| Network Working Group | B. Greevenbosch |
| Internet-Draft | Huawei Technologies |
| Intended status: Informational | February 14, 2014 |
| Expires: August 18, 2014 |
CBOR data definition language: a notational convention to express CBOR data structures.
draft-greevenbosch-appsawg-cbor-cddl-01
This document proposes a notational convention to express CBOR data structures. Its main goal is to make it easy to express message structures for protocols that use CBOR.
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at http://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."
This Internet-Draft will expire on August 18, 2014.
Copyright (c) 2014 IETF Trust and the persons identified as the document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119].
In this document, a notational convention to express CBOR [RFC7049] data structures is defined.
The main goal for the convention is to provide a unified notation that can be used when defining protocols that use CBOR.
The CBOR notational convention has the following goals:
The following contains a list of used words in this document:
The basic syntax is as follows:
The following keywords for primitive datatypes are defined:
In addition, Section 4.6 defines datatypes associated with CBOR tags.
Arrays can be of fixed length or of variable length. Both fixed length and variable length arrays can be implemented as definite and indefinite length arrays.
A fixed length array is is indicated by '[' and ']' characters behind its type, where number in between specifies the number of elements.
A variable length array can be indicated with a "*" behind its type.
The following is an example of an array of 4 integers:
fourNumbers: int[4];
fibonacci : uint*;
The following is an example of a variable length array:
Structures are a logical grouping of CBOR fields.
A structure has a name, which can be used as a value type for other fields. The name is followed by a '{' character and the definitions of the variables inside of the structure. The structure is closed by a '}' character.
A structure MAY be encoded as an array, in which case its name is preceded by a '*' character. Otherwise there is no CBOR encoding for the grouping.
*Geography {
city : tstr;
gpsCoordinates : GpsCoordinates;
}
GpsCoordinates {
longitude : uint; # multiplied by 10^7
lattitude : uint; # multiplied by 10^7
}
The following is an example of a structure:
When encoding, the Geography structure is encoded using a CBOR array, whereas the GpsCoordinates do not have their own encompassing array.
map( x, y )
If an entity is a map (major type 5), its datatype has the form
If either x or y is unspecified (i.e. free to choose per entry), it is replaced by a '.'.
It is also possible to define a map with predefined keys as a type. In this case, type declaration is as follows:
x: map( y ) {
key1: type1;
key2: type2;
...
}
y is the datatype of the keys, and type1, type2, etc the datatype of the value associated with keys key1, key2 etc.
The name of an optional map element is preceded by a '?' character.
The example below the defines a map with display name (as a string), optionally the name components first name and family name (see Section 4.7 for more on optional variables), and age information (as an unsigned int).
PersonalData: map( tstr ) {
"displayName": tstr;
?"nameComponents": NameComponents;
"age": uint;
}
NameComponents: map( tstr, tstr ) {
"firstName": tstr;
"familyName" : tstr;
}
It is up to the application how to handle unknown tags, however, it is RECOMMENDED to ignore them.
A variable can have an associated CBOR tag (major type 6). This is indicated by the tag encapsulated between the square brackets '[' and ']', just before the variable's datatype definition.
N: [2]bstr;
For example, the following defines a positive bignum N:
[RFC7049] defines several tags. These tags can be also written using the datatypes from Table 1. For table rows with an empty "possible tag notation" entry, we refer to Table 3 in [RFC7049] and associated references for the possible encodings.
For example, the following is another way to define the bignum:
N: bignum;
| datatype | possible tag notation | description |
|---|---|---|
| b64 | [34]tstr | Base 64 (tag 34) |
| b64url | [33]tstr | Base 64 URL (tag 33) |
| bigfloat | bigfloat (tag 5) | |
| bignum | [2]bstr or [3]bstr | positive (tag 2) or negative (tag 3) bignum |
| cbor | [24]bstr | Encoded CBOR data item (tag 24) |
| decfrac | decimal fraction (tag 4) | |
| eb16 | Expected conversion to base16 encoding (tag 23) | |
| eb64 | Expected conversion to base64 encoding (tag 22) | |
| eb64url | Expected conversion to base64 url encoding (tag 21) | |
| epochdt | epoch date/time (tag 1) | |
| mime | [36]tstr | Mime message (tag 36) |
| nbignum | [3]bstr | negative bignum (tag 3) |
| regex | [35]tstr | regular expression (tag 35) |
| standarddt | [0]tstr | standard date/time string (tag 0) |
| ubignum | [2]bstr | positive bignum (tag 2) |
| uri | [32]tstr | URI (tag 32) |
There may be variables or structures whose inclusion is optional. In this case, the name of the variable is preceded by a '?' character
*MainStruct {
whichForm : bool;
?data1 : Form1; # when whichForm == true
?data2 : Form2; # when whichForm == false
}
Form1 {
anInteger : int;
aTextString : tstr;
}
Form2 {
aFloat : float(16);
aBinaryString : bstr;
}
For example, the following defines a CBOR structure that is dependent on a boolean value.
