Why do look up tables speed things up compared to brute force?












4












$begingroup$


I'm currently reading up on lookup tables and efficiency. In my uni script it says the following:



For Brute Force:




  • Preparation time: $O(1)$


  • Disk space requirement: $O(1)$


  • Time required to crack the password: $O(2^n)$



Full lookup table:




  • Preparation time: $O(2^n)$


  • Disk space requirement: $O(2^n)$


  • Time required to crack the password: $O(1)$



I'm not sure I'm understanding this correctly.



As far as I understand, for Brute Force:



$O(1)$ is the time required to look up the password in my table of all possible passwords.
The disk space requirement is simply as large as needed for a list of all possible passwords, so $O(1)$ as well



My questions:



Why is the time needed to crack the password $O(2^n)$? How is it determined?



It seems I don't understand the concept of a lookup table either. As far as I see, a lookup table will simply hash the full list of possible passwords (so required disk space is larger) but what then? Why is the cracking of the password per se faster?



I think I'm missing something crucial here. Any help will be appreciated.










share|improve this question









New contributor




Fang is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.







$endgroup$








  • 1




    $begingroup$
    n-bit input so you need to look for $2^n$ possible passwords.
    $endgroup$
    – kelalaka
    2 days ago












  • $begingroup$
    @kelalaka: Thanks! I can follow that, at least.
    $endgroup$
    – Fang
    2 days ago
















4












$begingroup$


I'm currently reading up on lookup tables and efficiency. In my uni script it says the following:



For Brute Force:




  • Preparation time: $O(1)$


  • Disk space requirement: $O(1)$


  • Time required to crack the password: $O(2^n)$



Full lookup table:




  • Preparation time: $O(2^n)$


  • Disk space requirement: $O(2^n)$


  • Time required to crack the password: $O(1)$



I'm not sure I'm understanding this correctly.



As far as I understand, for Brute Force:



$O(1)$ is the time required to look up the password in my table of all possible passwords.
The disk space requirement is simply as large as needed for a list of all possible passwords, so $O(1)$ as well



My questions:



Why is the time needed to crack the password $O(2^n)$? How is it determined?



It seems I don't understand the concept of a lookup table either. As far as I see, a lookup table will simply hash the full list of possible passwords (so required disk space is larger) but what then? Why is the cracking of the password per se faster?



I think I'm missing something crucial here. Any help will be appreciated.










share|improve this question









New contributor




Fang is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.







$endgroup$








  • 1




    $begingroup$
    n-bit input so you need to look for $2^n$ possible passwords.
    $endgroup$
    – kelalaka
    2 days ago












  • $begingroup$
    @kelalaka: Thanks! I can follow that, at least.
    $endgroup$
    – Fang
    2 days ago














4












4








4


2



$begingroup$


I'm currently reading up on lookup tables and efficiency. In my uni script it says the following:



For Brute Force:




  • Preparation time: $O(1)$


  • Disk space requirement: $O(1)$


  • Time required to crack the password: $O(2^n)$



Full lookup table:




  • Preparation time: $O(2^n)$


  • Disk space requirement: $O(2^n)$


  • Time required to crack the password: $O(1)$



I'm not sure I'm understanding this correctly.



As far as I understand, for Brute Force:



$O(1)$ is the time required to look up the password in my table of all possible passwords.
The disk space requirement is simply as large as needed for a list of all possible passwords, so $O(1)$ as well



My questions:



Why is the time needed to crack the password $O(2^n)$? How is it determined?



It seems I don't understand the concept of a lookup table either. As far as I see, a lookup table will simply hash the full list of possible passwords (so required disk space is larger) but what then? Why is the cracking of the password per se faster?



I think I'm missing something crucial here. Any help will be appreciated.










share|improve this question









New contributor




Fang is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.







