Torrent File Structure Resolution

The data structure in the torrent file is divided into the following sections:

Announce:tracker's primary server

Announce-list:tracker Server List

Comment: Comments on seed files

Comment.utf-8: Utf-8 encoding for seed file annotations

Creation Date: The time that the seed file was established is the number of seconds from January 1, 1970 00:00:00 to the present.

Encoding: The default encoding for seed files, such as Gb2312,big5,utf-8

Info: All information about downloaded files is in this field, it includes multiple child fields, and depending on whether you're downloading a single file or multiple files, the child field items will be different.

When the seed contains more than one file, the Info field includes the following child fields:

Files: Represents the file name, size, which contains the following three child fields:

Lenghth: Size of file, in byte

Path: The name of the file that cannot be changed at download time

Path.utf-8: UTF-8 code for filename, ibid.

The above three fields have a set of values for each file.

Name: The recommended folder name, which can be changed at download time.

Name.utf-8: Recommended Utf-8 encoding for the folder name, ditto.

Piece Length: The size of each file block, calculated in byte

Pieces: File feature information, the field is larger, is actually the seed contains all the file segments of the SHA1 checksum value of the connection, will all the files according to the piece length of the byte size into blocks, each block to calculate a SHA1 value, These values are then concatenated to form the pieces field, which is always 20 times the size of an integer, because the SHA1 checksum value is 20Byte. This field is the largest portion of the torrent file, which can be seen if large files are small chunks, resulting in large torrent file size.

Publisher: Name of File publisher

Publisher.utf-8: Utf-8 encoding of the name of the file publisher

Publisher-url: Web site for publishers of documents

Publisher-url.utf-8: Utf-8 encoding of the file Publisher URL.

Also, when you publish a single file, the Files field is not, and the


Lenghth:
Name
Name.utf-8:

These three fields are responsible for describing the attributes of a single file: size, name, utf-8 encoding of the name. Other items are the same as multiple files.

The above items are all of the info field.

When it comes to info, I have to say Info_hash, this value is the hash value of the info field, 20 Byte, also uses SHA1 as the hash function. Since the Info field is made up of published file information, Info_hash is used to identify different seed files in the BT protocol. Basically, the info_hash of each seed file is different (at least no one has found the collision of Sha), so both the BT server and the client identify the different seed files with this value.

The scope of the calculation is to start with the Info field (not including the four bytes of "info") until the nodes field (not including the 5 bytes "nodes" and the two bytes of "5:" in front of the nodes representing the nodes field length). In addition, the Info_hash value is computed instantaneously and is not included in the torrent file.

Nodes: The last field is the Nodes field, which contains a list of IP and corresponding ports that are used to connect to the DHT initial node.

In summary, the structure of the multi-file torrent is:

Multi-file Torrent
├─announce
├─announce-list
├─comment
├─comment.utf-8
├─creation Date
├─encoding
├─info
│├─files
││├─length
││├─path
││└─path.utf-8
│├─name
│├─name.utf-8
│├─piece length
│├─pieces
│├─publisher
│├─publisher-url
│├─publisher-url.utf-8
│└─publisher.utf-8
└─nodes

The tree of the torrent structure of a single file is:

Single-file Torrent
├─announce
├─announce-list
├─comment
├─comment.utf-8
├─creation Date
├─encoding
├─info
│├─length
│├─name
│├─name.utf-8
│├─piece length
│├─pieces
│├─publisher
│├─publisher-url
│├─publisher-url.utf-8
│└─publisher.utf-8
└─nodes

The following is a practical example to illustrate the actual structure of the torrent file:

First, we need to understand the logical representation of data in the torrent file.

String: String by number of characters: the form of a string, such as 5:files means "files"
Integer: Torrent in 10 to denote an integer, consisting of the form of the I-digit e, as i50e means 50
List: The list starts with L, E ends with multiple strings in the middle, such as L5:files4:infoe ["Files", "info"]
Dictionary: Start with D, end of E, the middle can be any structure, such as d5:filesl4:info3:eggee means {files, ["Info," egg "]}


The sample file is a torrent containing three PDF files, within the directory mysampletorrent, the directory structure is:

Mysampletorrent
├─protocol.pdf
├─draft_dht_protocol.pdf
└─fast_extensions.pdf

The torrent file named Mysampletorrent.torrent is made by BitComet, the size is 976 Byte, the block size is 8M, and the following is the display of winhex opening the torrent:

The information that can be obtained from the above diagram corresponds to the structure tree:

├─announce->http://www.sample.org/announce
├─announce-list
├─comment->this is a sample comment of thsi torrent file.
├─comment.utf-8->this is a sample comment of thsi torrent file.
├─creation Date->1182679174
├─encoding->shift-jis
├─info
│├─files
││├─length->59126
││├─path->draft_dht_protocol.pdf
││└─path.utf-8->draft_dht_protocol.pdf
│├─name->mysampletorrent
│├─name.utf-8->mysampletorrent
│├─piece length->8388608
│├─pieces->751ce8fd85d69d5edeb31f6b15f3c729587bddd6