording will be published here as soon as it’s available.

      About the Talk

      Built on top of popular open-sourced libraries such as OpenCV, OpenVINO is a truly open-sourced platform that enables deep learning deployments across platforms, including CPUs, integrated GPUs, FPGAs and AI with a write-once, deploy-anywhere simplicity. We discuss what it takes to build a sustainable, open-sourced deep learning inference platform for everyone. This talk highlights how you can get involved in the community, and educational resources you can use to learn more.

      What You’ll Learn

      • How collaboration and the open source community are shaping deep learning today
      • Ways to contribute to the future of deep learning
      • Ready-to-go tools, educational resources, and a vibrant community to help you get started quickly

      This Talk is Designed For

      • AI/DL app developers of all levels, from beginners to experts

      About the Presenters

      • Zoe Cayetano – Product Manager for OpenVINO, Intel
      • Raymond Lo – Software Evangelist for OpenVINO, Intel

      Zoe Cayetano is a Product Manager at Intel working on advancing the deployment of AI/DL technologies from edge to cloud. Prior to Intel, she founded an emotion-sensing headphones startup and was a data science researcher for a particle accelerator at Arizona State University, where she analyzed electron beam dynamics of novel x-ray lasers. She holds Bachelor’s degrees in Applied Physics and Business.

      Raymond Lo is a Software Evangelist at Intel for AI and deep learning. Prior to joining Intel, Raymond was the founder and CTO of Meta (a YCombinator-backed augmented reality company) and the Technology Evangelist for Samsung NEXT. During his PhD, Raymond worked with Prof. Steve Mann, who is widely recognized as the father of wearable computing.

      To join the live Tech Talk, register here.



      Source link

      Learning Go Functions, Loops, and Errors – A Tutorial


      Updated by Linode Contributed by Mihalis Tsoukalos

      Introduction

      Go is a modern, open source, and general-purpose programming language that began as an internal Google project and was officially announced at the end of 2009. Go was inspired by many other programming languages including C, Pascal, Alef, and Oberon. Its spiritual fathers were Robert Griesemer, Ken Thomson, and Rob Pike, who all designed Go as a language for professional programmers that want to build reliable, robust, and efficient software. Apart from its syntax and its standard functions, Go comes with a rich standard library.

      In this Guide

      This guide will cover the following topics:

      Note

      This guide was written with Go version 1.13.

      Before You Begin

      1. You will need Go installed on your computer. To get it, go to Go’s official download page and get the installer for your operating system, or you can install it from source. Follow the installation instructions for your operating system.

      2. Add /usr/local/go/bin to the PATH environment variable:

        export PATH=$PATH:/usr/local/go/bin
        

        You may need to restart your shell for this change to apply.

      The Advantages of Go

      Although Go is not perfect, it has many advantages, including the following:

      • It is a modern programming language that was made by experienced developers for developers.
      • The code is easy to read.
      • Go keeps concepts orthogonal, or simple, because a few orthogonal features work better than many overlapping ones.
      • The compiler prints practical warnings and error messages that help you solve the actual problem.
      • It has support for procedural, concurrent, and distributed programming.
      • Go supports garbage collection so you do not have to deal with memory allocation and deallocation.
      • Go can be used to build web applications and it provides a simple web server for testing purposes.
      • The standard Go library offers many packages that simplify the work of the developer.
      • It uses static linking by default, which means that the produced binary files can be easily transferred to other machines with the same OS and architecture. As a consequence, once a Go program is compiled successfully and the executable file is generated, the developer does not need to worry about dependencies and library versions.
      • The code is portable, especially among UNIX machines.
      • Go can be used for writing UNIX systems software.
      • It supports Unicode by default which means that you do not need any extra code for printing characters from multiple human languages or symbols.

      Executing Go code

      There are two kinds of Go programs: autonomous programs that are executable, and Go libraries. Go does not care about an autonomous program’s file name. What matters is that the package name is main and that there is a single main() function in it. This is because the main() function is where program execution begins. As a result, you cannot have multiple main() functions in the files of a single project.

      A Simple Go program

      This is the Go version of the Hello World program:

      ./helloworld.go
      1
      2
      3
      4
      5
      6
      7
      8
      9
      
      package main
      
      import (
          "fmt"
      )
      
      func main() {
          fmt.Println("Hello World!")
      }
      • All Go code is delivered within Go packages. For executable programs, the package name should be main. Package declarations begin with the package keyword.

      • Executable programs should have a function named main() without any function parameters. Function definitions begin with the func keyword.

      • Go packages might include import statements for importing Go packages. However, Go demands that you use some functionality from each one of the packages that you import. There is a way to bypass this rule, however, it is considered a bad practice to do this.

        The helloworld.go file above imports the fmt package and uses the fmt.Println() function from that package.

        Note

        All exported package functions begin with an uppercase letter. This follows the Go rule: if you export something outside the current package, it should begin with an uppercase letter. This rule applies even if the field of the Go structure or the global variable is included in a Go package.

      • Go statements do not need to end with a semicolon. However, you are free to use semicolons if you wish. For more information on formatting with curly braces, see the section below.

      1. Now that you better understand the helloworld.go program, execute it with the go run command:

        go run helloworld.go
        

        You will see the following output:

          
        Hello World!
        