$endgroup$




I'm currently reading up on lookup tables and efficiency. In my uni script it says the following:



For Brute Force:




  • Preparation time: $O(1)$


  • Disk space requirement: $O(1)$


  • Time required to crack the password: $O(2^n)$



Full lookup table:




  • Preparation time: $O(2^n)$


  • Disk space requirement: $O(2^n)$


  • Time required to crack the password: $O(1)$



I'm not sure I'm understanding this correctly.



As far as I understand, for Brute Force:



$O(1)$ is the time required to look up the password in my table of all possible passwords.
The disk space requirement is simply as large as needed for a list of all possible passwords, so $O(1)$ as well



My questions:



Why is the time needed to crack the password $O(2^n)$? How is it determined?



It seems I don't understand the concept of a lookup table either. As far as I see, a lookup table will simply hash the full list of possible passwords (so required disk space is larger) but what then? Why is the cracking of the password per se faster?



I think I'm missing something crucial here. Any help will be appreciated.







brute-force-attack complexity






share|improve this question









New contributor




Fang is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.











share|improve this question









New contributor




Fang is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.









share|improve this question




share|improve this question








edited yesterday









Journeyman Geek

1032




1032






New contributor




Fang is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.









asked 2 days ago









FangFang

12316




12316




New contributor




Fang is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.





New contributor





Fang is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.






Fang is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.








  • 1




    $begingroup$
    n-bit input so you need to look for $2^n$ possible passwords.
    $endgroup$
    – kelalaka
    2 days ago












  • $begingroup$
    @kelalaka: Thanks! I can follow that, at least.
    $endgroup$
    – Fang
    2 days ago














  • 1




    $begingroup$
    n-bit input so you need to look for $2^n$ possible passwords.
    $endgroup$
    – kelalaka
    2 days ago












  • $begingroup$
    @kelalaka: Thanks! I can follow that, at least.
    $endgroup$
    – Fang
    2 days ago








1




1




$begingroup$
n-bit input so you need to look for $2^n$ possible passwords.
$endgroup$
– kelalaka
2 days ago






$begingroup$
n-bit input so you need to look for $2^n$ possible passwords.
$endgroup$
– kelalaka
2 days ago














$begingroup$
@kelalaka: Thanks! I can follow that, at least.
$endgroup$
– Fang
2 days ago




$begingroup$
@kelalaka: Thanks! I can follow that, at least.
$endgroup$
– Fang
2 days ago










2 Answers
2






active

oldest

votes


















14












$begingroup$

Suppose you have an $n$-bit key. Suppose further you have some reliable predicate $P(k,m)$ which decides whether a key $k$ is the key you are looking for given the reference $m$. Furthermore, suppose you have a "successor" function $F$, that takes a key and returns you the "next" key so that eventually all keys are traversed.



Now what a brute-force attack does is, given $m$, it takes a starting key $k$, evaluates $P(k,m)$ if it returns false, sets $kgets F(k)$. This will take $mathcal O(2^n)$ evaluations of $P$ and $F$. Assuming both can be computed in $mathcal O(1)$ the entire calculation takes $mathcal O(2^n)$.



Now for the look-up table. For this you simply start at some key $k$, and store the $m$ such that $P(k,m)$ yields true as a pair $(m,k)$ in a giant array (or hashmap). Now you repeat this with $F(k)$. When you then get a concrete $m$, you simply look it up in the table and find the corresponding $k$. For hashmaps and continuous arrays, this takes time $mathcal O(1)$ (ie the lookup time is independent of the number of elements). Clearly, you need storage for all the $2^n$ pairs of $(m,k)$. Now the preparation time assumes that you can find $m$ such that $P(k,m)$ is true, but this is usually possible in constant time, e.g. by evaluating a hash function or an encryption cipher.






share|improve this answer











$endgroup$













  • $begingroup$
    For Hash table hash table runtime complexity; insert, search and delete
    $endgroup$
    – kelalaka
    2 days ago



















2












$begingroup$

Brute Force



Let's say your password was all digits (e.g. 607552)



If your password has length 1, it will take up to 10 attempts to crack:



0, 1, 2, 3 ..., 9


If your password has length 2, it will take up to 100 attempts to crack



00, 01, 02, ... 10, 11, 12, ... 20, 21, 22, ... 99


If your password has length 3 it will take up to 1000 attempts to crack



See the pattern? The number of attempts is $10^n$. It's the same idea if you are cracking a password with $n$ bits. (Since a bit can be 0 or 1, instead of 0-9, it's a power of 2 instead of 10).