Protocol designers should exhibit utmost care when defining CBOR structures with optional variables, especially when some of these variables have the same datatype.
*DataStruct {
?OptionalVariable : uint;
MandatoryVariable : uint;
?AnotherOptionalVariable : uint;
}
For example, the following CBOR data structure is ambiguous:
Since optional variables are often detected from their datatype, it is RECOMMENDED to not have a following of multiple variables of the same datatype, when some of these variables are optional.
This section contains various examples of structures defined using the CBOR notational convention.
A multiplayer computer game uses CBOR to exchange moves between the players. To ensure a good gaming experience, the move information needs to be exchanged quickly and frequently. Therefore, the game uses CBOR to send its information in a compact format. Figure 1 shows definition of the CBOR information exchange format.
*UpdateMsg {
move_no : uint; # increases for each move
player_info : PlayerInfo; # general information
moves : Moves*; # moves in this message
}
PlayerInfo {
alias : tstr;
player_id : uint;
experience : uint; # beginner: 0; expert: 3
gold : uint;
supplies : Supplies;
avg_strength : float(16);
}
Supplies : map( uint ) {
0 : uint; # wood
1 : uint; # iron
2 : uint; # grain
}
*Moves {
unit_id : uint;
unit_strength : uint; # between 0 and 100
source_pos : uint[2]; # (x,y)
target_pos : uint[2]; # (x,y)
}
Figure 1: CBOR definition of an information exchange format for a computer game
Notice that the supplies have been encoded as a map with integer keys. In this example, using string keys would also have been suitable. However, the example illustrates the possibility to use other datatypes for keys, leading to more efficient encoding.
Player "Johnny" does two moves. The game server has assigned Johnny the ID 0x7a3b871f. Johnny is an amateur player, so has experience 1. He currently has 1200 gold, 13 units of wood, 70 units of iron and 29 units of grain. He has several units, with a total average strength of 30.25.
The units Johnny plays in move 250 are the unit with ID 19, strength 20 from (5,7) to (6,9), and the unit with ID 87, strength 40 from (7,10) to (6,10).
9F
18 FA # move 250
66 4A 6F 68 6E 6E 79 # "Johnny"
1A 7A 3B 87 1F # player_id
01 # experience
19 04 B0 # 1200 gold as uint
A3 # begin map "supplies" with 3 elements
00 # wood:
0C # 13 as uint
01 # iron:
18 86 # 70 as uint
02 # grain:
18 1D # 29 as uint
F9 4F 90 # average strength 30.25 half-precision float
9F # indefinite length "moves" array
84 # 4-element array Moves
13 # unit id 19 as uint
14 # strength 20 as uint
82 # 2-element array source_pos
05 # source_pos.x=5
07 # source_pos.y=7
82 # 2-element array target_pos
06 # target_pos.x=6
09 # target_pos.y=9
84 # 4-element array Moves
18 57 # unit id 87
18 28 # strength 40
82 # 2-element array source_pos
07 # source_pos.x=7
0a # source_pos.y=10
82 # 2-element array target_pos
06 # target_pos.x=6
0a # target_pos.y=10
FF # end of "moves" array
FF
Figure 2: CBOR instance for game example
This information is coded in CBOR as depicted in Figure 2.
fruitlist : Fruit*;
*Fruit {
name : FruitName;
colour : uint;
avg_weight : float( 16 );
price : uint;
international_names : map( Lang, tstr );
rfu : bstr; # reserved for future use
}
FruitName : map( tstr ) {
"CN" : tstr; # Chinese
"NL" : tstr; # Dutch
"EN" : tstr; # English
"FR" : tstr; # French
"DE" : tstr; # German
}
Figure 3: Example CBOR structure
Figure 3 contains an example for a CBOR structure that contains information about fruit.