        

        This is the simplest of two ways that you can execute Go code. The go run command compiles the code and creates a temporary executable file that is automatically executed and then it deletes that temporary executable file. This is similar to using a scripting programming language.

      2. The second method to execute Go code is to use the build command. Run the following command to use this method:

        go build helloworld.go
        

        The result of that command is a binary executable file that you have to manually execute. This method is similar to the way you execute C code on a UNIX machine. The executable file is named after the Go source filename, which means that in this case the result will be an executable file named helloworld. Go creates statically linked executable files that have no dependencies to external libraries.

      3. Execute the helloworld file:

        ./helloworld
        

        You will see the following output:

          
        Hello World!
        
        

        Note

        The go run command is usually used while experimenting and developing new Go projects. However, if you need to transfer an executable file to another system with the same architecture, you should use go build.

      Formatting Curly Braces

      The following version of the “Hello World” program will not compile:

      ./curly.go
       1
       2
       3
       4
       5
       6
       7
       8
       9
      10
      
      package main
      
      import (
          "fmt"
      )
      
      func main()
      {
          fmt.Println("Hello World!")
      }
      1. Execute the program above, and observer the error message generated by the compiler:

        go run curly.go
        
          
        # command-line-arguments
        ./curly.go:7:6: missing function body
        ./curly.go:8:1: syntax error: unexpected semicolon or newline before {
        
        
      • This error message is generated because Go requires the use of semicolons as statement terminators in many contexts and the compiler automatically inserts the required semicolons when it thinks that they are necessary. Putting the opening curly brace ({) on its own line makes the Go compiler look for a semicolon at the end of the previous line (func main()), which is the cause of the error message.

      • There is only one way to format curly braces in Go; the opening curly brace must not appear on it’s own line. Additionally, you must use curly braces even if a code block contains a single Go statement, like in the body of a for loop. You can see an example of this in the first version of the helloworld.go program or in the Loops in Go section.

      The Assignment Operator and Short Variable Declarations

      • Go supports assignment (=) operators and short variable declarations (:=).
      • With := you can declare a variable and assign a value to it at the same time. The type of the variable is inferred from the given value.
      • You can use = in two cases. First, to assign a new value to an existing variable and second, to declare a new variable, provided that you also give its type.

        For example, var aVariable int = 10, is equivalent to aVariable := 10 assuming aVariable is an int.

      • When you specifically want to control a variable’s type, it is safer to declare the variable and its type using var and then assign a value to it using =.

      Loops in Go

      The file loops.go demonstrates loops in Go:

      ./loops.go
       1
       2
       3
       4
       5
       6
       7
       8
       9
      10
      11
      12
      13
      14
      15
      16
      17
      18
      19
      20
      21
      22
      23
      24
      25
      26
      27
      28
      29
      30
      31
      32
      33
      34
      35
      36
      37
      38
      39
      40
      41
      42
      43
      44
      45
      46
      47
      48
      
      package main
      
      import (
          "fmt"
      )
      
      func main() {
          for loopIndex := 0; loopIndex < 20; loopIndex++ {
              if loopIndex%10 == 0 {
                  continue
              }
      
              if loopIndex == 19 {
                  break
              }
              fmt.Print(loopIndex, " ")
          }
          fmt.Println()
      
          // Use break to exit the for loop
          loopIndex := 10
          for {
              if loopIndex < 0 {
                  break
              }
              fmt.Print(loopIndex, " ")
              loopIndex--
          }
          fmt.Println()
      
          // This is similar to a while(true) do something loop
          loopIndex = 0
          anExpression := true
          for ok := true; ok; ok = anExpression {
              if loopIndex > 10 {
                  anExpression = false
              }
      
              fmt.Print(loopIndex, " ")
              loopIndex++
          }
          fmt.Println()
      
          anArray := [5]int{0, 1, -1, 2, -2}
          for loopIndex, value := range anArray {
              fmt.Println("index:", loopIndex, "value: ", value)
          }
      }
      • There are two types of for loops in Go. Traditional for loops that use a control variable initialization, condition, and afterthought; and those that iterate over the elements of a Go data type such as an array or a map using the range keyword.

      • Go has no direct support for while loops. If you want to use a while loop, you can emulate it with a for loop.

      • In their simplest form, for loops allow you to iterate, a predefined number of times, for as long as a condition is valid, or according to a value that is calculated at the beginning of the for loop. Such values include the size of a slice or an array, or the number of keys on a map. However, range is more often used for accessing all the elements of a slice, an array, or a map because you do not need to know the object’s cardinality in order to process its elements one by one. For simplicity, this example uses an array, and a later example will use a slice.

      • You can completely exit a for loop using the break keyword. The break keyword also allows you to create a for loop without an exit condition because the exit condition can be included in the code block of the for loop. You are also allowed to have multiple exit conditions in a for loop.

      • You can skip a single iteration of a for loop using the continue keyword.

      1. Execute the loops.go program:

        go run loops.go
        

        You will see the following output:

          
        1 2 3 4 5 6 7 8 9 11 12 13 14 15 16 17 18
        10 9 8 7 6 5 4 3 2 1 0
        0 1 2 3 4 5 6 7 8 9 10 11
        index: 0 value:  0
        index: 1 value:  1
        index: 2 value:  -1
        index: 3 value:  2
        index: 4 value:  -2
            
        

      Functions in Go

      Functions are first class citizens in Go, which means that functions can be parameters to other functions as well as returned by functions. This section will illustrate various types of functions.