Lookup Table



While a brute-force attack tries all possible passwords to see if they produce the right hash, a lookup table is a fast way to do the opposite: get a password based on a hash. It does this by hashing all the possible passwords ahead of time, and storing the results in a table.



Method 1:



Store the items in an array, sorted by hash. When given a hash, guess the position in the array of this hash, based on the numerical value of the hash. (I.e. if the numerical value is 25% of the maximum value the hash could be, then you would guess the index that is 25% of the total array size.) Your guess will be very close, because any good hash is evenly distributed. Go forwards or backwards until you find a match. This is freakishly close to $O(1)$, especially if you refine your guesses (binary-search-style) rather than just incrementing or decrementing.



The main disadvantage of this method is that initially sorting the table is asymptotically slower than generating it, namely $O(2^n times log (2^n)) = O(n 2^n)$.



Method 2:



This is essentially a hashmap, except using a library hashmap is silly when you already have a hash! Just use the first $n$ digits of your hash. That means: create an array of size $2^n$. In each slot goes a list of all hashes whose first $n$ digits equal that array index. (And their corresponding passwords.)



To look up a hash, take its first $n$ digits, and get the corresponding list out of the array. Traverse the list to find the hash. Like a hashmap, this takes $O(1)$ constant amortised time.






share|improve this answer








New contributor




Artelius is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.






$endgroup$













    Your Answer





    StackExchange.ifUsing("editor", function () {
    return StackExchange.using("mathjaxEditing", function () {
    StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix) {
    StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
    });
    });
    }, "mathjax-editing");

    StackExchange.ready(function() {
    var channelOptions = {
    tags: "".split(" "),
    id: "281"
    };
    initTagRenderer("".split(" "), "".split(" "), channelOptions);

    StackExchange.using("externalEditor", function() {
    // Have to fire editor after snippets, if snippets enabled
    if (StackExchange.settings.snippets.snippetsEnabled) {
    StackExchange.using("snippets", function() {
    createEditor();
    });
    }
    else {
    createEditor();
    }
    });

    function createEditor() {
    StackExchange.prepareEditor({
    heartbeatType: 'answer',
    autoActivateHeartbeat: false,
    convertImagesToLinks: false,
    noModals: true,
    showLowRepImageUploadWarning: true,
    reputationToPostImages: null,
    bindNavPrevention: true,
    postfix: "",
    imageUploader: {
    brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
    contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
    allowUrls: true
    },
    noCode: true, onDemand: true,
    discardSelector: ".discard-answer"
    ,immediatelyShowMarkdownHelp:true
    });


    }
    });






    Fang is a new contributor. Be nice, and check out our Code of Conduct.










    draft saved

    draft discarded


















    StackExchange.ready(
    function () {
    StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fcrypto.stackexchange.com%2fquestions%2f66461%2fwhy-do-look-up-tables-speed-things-up-compared-to-brute-force%23new-answer', 'question_page');
    }
    );

    Post as a guest















    Required, but never shown

























    2 Answers
    2






    active

    oldest

    votes








    2 Answers
    2






    active

    oldest

    votes









    active

    oldest

    votes






    active

    oldest

    votes









    14












    $begingroup$

    Suppose you have an $n$-bit key. Suppose further you have some reliable predicate $P(k,m)$ which decides whether a key $k$ is the key you are looking for given the reference $m$. Furthermore, suppose you have a "successor" function $F$, that takes a key and returns you the "next" key so that eventually all keys are traversed.



