Let me warn you ahead of time this is more of an algorithm question than a pure Perl question.

I recently got handed the job of maintaining a web application that registers visitors to a site for sweepstakes and free samples. It uses Apache::Session to store the visitor's information. This is the routine from Session.pm that creates a session id to use as a key.

sub generate_id {
  return substr(MD5->hexhash(time(). {}. rand(). $$. blah'), 0, 16);
}

[download]

To be a good key for this use on the web, I'm primarily worried about two things: is it unique, and how likely is it that given one key a person can guess another valid one.

The string normally returned from hexhash is 32 chars long. This routine is tossing 16 of those characters right out the window (64 of 128 bits).

If I convert to a base 64 representation for the 128 bit quantity (22 chars) I only lose 36 bits by truncation. If I do what seems the logical thing and override the truncation being done by the module, our current database column definitions are too short and several of the reporting tools have to be examined for bugs (best case) or partially rewritten (worst case). In order to get management sign-off for that, I need to present a pretty strong case for change.

My questions are:

  1. How much more likely is it that two data strings like those being passed into hexhash() will collide if you only use the first 64 bits rather than all 128? I can't find anything about predictability of MD5 substrings. Maybe I'm worried about nothing?
  2. Some of the applications we build require the users to type in a URL (containing a session id) that was sent to them on a postcard. Is there a reliable way in Perl to generate a shorter identifier that meets both the uniqueness and the "very difficult to guess" tests?

If I'm doing the math correctly, 28 bits gives me almost 270 million buckets (a very rough 1 to 1 mapping to the 285 million in the US). I could code that in hex as 7 characters or base64 as 5 chars (although then the user has to get the case of the letters correct when they type them in).

Maybe a better second question would be:
"Does anyone know where I can find a digest function that hashes uniformly into a 28-32 bit address space?"

BTW, if this offends you because it's not directly a Perl question, I'd love to take it elsewhere but I don't have access to newsgroups/IRC from work and don't know of a more appropriate place to ask. If you do, please suggest it.

In reply to How safe is truncating an MD5 digest string? by lemmett

Title:
Use:  <p> text here (a paragraph) </p>
and:  <code> code here </code>
to format your post, it's "PerlMonks-approved HTML":


  • Posts are HTML formatted. Put <p> </p> tags around your paragraphs. Put <code> </code> tags around your code and data!
  • Titles consisting of a single word are discouraged, and in most cases are disallowed outright.
  • Read Where should I post X? if you're not absolutely sure you're posting in the right place.
  • Please read these before you post! —
  • Posts may use any of the Perl Monks Approved HTML tags:
    a, abbr, b, big, blockquote, br, caption, center, col, colgroup, dd, del, details, div, dl, dt, em, font, h1, h2, h3, h4, h5, h6, hr, i, ins, li, ol, p, pre, readmore, small, span, spoiler, strike, strong, sub, summary, sup, table, tbody, td, tfoot, th, thead, tr, tt, u, ul, wbr
  • You may need to use entities for some characters, as follows. (Exception: Within code tags, you can put the characters literally.)
            For:     Use:
    & &amp;
    < &lt;
    > &gt;
    [ &#91;
    ] &#93;
  • Link using PerlMonks shortcuts! What shortcuts can I use for linking?
  • See Writeup Formatting Tips and other pages linked from there for more info.