Ed2k-hash

What's an ed2k hash?

A hash is a <insert non-laziness here>

The ed2k hash is based on the md4 algorithm, but rather than providing a single hash of the entire file, it breaks the file up into 9500kb chunks and produces a final hash based on the md4 sums of the chunks. While no longer considered secure from a cryptographic perspective, for the purpose of uniquely identifying files it's more than adequate. It is often listed as part of an ed2k link, which also includes a size in bytes, and a name.

Why does anidb require ed2k-hashes?

The main reason for this is that it avoids adding of double db entries. AniDB will not allow you to add a file with the same ed2k-hash as an existing one.

The file size and ed2k-hash of a file is used to identify it globally.

You are allowed to add files without ed2k-hashes to AniDB, however you should edit those files later and add the missing ed2k-hashes. Once you added a certain number of files without ed2k-hashes you may no longer add new files without ed2k-hashes until you edit your old files first.

Why use ed2k instead of another type of hash if you do not support downloads?

• The combination of ed2k-hash and file size makes ed2k effective for uniquely identifying files.
• Since ed2k-hashes can be passed back and forth within a URL with both the hash and file size in a widely recognized format it is a convenient method for adding or checking files.
• Other hashes are good for validating if a file is corrupt if you already know what file you are comparing it against, but cannot necessarily globally identify a file in the system like the ed2k.
• AniDB was designed around ed2k, although other hashes have been added to the file records for validation, the internal structure is based on ed2k. If the site goes through a complete redesign then maybe another hash will be made the primary hash, but at this point, this is not likely to change.

Which software can be used to generate them?

If you have the file(s) on your hard disk or on CD you can use all kinds of tools to generate the ed2k-hash and other additional info:

• AniDB_O'Matic by BennieB/PetriW can generate ed2k-hashes. It generates ed2k-hashes, md5, sha-1, crc32 in one go and also lists stuff like codec, resolution, bitrates, ...
• ed2k_hash another tool which is commandline based and also available for Linux.
• Filehash a little java program written by Malich. for further info on it read here
• Hashcalc can create md5, sha-1, crc32, ed2k-hashes and various other hashes in 1 go. though at least 1 case is known in which it created a wrong ed2k!

How is an ed2k hash calculated exactly?

A file is hashed in 9728000 byte chunks, using the md4 algorithm, and produces a 128 bit hash for each chunk. For files with only one chunk, the ed2k hash is the md4 of the file, however for hashes with 2 or more chunks the the hash of each chunk is appended to those before it, and an further md4 of the hashes themselves provides the ed2k hash of the file. Pseudo code is given below:

if filesize is less than or equal to 9728000:

return md4 of file

for chunk of size upto 9728000 in file:

append md4 of chunk to hashlist

if filesize is a multiple of 9728000:

append md4 of null to hashlist

return md4 of hashlist

Note that there are two different ways in practice that implimentations treat the 9728000 byte boundary, given as either the red code or the blue code above, black is common to both. In practice this difference only affects a tiny number of files, however is the one case where two 'valid' ed2k hashes might be produced from one file.

<list of which clients use which method>

• edonkey2000 v1.4.3
• mldonkey-2.5.30.17
• HashCalc Version 2.01
• edonkey-tool-hash-0.4.0
• fsum 2.51
• emule 0.46c
• AOM 0.5.5.239
• webaom v1.13
• ed2k (Stephane D'Alu) 1.4

<list of affected files, filesize>

File	Size (bytes)	Blue method	Red method
File of zeros	9728000	d7def262a127cd79096a108e7a9fc138	fc21d9af828f92a8df64beac3357425d
File of zeros	19456000	194ee9e4fa79b2ee9f8829284c466051	114b21c63a74b6ca922291a11177dd5c
http://anidb.info/f7047	145920000
http://anidb.info/f24359	136192000
http://anidb.info/f31383	136192000
http://anidb.info/f48530	175104000
http://anidb.info/f51131	107008000
http://anidb.info/f51330	175104000
http://anidb.info/f55744	97280000
http://anidb.info/f56411	165376000
http://anidb.info/f57766	145920000
http://anidb.info/f73921	184832000
http://anidb.info/f78552	194560000
http://anidb.info/f80216	165376000
http://anidb.info/f92884	68096000
http://anidb.info/f123554	243200000
http://anidb.info/f126410	184832000
http://anidb.info/f130233	184832000
http://anidb.info/f142402	155648000
http://anidb.info/f165143	165376000	0ced631bb9010d3ccd331689a2fb02de	6a092c056bc46e7a08d63408f918ba52
http://anidb.info/f166304	184832000	aab2ce19d5b786af20d6e4a15f63552f	aa9930ccd300a2feac30b0e49830c321
http://anidb.info/f174421	223744000