    Now what a brute-force attack does is, given $m$, it takes a starting key $k$, evaluates $P(k,m)$ if it returns false, sets $kgets F(k)$. This will take $mathcal O(2^n)$ evaluations of $P$ and $F$. Assuming both can be computed in $mathcal O(1)$ the entire calculation takes $mathcal O(2^n)$.



    Now for the look-up table. For this you simply start at some key $k$, and store the $m$ such that $P(k,m)$ yields true as a pair $(m,k)$ in a giant array (or hashmap). Now you repeat this with $F(k)$. When you then get a concrete $m$, you simply look it up in the table and find the corresponding $k$. For hashmaps and continuous arrays, this takes time $mathcal O(1)$ (ie the lookup time is independent of the number of elements). Clearly, you need storage for all the $2^n$ pairs of $(m,k)$. Now the preparation time assumes that you can find $m$ such that $P(k,m)$ is true, but this is usually possible in constant time, e.g. by evaluating a hash function or an encryption cipher.






    share|improve this answer











    $endgroup$













    • $begingroup$
      For Hash table hash table runtime complexity; insert, search and delete
      $endgroup$
      – kelalaka
      2 days ago
















    14












    $begingroup$

    Suppose you have an $n$-bit key. Suppose further you have some reliable predicate $P(k,m)$ which decides whether a key $k$ is the key you are looking for given the reference $m$. Furthermore, suppose you have a "successor" function $F$, that takes a key and returns you the "next" key so that eventually all keys are traversed.



    Now what a brute-force attack does is, given $m$, it takes a starting key $k$, evaluates $P(k,m)$ if it returns false, sets $kgets F(k)$. This will take $mathcal O(2^n)$ evaluations of $P$ and $F$. Assuming both can be computed in $mathcal O(1)$ the entire calculation takes $mathcal O(2^n)$.



    Now for the look-up table. For this you simply start at some key $k$, and store the $m$ such that $P(k,m)$ yields true as a pair $(m,k)$ in a giant array (or hashmap). Now you repeat this with $F(k)$. When you then get a concrete $m$, you simply look it up in the table and find the corresponding $k$. For hashmaps and continuous arrays, this takes time $mathcal O(1)$ (ie the lookup time is independent of the number of elements). Clearly, you need storage for all the $2^n$ pairs of $(m,k)$. Now the preparation time assumes that you can find $m$ such that $P(k,m)$ is true, but this is usually possible in constant time, e.g. by evaluating a hash function or an encryption cipher.






    share|improve this answer











    $endgroup$













    • $begingroup$
      For Hash table hash table runtime complexity; insert, search and delete
      $endgroup$
      – kelalaka
      2 days ago














    14












    14








    14





    $begingroup$

    Suppose you have an $n$-bit key. Suppose further you have some reliable predicate $P(k,m)$ which decides whether a key $k$ is the key you are looking for given the reference $m$. Furthermore, suppose you have a "successor" function $F$, that takes a key and returns you the "next" key so that eventually all keys are traversed.



    Now what a brute-force attack does is, given $m$, it takes a starting key $k$, evaluates $P(k,m)$ if it returns false, sets $kgets F(k)$. This will take $mathcal O(2^n)$ evaluations of $P$ and $F$. Assuming both can be computed in $mathcal O(1)$ the entire calculation takes $mathcal O(2^n)$.



    Now for the look-up table. For this you simply start at some key $k$, and store the $m$ such that $P(k,m)$ yields true as a pair $(m,k)$ in a giant array (or hashmap). Now you repeat this with $F(k)$. When you then get a concrete $m$, you simply look it up in the table and find the corresponding $k$. For hashmaps and continuous arrays, this takes time $mathcal O(1)$ (ie the lookup time is independent of the number of elements). Clearly, you need storage for all the $2^n$ pairs of $(m,k)$. Now the preparation time assumes that you can find $m$ such that $P(k,m)$ is true, but this is usually possible in constant time, e.g. by evaluating a hash function or an encryption cipher.






    share|improve this answer











    $endgroup$



    Suppose you have an $n$-bit key. Suppose further you have some reliable predicate $P(k,m)$ which decides whether a key $k$ is the key you are looking for given the reference $m$. Furthermore, suppose you have a "successor" function $F$, that takes a key and returns you the "next" key so that eventually all keys are traversed.