      Go also supports anonymous functions. These can be defined inline without the need for a name and they are usually used for implementing operations that require a small amount of code. In Go, a function can return an anonymous function or take an anonymous function as one of its arguments. Additionally, anonymous functions can be attached to Go variables. In functional programming terminology anonymous functions are called closures. It is considered a good practice for anonymous functions to have a small implementation and a local focus.

      Regular functions

      This section will present the implementation of some traditional functions.

      ./functions.go
       1
       2
       3
       4
       5
       6
       7
       8
       9
      10
      11
      12
      13
      14
      15
      16
      17
      18
      19
      20
      21
      22
      23
      24
      25
      26
      27
      28
      29
      30
      31
      32
      33
      34
      35
      36
      37
      38
      39
      40
      41
      42
      43
      44
      45
      46
      47
      48
      49
      50
      51
      52
      53
      54
      55
      56
      57
      58
      
      package main
      
      import (
          "fmt"
      )
      
      func doubleSquare(firstNum int) (int, int) {
          return firstNum * 2, firstNum * firstNum
      }
      
      func namedMinMax(firstNum, secondNum int) (min, max int) {
          if firstNum > secondNum {
              min = secondNum
              max = firstNum
          } else {
              min = firstNum
              max = secondNum
          }
          return
      }
      
      func minMax(firstNum, secondNum int) (min, max int) {
          if firstNum > secondNum {
              min = secondNum
              max = firstNum
          } else {
              min = firstNum
              max = secondNum
          }
          return min, max
      }
      
      func main() {
          secondNum := 10
      
          square := func(numberToSquare int) int {
              return numberToSquare * numberToSquare
          }
          fmt.Println("The square of", secondNum, "is", square(secondNum))
      
          double := func(numberToDouble int) int {
              return numberToDouble + numberToDouble
          }
          fmt.Println("The double of", secondNum, "is", double(secondNum))
      
          fmt.Println(doubleSquare(secondNum))
          doubledNumber, squaredNumber := doubleSquare(secondNum)
          fmt.Println(doubledNumber, squaredNumber)
      
          value1 := -10
          value2 := -1
          fmt.Println(minMax(value1, value2))
          min, max := minMax(value1, value2)
          fmt.Println(min, max)
          fmt.Println(namedMinMax(value1, value2))
          min, max = namedMinMax(value1, value2)
          fmt.Println(min, max)
      }
      • The main() function takes no arguments and returns no arguments. Once the special function main() exits, the program automatically ends.

      • The doubleSquare() function requires a single int parameter and returns two int values, which is defined as (int, int).

      • All function arguments must have a name – variadic functions are the only exception to this rule.

      • If a function returns a single value, you do not need to put parenthesis around its type.

      • Because namedMinMax() has named return values in its signature, the min and max parameters are automatically returned in the order in which they were put in the function definition. Therefore, the function does not need to explicitly return any variables or values in its return statement at the end, and does not. minMax() function has the same functionality as namedMinMax() but it explicitly returns its values demonstrating that both ways are valid.

      • Both square and double variables in main() are assigned an anonymous function. However, nothing stops you from changing the value of square, double, or any other variable that holds the result of an anonymous function, afterwards. This means that both variables may have a different value in the future.

      1. Execute the functions.go program.

        go run functions.go
        

        Your output will resemble the following:

          
        The square of 10 is 100
        The double of 10 is 20
        20 100
        20 100
        -10 -1
        -10 -1
        -10 -1
        -10 -1
            
        

      Variadic functions

      Variadic functions are functions that accept a variable number of arguments. The most popular variadic functions in Go can be found in the fmt package. The code of variadic.go illustrates the creation and the use of variadic functions.

      ./variadic.go
       1
       2
       3
       4
       5
       6
       7
       8
       9
      10
      11
      12
      13
      14
      15
      16
      17
      18
      19
      20
      21
      22
      23
      24
      25
      26
      27
      28
      29
      30
      31
      
      package main
      
      import (
          "fmt"
      )
      
      func varFunc(input ...string) {
          fmt.Println(input)
      }
      
      func oneByOne(message string, sliceOfNumbers ...int) int {
          fmt.Println(message)
          sum := 0
          for indexInSlice, sliceElement := range sliceOfNumbers {
              fmt.Print(indexInSlice, sliceElement, "t")
              sum = sum + sliceElement
          }
          fmt.Println()
          sliceOfNumbers[0] = -1000
          return sum
      }
      
      func main() {
          many := []string{"12", "3", "b"}
          varFunc(many...)
          sum := oneByOne("Adding numbers...", 1, 2, 3, 4, 5, -1, 10)
          fmt.Println("Sum:", sum)
          sliceOfNumbers := []int{1, 2, 3}
          sum = oneByOne("Adding numbers...", sliceOfNumbers...)
          fmt.Println(sliceOfNumbers)
      }
      • The ... operator used as a prefix to a type like ...int is called the pack operator, whereas the unpack operator appends a slice like sliceOfNumbers.... A slice is a Go data type that is essentially an abstraction of an array of unspecified length.

      • Each variadic function can use the pack operator once. The oneByOne() function accepts a single string and a variable number of integer arguments using the sliceOfNumbers slice.