The colour integer can have the values from Table 2.
| Colour | Value |
|---|---|
| black | 0 |
| red | 1 |
| green | 2 |
| yellow | 3 |
| blue | 4 |
| magenta | 5 |
| cyan | 6 |
| white | 7 |
| orange | 8 |
| pink | 9 |
| purple | 10 |
| brown | 11 |
| grey | 12 |
For example, apples can be red, yellow or green. They have an average weight of 0.195kg and a price of 30 cents. Chinese for "apple" in UTF-8 is [ E8 8B B9 E6 9E 9C ], the Dutch word is "appel" and the French word "pomme".
For simplicity, let's assume that the colour of oranges can only be orange. They have an average weight of 0.230kg and a price of 50 cents. Chinese for "orange" in UTF-8 is [ E6 A9 99 E5 AD 90 ], the Dutch word is "sinaasappel" and the German word "Orange".
9F # indefinite length "fruitlist" array
86 # First "Fruit" instance, 6 elements
65 # text string "name" length 5
61 70 70 6C 65 # "apple"
83 # array for "Colour", 3 elements
01 # "red" as uint
02 # "green" as uint
03 # "yellow" as uint
F9 # Floating point half precision
32 3D # "avg_weight" 0.195
18 1E # "price" 30 as uint
A3 # map "international_names", 3 pairs
62 43 4E # text string length 2, "CN"
66 E8 8B B9 E6 9E 9C # Chinese word for apple
62 4E 4C # "NL"
65 61 70 70 65 6C # "appel"
62 46 52 # "FR"
65 70 6F 6D 6D 65 # "pomme"
40 # byte string "rfu", 0 bytes length
86 # Second "Fruit" instance
66 # text string "name" length 6
6F 72 61 6E 67 65 # "orange"
81 # array for "Colour", 3 elements
08 # "orange" as uint
F9 # Floating point half precision
33 5C # "avg_weight" 0.230
18 32 # "price" 50 as uint
A3 # map "international_names", 3 pairs
62 43 4E # text string length 2, "CN"
66 E6 A9 99 E5 AD 90 # Chinese word for orange
62 4E 4C # "NL"
6B 73 69 6E 61 61 73 61 70 70 65 6C # "sinaasappel"
62 44 45 # "DE"
66 4F 72 61 6E 67 65 # "Orange"
40 # byte string "rfu", 0 bytes length
FF # end of "fruitlist" array
Figure 4: Example CBOR instance
This information would be encoded as depicted in Figure 4.
Notice that if the "Fruit" structure did not have the preceding "*", the two "Fruit" instance arrays would have been omitted. In addition, the "fruitlist" array would have had 12 elements instead of 2. (Although for "fruitlist" the indefinite length approach was chosen, such that the number of elements is not explicitely signalled.)
The CBOR notational convention can be used to efficiently define the layout of CBOR data.
In addition, it has been specified such that a machine can verify whether or not CBOR data is compliant to its definition. The thoroughness of this compliance verification depends on the application.
For example, an application may decide not to verify the data structure at all, and use the CBOR content rules solely as a means to indicate the structure of the data to the programmer.
On the other end, the application may also implement a verification method that goes as far as verifying that all mandatory map keys are available.
The matter in how far the data description must be enforced by an application is left to the implementers and specifiers of that application.
At least the following issues need further consideration:
Changes from version 00 to version 01
This document presents a content rules language for expressing CBOR data structures. As such, it does not bring any security issues on itself, although specification of protocols that use CBOR naturally need security analysis when defined.
This document does not require any IANA registrations.
For this draft, there has been inspiration from the C and Pascal languages, MPEG's conventions for describing structures in the ISO base media file format, and Andrew Lee Newton's "JSON Content Rules" draft.
Useful feedback came from Carsten Bormann and Joe Hildebrand.
| [RFC2119] | Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. |
| [RFC7049] | Bormann, C. and P. Hoffman, "Concise Binary Object Representation (CBOR)", RFC 7049, October 2013. |