    Now what a brute-force attack does is, given $m$, it takes a starting key $k$, evaluates $P(k,m)$ if it returns false, sets $kgets F(k)$. This will take $mathcal O(2^n)$ evaluations of $P$ and $F$. Assuming both can be computed in $mathcal O(1)$ the entire calculation takes $mathcal O(2^n)$.



    Now for the look-up table. For this you simply start at some key $k$, and store the $m$ such that $P(k,m)$ yields true as a pair $(m,k)$ in a giant array (or hashmap). Now you repeat this with $F(k)$. When you then get a concrete $m$, you simply look it up in the table and find the corresponding $k$. For hashmaps and continuous arrays, this takes time $mathcal O(1)$ (ie the lookup time is independent of the number of elements). Clearly, you need storage for all the $2^n$ pairs of $(m,k)$. Now the preparation time assumes that you can find $m$ such that $P(k,m)$ is true, but this is usually possible in constant time, e.g. by evaluating a hash function or an encryption cipher.







    share|improve this answer














    share|improve this answer



    share|improve this answer








    edited 2 days ago









    kelalaka

    6,23022142




    6,23022142










    answered 2 days ago









    SEJPMSEJPM

    28.7k555134




    28.7k555134












    • $begingroup$
      For Hash table hash table runtime complexity; insert, search and delete
      $endgroup$
      – kelalaka
      2 days ago


















    • $begingroup$
      For Hash table hash table runtime complexity; insert, search and delete
      $endgroup$
      – kelalaka
      2 days ago
















    $begingroup$
    For Hash table hash table runtime complexity; insert, search and delete
    $endgroup$
    – kelalaka
    2 days ago




    $begingroup$
    For Hash table hash table runtime complexity; insert, search and delete
    $endgroup$
    – kelalaka
    2 days ago











    2












    $begingroup$

    Brute Force



    Let's say your password was all digits (e.g. 607552)



    If your password has length 1, it will take up to 10 attempts to crack:



    0, 1, 2, 3 ..., 9


    If your password has length 2, it will take up to 100 attempts to crack



    00, 01, 02, ... 10, 11, 12, ... 20, 21, 22, ... 99


    If your password has length 3 it will take up to 1000 attempts to crack



    See the pattern? The number of attempts is $10^n$. It's the same idea if you are cracking a password with $n$ bits. (Since a bit can be 0 or 1, instead of 0-9, it's a power of 2 instead of 10).



    Lookup Table



    While a brute-force attack tries all possible passwords to see if they produce the right hash, a lookup table is a fast way to do the opposite: get a password based on a hash. It does this by hashing all the possible passwords ahead of time, and storing the results in a table.



    Method 1:



    Store the items in an array, sorted by hash. When given a hash, guess the position in the array of this hash, based on the numerical value of the hash. (I.e. if the numerical value is 25% of the maximum value the hash could be, then you would guess the index that is 25% of the total array size.) Your guess will be very close, because any good hash is evenly distributed. Go forwards or backwards until you find a match. This is freakishly close to $O(1)$, especially if you refine your guesses (binary-search-style) rather than just incrementing or decrementing.



    The main disadvantage of this method is that initially sorting the table is asymptotically slower than generating it, namely $O(2^n times log (2^n)) = O(n 2^n)$.



    Method 2:



    This is essentially a hashmap, except using a library hashmap is silly when you already have a hash! Just use the first $n$ digits of your hash. That means: create an array of size $2^n$. In each slot goes a list of all hashes whose first $n$ digits equal that array index. (And their corresponding passwords.)



    To look up a hash, take its first $n$ digits, and get the corresponding list out of the array. Traverse the list to find the hash. Like a hashmap, this takes $O(1)$ constant amortised time.






    share|improve this answer








    New contributor




    Artelius is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
    Check out our Code of Conduct.