      • The varFunc function accepts a single argument and just calls the fmt.Println() function.

      • Another note about slices: the second call to oneByOne() is using a slice. Any changes you make to that slice inside the variadic function will persist after the function exits because this is how slices work in Go.

      1. Execute the variadic.go program:

        go run variadic.go
        

        The output will resemble the following

          
        [12 3 b]
        Adding numbers...
        0 1     1 2     2 3     3 4     4 5     5 -1     6 10
        Sum: 24
        Adding numbers...
        0 1     1 2     2 3
        [-1000 2 3]
            
        

      Functions and pointer variables

      Go supports pointers and this section will briefly present how functions can work with pointers. A future Go guide will talk about pointers in more detail, but here is a brief overview.

      ./fPointers.go
       1
       2
       3
       4
       5
       6
       7
       8
       9
      10
      11
      12
      13
      14
      15
      16
      17
      18
      19
      20
      21
      22
      23
      24
      25
      
      package main
      
      import (
          "fmt"
      )
      
      func getPointer(varToPointer *float64) float64 {
          return *varToPointer * *varToPointer
      }
      
      func returnPointer(testValue int) *int {
          squareTheTestValue := testValue * testValue
          return &squareTheTestValue
      }
      
      func main() {
          testValue := -12.12
          fmt.Println(getPointer(&testValue))
          testValue = -12
          fmt.Println(getPointer(&testValue))
      
          theSquare := returnPointer(10)
          fmt.Println("sq value:", *theSquare)
          fmt.Println("sq memory address:", theSquare)
      }
      • The getPointer() function takes a pointer argument to a float64, which is defined as varToPointer *float64, where returnPointer() returns a pointer to an int, which is declared as *int.
      1. Execute the fPointers.go program:

        go run fPointers.go
        

        The output will resemble the following:

          
        146.8944
        144
        sq value: 100
        sq memory address: 0xc00001a0b8
            
        

      Functions with Functions as Parameters

      Go functions can have functions as parameters.

      ./fArgF.go
       1
       2
       3
       4
       5
       6
       7
       8
       9
      10
      11
      12
      13
      14
      15
      16
      17
      18
      19
      20
      21
      
      package main
      
      import "fmt"
      
      func doubleIt(numToDouble int) int {
          return numToDouble + numToDouble
      }
      
      func squareIt(numToSquare int) int {
          return numToSquare * numToSquare
      }
      
      func funFun(functionName func(int) int, variableName int) int {
          return functionName(variableName)
      }
      
      func main() {
          fmt.Println("funFun Double:", funFun(doubleIt, 12))
          fmt.Println("funFun Square:", funFun(squareIt, 12))
          fmt.Println("Inline", funFun(func(numToCube int) int { return numToCube * numToCube * numToCube }, 12))
      }
      • The funFun() function accepts two parameters, a function parameter named functionName and an int value. The functionName parameter should be a function that takes one int argument and returns an int value.

      • The first fmt.Println() call in main() uses funFun() and passes the doubleIt function, without any parentheses, as its first parameter.

      • The second fmt.Println() call uses funFun() with squareIt as its first parameter.

      • In the last fmt.Println() statement the implementation of the function parameter is defined inside the call to funFun() using an anonymous function.

      1. Execute the fArgF.go program:

        go run fArgF.go
        

        The output will resemble the following:

          
        function1: 24
        function2: 144
        Inline 1728
            
        

      Functions Returning Functions

      Go functions can return functions.

      ./fRetF.go
       1
       2
       3
       4
       5
       6
       7
       8
       9
      10
      11
      12
      13
      14
      15
      16
      17
      18
      19
      20
      21
      22
      23
      
      package main
      
      import (
          "fmt"
      )
      
      func squareFunction() func() int {
          numToSquare := 0
          return func() int {
              numToSquare++
              return numToSquare * numToSquare
          }
      }
      
      func main() {
          square1 := squareFunction()
          square2 := squareFunction()
      
          fmt.Println("First Call to square1:", square1())
          fmt.Println("Second Call to square1:", square1())
          fmt.Println("First Call to square2:", square2())
          fmt.Println("Third Call to square1:", square1())
      }
      • squareFunction() returns an anonymous function with the func() int signature.

      • As squareFunction() is called two times, you will need to use two separate variables, square1 and square2 to keep the two return values.

      1. Execute the fRetF.go program:

        go run fRetF.go
        

        Your output will resemble the following:

          
        First Call to square1: 1
        Second Call to square1: 4
        First Call to square2: 1
        Third Call to square1: 9
            
        

        Notice that the values of square1 and square2 are not connected even though they both came from squareFunction().

      Errors in Go

      Errors and error handling are two important topics in Go. Go puts so much importance on error messages that it has a dedicated data type for errors, aptly named error. This also means that you can easily create your own error messages if you find that what Go gives you is not adequate. You will most likely need to create and handle your own errors when you are developing your own Go packages.

      Recognizing an error condition is one task, while deciding how to react to an error condition is another task. Therefore, some error conditions might require that you immediately stop the execution of the program, whereas in other error situations, you might just print a warning message and continue.

      ./errors.go
       1
       2
       3
       4
       5
       6
       7
       8
       9
      10
      11
      12
      13
      14
      15
      16
      17
      18
      19
      20
      21
      22
      23
      24
      25
      26
      27
      28
      29
      
      package main
      
      import (
          "errors"
          "fmt"
          "strconv"
      )
      
      func main() {
      
          customError := errors.New("My Custom Error!")
          if customError.Error() == "My Custom Error!" {
              fmt.Println("!!")
          }
      
          stringToConvert1 := "123"
          stringToConvert2 := "43W"
          _, err := strconv.Atoi(stringToConvert1)
          if err != nil {
              fmt.Println(err)
              return
          }
      
          _, err = strconv.Atoi(stringToConvert2)
          if err != nil {
              fmt.Println(err)
              return
          }
      }
      • The strconv.Atoi() function tries to convert a string into an integer, provided that the string is a valid integer, and returns two things, an integer value and an error variable. If the error variable is nil, then the conversion was successful and you get a valid integer. The _ character tells Go to ignore one, as in this case, or more of the return values of a function.

      • Most of the time, you need to check whether an error variable is equal to nil and then act accordingly. This kind of Go code is very popular in Go programs and you will see it and use it multiple times.

      • Also presented here is the errors.New() function that allows you to create a custom error message and errors.Error() function that allows you to convert an error variable into a string variable.

      1. Execute the errors.go program:

        go run errors.go
        

        Your output will resemble the following:

          
        !!
        strconv.Atoi: parsing "43W": invalid syntax
            
        

      Summary

      In this guide you learned the basics about the Go programming language, how to execute programs, how to write loops, how to handle errors, and you saw examples for various function types.

      More Information

      You may wish to consult the following resources for additional information on this topic. While these are provided in the hope that they will be useful, please note that we cannot vouch for the accuracy or timeliness of externally hosted materials.

      Find answers, ask questions, and help others.

      This guide is published under a CC BY-ND 4.0 license.



      Source link

      Learning to Use the ss Tool to its Full Potential


      Updated by Linode

      Contributed by

      Mihalis Tsoukalos

      Introduction

      The study of socket connections is important for every
      UNIX and network administrator because it allows you to better understand your Linux system’s status. Written by Alexey Kuznetosv to replace the famous netstat utility , the more capable ss (socket statistics) utility allows you to monitor TCP, UDP, and UNIX sockets. The purpose of this guide is to help you learn the ss utility and to use it as productively as possible.

      Note

      Running ss without using the sudo utility will result in different output. Practically, this means that running ss without root privileges will show the results available to the current user only. If you are not familiar with the sudo command,
      see the Users and Groups guide.

      Command Line Options

      The ss(8) binary supports many command line options, including the following:

      Option Definition
      -h The -h option shows a summary of all options.
      -V The -V option displays the version of ss
      -H The -H option tells ss to suppress the header line – this is useful when you want to process the generated output using a scripting language.
      -t The -t parameter tells ss to show TCP connections only.
      -u The –u parameter tells ss to show UDP connections only.
      -d The –d parameter tells ss to show DCCP sockets only.
      -S The –S parameter tells ss to show SCTP sockets only.
      -a The -a option tells ss to display both listening and non-listening sockets of every kind.
      -l The -l parameter tells ss to display listening sockets, which are omitted by default.
      -e The -e option tells ss to display detailed socket information.
      -x The -x parameter tells ss to display UNIX domain sockets only.
      -A The -A option allows you to select the socket types that you want to see. The -A option accepts the following set of identifiers that can be combined and separated by commas: all, inet, tcp, udp, raw, unix, packet, netlink, unix_dgram, unix_stream, unix_seqpacket, packet_raw and packet_dgram.
      -4 The -4 command line option tells ss to display IPv4 connections only.
      -6 The -6 command line option tells ss to display IPv6 connections only.
      -f FAMILY The -f tells ss to display sockets of type FAMILY. The supported values are unix, inet, inet6 and netlink.
      -s The -s option displays useful statistics about the current connections.
      -o The -o option displays timer information. There are five types of timers: on, which is either a TCP retrans timer, a TCP early retrans timer, or a tail loss probe timer; keepalive, which is the TCP keep alive timer; timewait, which is the timewait stage timer; persist, which is the zero window probe timer; and unknown, which is a timer that is none of the other timers.
      -n The -n option tells ss to disable the resolving of service names.
      -r The -r option tells ss to enable DNS resolving in the output, which is turned off by default.
      -m The -m parameter tells ss to display socket memory usage information.
      -p The -p parameter tells ss to display the process that is using a socket.
      -D FILE The -D parameter tells ss to save the output in the FILE file.

      Note

      The -A tcp option is equivalent to -t, the -A udp option is equivalent to -u and the –A unix
      option is equivalent to -x.

      Installing ss

      The ss tool is part of the IPROUTE2 Utility Suite. Since the ss command line tool is usually
      installed by default, you will not need to install it yourself. On a Debian Linux system, you can
      find the ss executable inside /bin.

      If for some reason ss is not installed on your Linux system, you should install the iproute2
      package using your favorite package manager.