    $endgroup$


















      2












      $begingroup$

      Brute Force



      Let's say your password was all digits (e.g. 607552)



      If your password has length 1, it will take up to 10 attempts to crack:



      0, 1, 2, 3 ..., 9


      If your password has length 2, it will take up to 100 attempts to crack



      00, 01, 02, ... 10, 11, 12, ... 20, 21, 22, ... 99


      If your password has length 3 it will take up to 1000 attempts to crack



      See the pattern? The number of attempts is $10^n$. It's the same idea if you are cracking a password with $n$ bits. (Since a bit can be 0 or 1, instead of 0-9, it's a power of 2 instead of 10).



      Lookup Table



      While a brute-force attack tries all possible passwords to see if they produce the right hash, a lookup table is a fast way to do the opposite: get a password based on a hash. It does this by hashing all the possible passwords ahead of time, and storing the results in a table.



      Method 1:



      Store the items in an array, sorted by hash. When given a hash, guess the position in the array of this hash, based on the numerical value of the hash. (I.e. if the numerical value is 25% of the maximum value the hash could be, then you would guess the index that is 25% of the total array size.) Your guess will be very close, because any good hash is evenly distributed. Go forwards or backwards until you find a match. This is freakishly close to $O(1)$, especially if you refine your guesses (binary-search-style) rather than just incrementing or decrementing.



      The main disadvantage of this method is that initially sorting the table is asymptotically slower than generating it, namely $O(2^n times log (2^n)) = O(n 2^n)$.



      Method 2:



      This is essentially a hashmap, except using a library hashmap is silly when you already have a hash! Just use the first $n$ digits of your hash. That means: create an array of size $2^n$. In each slot goes a list of all hashes whose first $n$ digits equal that array index. (And their corresponding passwords.)



      To look up a hash, take its first $n$ digits, and get the corresponding list out of the array. Traverse the list to find the hash. Like a hashmap, this takes $O(1)$ constant amortised time.






      share|improve this answer








      New contributor




      Artelius is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
      Check out our Code of Conduct.






      $endgroup$
















        2












        2








        2





        $begingroup$

        Brute Force



        Let's say your password was all digits (e.g. 607552)



        If your password has length 1, it will take up to 10 attempts to crack:



        0, 1, 2, 3 ..., 9


        If your password has length 2, it will take up to 100 attempts to crack



        00, 01, 02, ... 10, 11, 12, ... 20, 21, 22, ... 99


        If your password has length 3 it will take up to 1000 attempts to crack



        See the pattern? The number of attempts is $10^n$. It's the same idea if you are cracking a password with $n$ bits. (Since a bit can be 0 or 1, instead of 0-9, it's a power of 2 instead of 10).



        Lookup Table



        While a brute-force attack tries all possible passwords to see if they produce the right hash, a lookup table is a fast way to do the opposite: get a password based on a hash. It does this by hashing all the possible passwords ahead of time, and storing the results in a table.



        Method 1:



        Store the items in an array, sorted by hash. When given a hash, guess the position in the array of this hash, based on the numerical value of the hash. (I.e. if the numerical value is 25% of the maximum value the hash could be, then you would guess the index that is 25% of the total array size.) Your guess will be very close, because any good hash is evenly distributed. Go forwards or backwards until you find a match. This is freakishly close to $O(1)$, especially if you refine your guesses (binary-search-style) rather than just incrementing or decrementing.



        The main disadvantage of this method is that initially sorting the table is asymptotically slower than generating it, namely $O(2^n times log (2^n)) = O(n 2^n)$.



        Method 2:



        This is essentially a hashmap, except using a library hashmap is silly when you already have a hash! Just use the first $n$ digits of your hash. That means: create an array of size $2^n$. In each slot goes a list of all hashes whose first $n$ digits equal that array index. (And their corresponding passwords.)