      Examples

      Basic Usage

      The simplest way to use ss is without any command line parameters. When ss is
      used without any command line arguments, it prints all TCP, UDP and socket connections.
      The list might get big on busy machines, which means that it can become more difficult to parse – the output of wc(1), (a word count utility), shows that the list is long yet manageable:

      ss | wc
      
        
           94     750    7926
      
      

      If you also use the -a parameter to show all listening and non-listening sockets, the output will be much higher:

      ss -a | wc
      
        
          224    1682   19562
      
      

      Listing Sockets

      TCP

      The following command displays all listening and non-listening (-a) TCP (-t) sockets:

      ss -t -a
      
        
      State    Recv-Q  Send-Q  Local Address:Port   Peer Address:Port
      LISTEN   0       80      127.0.0.1:mysql      *:*
      LISTEN   0       128     *:ssh                *:*
      LISTEN   0       100     *:smtp               *:*
      ESTAB    0       204     109.74.193.253:ssh   2.86.7.61:55137
      LISTEN   0       128     :::http              :::*
      LISTEN   0       128     :::ssh               :::*
      LISTEN   0       128     :::https             :::*
      
      

      The output is separated into columns. The first column, state, shows the state of the TCP connection. As the example is using the -a
      option, both listening and non-listening states are included in the output.
      The second and the third columns, Recv-Q and Send-Q, show the amount of data queued for receive and
      transmit operations. The Local Address:Port column shows the IP address the process
      listens to as well as the port number that is used – you can connect the name of the
      service with a numeric value by looking at the /etc/services file. The last column, Peer Address:Port, is useful when there is an active connection
      because it shows the address and port number of the client machine, though here it is without any real values for TCP connections that are in the
      LISTEN state.
      As the -r option is not used, you only see IP addresses in the output.

      Running ss -t without –a will display established TCP connections only:

      ss -t
      
        
      State  Recv-Q  Send-Q  Local Address:Port  Peer Address:Port
      ESTAB  0       204     109.74.193.253:ssh  2.86.7.61:55137
      
      

      UDP

      The following command displays all UDP (-u) sockets:

      ss -u -a
      
        
      State    Recv-Q  Send-Q  Local Address:Port                  Peer Address:Port
      UNCONN   0       0       *:mdns                              *:*
      UNCONN   1536    0       109.74.193.253:syslog               *:*
      UNCONN   0       0       *:54087                             *:*
      UNCONN   0       0       *:bootpc                            *:*
      UNCONN   0       0       109.74.193.253:ntp                  *:*
      UNCONN   0       0       127.0.0.1:ntp                       *:*
      UNCONN   0       0       *:ntp                               *:*
      UNCONN   0       0       :::mdns                             :::*
      UNCONN   0       0       :::48582                            :::*
      UNCONN   0       0       fe80::f03c:91ff:fe69:1381%eth0:ntp  :::*
      UNCONN   0       0       2a01:7e00::f03c:91ff:fe69:1381:ntp  :::*
      UNCONN   0       0       ::1:ntp                             :::*
      UNCONN   0       0       :::ntp                              :::*
      
      

      Running ss -u without –a will display established UDP connections only. In this case there are no established UDP connections:

      ss -u
      
        
      Recv-Q Send-Q  Local Address:Port  Peer Address:Port
      
      

      Display Statistics

      You can display statistics about the current connections using the -s option:

      ss -s
      
        
      Total: 199 (kernel 228)
      TCP:   9 (estab 1, closed 2, orphaned 0, synrecv 0, timewait 0/0), ports 0
      
      Transport  Total  IP  IPv6
      *          228    -   -
      RAW        0      0   0
      UDP        13     7   6
      TCP        7      4   3
      INET       20     11  9
      FRAG       0      0   0
      
      

      Filter by TCP State

      ss allows you to filter its output by state using the state and exclude keywords
      followed by a state identifier. The state keyword displays output that matches the
      provided identifier, whereas the exclude keyword displays everything except the output
      that matches the identifier.

      The use of state is illustrated in the next example:

      ss -t4 state established
      
        
      Recv-Q  Send-Q  Local Address:Port  Peer Address:Port
      0       0       109.74.193.253:ssh  2.86.7.61:55137
      
      

      The use of exclude is illustrated in the next example:

      ss -t4 exclude established
      
        
      State      Recv-Q  Send-Q  Local Address:Port   Peer Address:Port
      LISTEN     0       80      127.0.0.1:mysql      *:*
      LISTEN     0       128     *:ssh                *:*
      LISTEN     0       100     *:smtp               *:*
      TIME-WAIT  0       0       109.74.193.253:smtp  103.89.91.73:55668
      
      

      The -t4 command option returns IPv4 TCP connections.

      Filter Output by IP Address and Port Number

      The more you filter the output of ss, the more accurate and relevant information you will receive. There exist two ss options that allow
      you to include connections from certain IP addresses and port numbers.

      The following command shows traffic from a given IP address only, using the
      dst keyword:

      ss -nt dst 2.86.7.61
      
        
      State        Recv-Q  Send-Q  Local Address:Port         Peer Address:Port
      ESTAB        0      0        109.74.193.253:22          2.86.7.61:55137
      FIN-WAIT-1   0       32      ::ffff:109.74.193.253:443  ::ffff:2.86.7.61:56075
      ESTAB        0       0       ::ffff:109.74.193.253:443  ::ffff:2.86.7.61:56077
      ESTAB        0       0       ::ffff:109.74.193.253:443  ::ffff:2.86.7.61:56074
      ESTAB        0       0       ::ffff:109.74.193.253:443  ::ffff:2.86.7.61:56078
      
      

      If you want to display traffic from an entire network, you can replace the IP address with
      a network address such as 2.86.7/24.

      The following command displays information about the HTTP and the HTTPS protocols, which
      are associated with port numbers 80 and 443 as defined in /etc/services:

      ss -at '( dport = :http or dport = :https or sport = :http or sport = :https )'
      
        
      State      Recv-Q  Send-Q  Local Address:Port           Peer Address:Port
      LISTEN     0       128     :::http                      :::*
      LISTEN     0       128     :::https                     :::*
      ESTAB      0       0       ::ffff:109.74.193.253:https  ::ffff:2.86.7.61:56046
      ESTAB      0       0       ::ffff:109.74.193.253:https  ::ffff:2.86.7.61:56055
      ESTAB      0       0       ::ffff:109.74.193.253:https  ::ffff:2.86.7.61:56047
      ESTAB      0       0       ::ffff:109.74.193.253:https  ::ffff:2.86.7.61:56054
      ESTAB      0       0       ::ffff:109.74.193.253:https  ::ffff:2.86.7.61:56056
      ESTAB      0       0       ::ffff:109.74.193.253:https  ::ffff:2.86.7.61:56057
      TIME-WAIT  0       0       ::ffff:109.74.193.253:http   ::ffff:54.39.151.52:59854
      
      

      dport means destination port and sport means source port.

      The following command is equivalent to the previous command:

      ss -at '( dport = :80 or dport = :443 or sport = :80 or sport = :443 )'
      

      Display Timer Information

      The -o option displays timer information:

      ss -nt dst 2.86.7.61 -o
      
        
      State  Recv-Q  Send-Q  Local Address:Port  Peer Address:Port
      ESTAB  0       0       109.74.193.253:22   2.86.7.61:55137     timer:(keepalive,72min,0)
      
      

      Enable IP Address Resolving

      The -r parameter enables IP address resolving, which returns the domain names of the IP addresses:

      ss -r -t
      
        
      State  Recv-Q  Send-Q    Local Address:Port                  Peer Address:Port
      ESTAB  0       168       li140-253.members.linode.com:ssh    ppp-2-86-7-61.home.otenet.gr:50939
      ESTAB  0       0         li140-253.members.linode.com:https  ::ffff:216.244.66.228:37668
      
      

      Note

      A side effect of the -r command line option is that it slows the execution of
      the ss command due to the DNS lookups that need to be performed.

      Display Detailed Socket Information

      The -e option tells ss to display detailed socket information. The -e option
      is illustrated in the following example:

      ss -t -e
      
        
      State  Recv-Q  Send-Q  Local Address:Port  Peer Address:Port
      ESTAB  0       0       109.74.193.253:ssh  2.86.7.61:62897    timer:(keepalive,54min,0) ino:10195329 sk:11e 
      
      

      Show a Connection’s UNIX Process

      The -p option displays the process ID(s) and the process name of a connection:

      ss -t -p
      
        
      State  Recv-Q  Send-Q  Local Address:Port           Peer Address:Port
      ESTAB  0       204     109.74.193.253:ssh           2.86.7.61:55137            users:(("sshd",pid=3964,fd=3),("sshd",pid=3951,fd=3))
      ESTAB  0       51      ::ffff:109.74.193.253:https  ::ffff:176.9.146.74:57536  users:(("apache2",pid=30871,fd=29))
      
      

      The following command shows SSH-related processes on the current machine:

      ss -t -p -a | grep ssh
      
        
      LISTEN  0  128  *:ssh               *:*                    users:(("sshd",pid=812,fd=3))
      ESTAB   0  36   109.74.193.253:ssh  2.86.7.61:55137        users:(("sshd",pid=3964,fd=3),("sshd",pid=3951,fd=3))
      ESTAB   0  0    109.74.193.253:ssh  138.197.140.194:41992  users:(("sshd",pid=8538,fd=3),("sshd",pid=8537,fd=3))
      LISTEN  0  128  :::ssh              :::*                   users:(("sshd",pid=812,fd=4))
      
      

      Find Which Process is Using a Given Port Number

      With the help of ss and grep(1), you can discover which process is using
      a given port number:

      ss -tunap | grep :80
      
        
      tcp  LISTEN  0  128  :::80 :::*  users:(("apache2",pid=8772,fd=4),("apache2",pid=8717,fd=4),("apache2",pid=8715,fd=4),("apache2",pid=8714,fd=4),("apache2",pid=8713,fd=4),("apache2",pid=8712,fd=4),("apache2",pid=8711,fd=4),("apache2",pid=8709,fd=4))
      
      

      As Apache uses multiple child processes, you receive a list of processes for port number 80.

      The next command will do exactly the same thing without using grep(1):

      ss -tup -a sport = :80
      
        
      Netid  State   Recv-Q  Send-Q  Local Address:Port  Peer Address:Port
      tcp    LISTEN  0       128     :::http             :::*               users:(("apache2",pid=8715,fd=4),("apache2",pid=8714,fd=4),("apache2",pid=8713,fd=4),("apache2",pid=8712,fd=4),("apache2",pid=8711,fd=4),("apache2",pid=8709,fd=4))
      
      

      Find Open Ports Above Port Number 1024

      ss supports ranges when working with port numbers. This feature is illustrated in
      the following example that finds open port above port number 1024:

      ss -t -a sport > :1024
      
        
      State   Recv-Q  Send-Q  Local Address:Port  Peer Address:Port
      LISTEN  0       80      127.0.0.1:mysql     *:*
      
      

      Note

      The ss -t -a sport > :1024 command can be also written as ss -t -a sport '> :1024'.