        To look up a hash, take its first $n$ digits, and get the corresponding list out of the array. Traverse the list to find the hash. Like a hashmap, this takes $O(1)$ constant amortised time.






        share|improve this answer








        New contributor




        Artelius is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
        Check out our Code of Conduct.






        $endgroup$



        Brute Force



        Let's say your password was all digits (e.g. 607552)



        If your password has length 1, it will take up to 10 attempts to crack:



        0, 1, 2, 3 ..., 9


        If your password has length 2, it will take up to 100 attempts to crack



        00, 01, 02, ... 10, 11, 12, ... 20, 21, 22, ... 99


        If your password has length 3 it will take up to 1000 attempts to crack



        See the pattern? The number of attempts is $10^n$. It's the same idea if you are cracking a password with $n$ bits. (Since a bit can be 0 or 1, instead of 0-9, it's a power of 2 instead of 10).



        Lookup Table



        While a brute-force attack tries all possible passwords to see if they produce the right hash, a lookup table is a fast way to do the opposite: get a password based on a hash. It does this by hashing all the possible passwords ahead of time, and storing the results in a table.



        Method 1:



        Store the items in an array, sorted by hash. When given a hash, guess the position in the array of this hash, based on the numerical value of the hash. (I.e. if the numerical value is 25% of the maximum value the hash could be, then you would guess the index that is 25% of the total array size.) Your guess will be very close, because any good hash is evenly distributed. Go forwards or backwards until you find a match. This is freakishly close to $O(1)$, especially if you refine your guesses (binary-search-style) rather than just incrementing or decrementing.



        The main disadvantage of this method is that initially sorting the table is asymptotically slower than generating it, namely $O(2^n times log (2^n)) = O(n 2^n)$.



        Method 2:



        This is essentially a hashmap, except using a library hashmap is silly when you already have a hash! Just use the first $n$ digits of your hash. That means: create an array of size $2^n$. In each slot goes a list of all hashes whose first $n$ digits equal that array index. (And their corresponding passwords.)



        To look up a hash, take its first $n$ digits, and get the corresponding list out of the array. Traverse the list to find the hash. Like a hashmap, this takes $O(1)$ constant amortised time.







        share|improve this answer








        New contributor




        Artelius is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
        Check out our Code of Conduct.









        share|improve this answer



        share|improve this answer






        New contributor




        Artelius is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
        Check out our Code of Conduct.









        answered yesterday









        ArteliusArtelius

        1211




        1211




        New contributor




        Artelius is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
        Check out our Code of Conduct.





        New contributor





        Artelius is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
        Check out our Code of Conduct.






        Artelius is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
        Check out our Code of Conduct.






















            Fang is a new contributor. Be nice, and check out our Code of Conduct.










            draft saved

            draft discarded


















            Fang is a new contributor. Be nice, and check out our Code of Conduct.













            Fang is a new contributor. Be nice, and check out our Code of Conduct.












            Fang is a new contributor. Be nice, and check out our Code of Conduct.
















            Thanks for contributing an answer to Cryptography Stack Exchange!


            • Please be sure to answer the question. Provide details and share your research!

            But avoid



            • Asking for help, clarification, or responding to other answers.

            • Making statements based on opinion; back them up with references or personal experience.


            Use MathJax to format equations. MathJax reference.


            To learn more, see our tips on writing great answers.




            draft saved


            draft discarded














            StackExchange.ready(
            function () {
            StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fcrypto.stackexchange.com%2fquestions%2f66461%2fwhy-do-look-up-tables-speed-things-up-compared-to-brute-force%23new-answer', 'question_page');
            }
            );

            Post as a guest















            Required, but never shown





















































            Required, but never shown














            Required, but never shown












            Required, but never shown







            Required, but never shown

































            Required, but never shown














            Required, but never shown












            Required, but never shown







            Required, but never shown







            Popular posts from this blog

            "Incorrect syntax near the keyword 'ON'. (on update cascade, on delete cascade,)

            Alcedinidae

            RAC Tourist Trophy