      Search for Specific TCP Characteristics

      The following command shows all TCP connections that use IPv4 that are in
      listening state, as well as the name of the process using the socket without
      resolving the IP addresses and the port number:

      ss -t -4nlp
      
        
      State   Recv-Q  Send-Q  Local Address:Port  Peer Address:Port
      LISTEN  0       80      127.0.0.1:3306      *:*                users:(("mysqld",pid=1003,fd=17))
      LISTEN  0       128     *:22                *:*                users:(("sshd",pid=812,fd=3))
      LISTEN  0       100     *:25                *:*                users:(("smtpd",pid=9011,fd=6),("master",pid=1245,fd=13))
      
      

      The following command shows all SSH related connections and sockets:

      ss -at '( dport = :ssh or sport = :ssh )'
      
        
      State   Recv-Q  Send-Q  Local Address:Port  Peer Address:Port
      LISTEN  0       128     *:ssh               *:*
      ESTAB   0       0       109.74.193.253:ssh  2.86.7.61:64363
      LISTEN  0       128     :::ssh              :::*
      
      

      Show Sockets in a Listening State

      The following command shows TCP sockets in listening (-l) state:

      ss -l -t
      
        
      State   Recv-Q  Send-Q  Local Address:Port  Peer Address:Port
      LISTEN  0       80      127.0.0.1:mysql     *:*
      LISTEN  0       128     *:ssh              *:*
      LISTEN  0       100     *:smtp              *:*
      LISTEN  0       128     :::http             :::*
      LISTEN  0       128     :::ssh              :::*
      LISTEN  0       128     :::https            :::*
      
      

      The following command shows IPv4 UDP sockets in listening state:

      ss -l -u -4
      
        
      State   Recv-Q  Send-Q  Local Address:Port      Peer Address:Port
      UNCONN  0       0       *:mdns                  *:*
      UNCONN  1536    0        109.74.193.253:syslog  *:*
      UNCONN  0       0        *:54087                *:*
      UNCONN  0       0        *:bootpc               *:*
      UNCONN  0       0        109.74.193.253:ntp     *:*
      UNCONN  0       0        127.0.0.1:ntp          *:*
      UNCONN  0       0        *:ntp                  *:*
      
      

      Advanced Filtering with ss

      The following ss command will list all TCP sockets that are in the ESTABLISHED state, use HTTP or HTTPS on the local machine and belong to the 2.86.7/24 network and display their timers:

      ss -o state established '( sport = :http or sport = :https )' dst 2.86.7/24
      
        
      Netid  Recv-Q  Send-Q  Local Address:Port           Peer Address:Port
      tcp    0       0       ::ffff:109.74.193.253:https  ::ffff:2.86.7.61:63057  timer:(keepalive,119min,0)
      tcp    0       0       ::ffff:109.74.193.253:https  ::ffff:2.86.7.61:63053  timer:(keepalive,119min,0)
      tcp    0       0       ::ffff:109.74.193.253:https  ::ffff:2.86.7.61:63055  timer:(keepalive,119min,0)
      tcp    0       0       ::ffff:109.74.193.253:https  ::ffff:2.86.7.61:63054  timer:(keepalive,119min,0)
      tcp    0       0       ::ffff:109.74.193.253:https  ::ffff:2.86.7.61:63052  timer:(keepalive,119min,0)
      tcp    0       0       ::ffff:109.74.193.253:https  ::ffff:2.86.7.61:63056  timer:(keepalive,119min,0)
      
      

      Apart from the standard TCP state names (established,
      syn-sent, syn-recv, fin-wait-1, fin-wait-2, time-wait, closed, close-wait, last-ack,
      listen and closing), you can also use the following states:

      • all: For all the states.
      • bucket: For TCP minisockets (TIME-WAIT|SYN-RECV) states.
      • big: For all states except for minisockets – this is the opposite of bucket.
      • connected: For the not closed and not listening states.
      • synchronized: For connected and not SYN-SENT states.

      Using AWK to Process ss Output

      The following command displays a summary of all sockets based on their state:

      ss -t -u -a | awk '{print $1}' | grep -v State | sort | uniq -c | sort -nr
      
        
           13 udp
            7 tcp
            1 Netid
      
      

      The following command displays a summary of all sockets based on their protocol:

      ss -a | awk '{print $1}' | grep -v State | sort | uniq -c | sort -nr
      
        
          133 u_str
           37 u_dgr
           34 nl
           13 udp
            8 tcp
            1 u_seq
            1 p_raw
            1 Netid
      
      

      The last command will create a summary of all IPv6 TCP connections that are in
      the CONNECTED state:

      ss -t6 state connected | awk '{print $1}' | grep -v State | sort | uniq -c | sort -nr
      

      More Information

      You may wish to consult the following resources for additional information on this topic. While these are provided in the hope that they will be useful, please note that we cannot vouch for the accuracy or timeliness of externally hosted materials.

      Find answers, ask questions, and help others.

      This guide is published under a CC BY-ND 4.0 license.



      Source link