DER file format: Encoding rules

DER, the Distinguished Encoding Rules, is the strict form of BER, the Basic Encoding Rules, both defined in ITU-T X.690 for data described in ASN.1 (X.680). Every value is encoded as identifier, length and contents octets (“TLV”); a constructed value’s contents are more TLVs. DER allows exactly one encoding for each value, which is why signatures are computed over it.

This page follows Rec. ITU-T X.690 (02/2021), the free ITU-T edition of ISO/IEC 8825-1, clauses 8 (BER), 10 (DER) and 11 (restrictions shared by CER and DER), and takes the universal tag numbers from Rec. ITU-T X.680 (02/2021), Table 1 and Table 8.

Identifier octets

Bits	Name	Meaning
8–7	Class	`00` universal (types defined by ASN.1 itself), `01` application, `10` context-specific (the `[0]`, `[1]` … tags of a structure), `11` private.
6	Constructed	1 if the contents are further TLVs (SEQUENCE, SET, explicit tags); 0 for a primitive value.
5–1	Tag number	0 to 30. The value 31 (`11111`) means the tag number follows in further octets, 7 bits per octet, most significant first, with bit 8 set on every octet but the last; bits 7–1 of the first further octet must not all be zero (X.690 8.1.2.4).

So 30 is a universal constructed tag 16 (SEQUENCE), A0 a context-specific constructed tag 0 ([0]), and 80 a context-specific primitive tag 0.

Universal tags

Number	Byte (primitive / constructed)	Type
0	`00`	Reserved for the encoding rules; BER uses `00 00` as end-of-contents after an indefinite length (X.690 8.1.5)
1	`01`	BOOLEAN
2	`02`	INTEGER
3	`03`	BIT STRING
4	`04`	OCTET STRING
5	`05`	NULL
6	`06`	OBJECT IDENTIFIER
7	`07`	ObjectDescriptor
8	`28`	EXTERNAL, INSTANCE OF
9	`09`	REAL
10	`0A`	ENUMERATED
11	`2B`	EMBEDDED PDV
12	`0C`	UTF8String
13	`0D`	RELATIVE-OID
14	`0E`	TIME
15	–	Reserved for future editions
16	`30`	SEQUENCE, SEQUENCE OF
17	`31`	SET, SET OF
18	`12`	NumericString
19	`13`	PrintableString
20	`14`	TeletexString (T61String)
21	`15`	VideotexString
22	`16`	IA5String (ASCII)
23	`17`	UTCTime
24	`18`	GeneralizedTime
25	`19`	GraphicString
26	`1A`	VisibleString (ISO646String)
27	`1B`	GeneralString
28	`1C`	UniversalString (4-octet ISO/IEC 10646 form)
29	`1D`	CHARACTER STRING (X.680 44.2)
30	`1E`	BMPString (2-octet ISO/IEC 10646 BMP form)
31	`1F 1F`	DATE
32	`1F 20`	TIME-OF-DAY
33	`1F 21`	DATE-TIME
34	`1F 22`	DURATION
35	`1F 23`	OID-IRI
36	`1F 24`	RELATIVE-OID-IRI
37 on	–	Reserved for addenda

The bytes are worked out from X.690 8.1.2: SEQUENCE, SET, EXTERNAL and EMBEDDED PDV are always constructed (8.9–8.12, 8.17, 8.18); the others are shown primitive, which DER requires for strings (10.2).

Length octets

First octet	Form	Meaning
`00`–`7F`	Short	The length itself, 0 to 127.
`81`–`FE`	Long	The low 7 bits say how many octets of length follow, big-endian: `81 C8` is 200, `82 03 5A` is 858.
`80`	Indefinite	BER only, and only for constructed encodings: the contents run until an end-of-contents `00 00`. Not allowed in DER.
`FF`	–	Shall not be used (kept for future extension, 8.1.3.5).

Contents of common types

Type	Contents
BOOLEAN	One octet: `00` false; BER allows any non-zero octet for true (8.2), DER and CER require `FF` (11.1).
INTEGER, ENUMERATED	Two’s complement, big-endian (8.3, 8.4). If there is more than one octet, the first octet and bit 8 of the second must not be all ones or all zeros, so the fewest octets are used: 128 is `02 02 00 80`.
BIT STRING	An octet giving the number of unused bits in the last octet (0–7), then the bits, the leading bit in bit 8 of the first octet (8.6.2). An empty bit string is the single octet `00`.
OCTET STRING	The octets.
NULL	Nothing: `05 00` (8.8).
OBJECT IDENTIFIER	The first two arcs as 40 × first + second, then each further arc, each in base 128 in the fewest octets, bit 8 set on all but the last octet of each (8.19). 1.2.840.113549 is `06 06 2A 86 48 86 F7 0D`; X.690’s own example, 2.999.3, is `06 03 88 37 03`.
UTCTime	ASCII; in DER it ends with `Z` and the seconds are always present (11.8): `YYMMDDHHMMSSZ`.
GeneralizedTime	ASCII; in DER it ends with `Z`, the seconds are always present, fractions of a second drop trailing zeros (and the point, if nothing is left), and the decimal point is `.` (11.7). Midnight is `YYYYMMDD000000Z` of the following day.
Strings	UTF8String in UTF-8, UniversalString in the 4-octet and BMPString in the 2-octet ISO/IEC 10646 forms (8.23.7–8.23.10); the other string types by the registered character sets X.680 Table 8 lists for them.

DER restrictions

Lengths are always definite and in the fewest octets (no 81 05 for 5) (10.1).
Bit strings, octet strings and restricted character strings are always primitive, never split into constructed pieces (10.2).
The components of a SET appear in X.680’s canonical tag order: universal, then application, then context-specific, then private, each by ascending tag number (10.3; X.680 8.6).
A SET or SEQUENCE component equal to its DEFAULT is not encoded (11.5).
The elements of a SET OF are in ascending order of their encodings, compared as octet strings with the shorter padded with zero octets for the comparison (11.6).
Unused bits of a BIT STRING are zero, and where X.680 22.7 applies (named-bit lists) trailing zero bits are removed (11.2).

Example

00000000  30 82 03 5A 30 82 02 42 A0 03 02 01 02 02 14 …  0..Z0..B.......

The start of a DER certificate.

Reading it: a SEQUENCE of 858 bytes (the certificate); inside it a SEQUENCE of 578 bytes (the to-be-signed part); a context tag [0] of 3 bytes holding the INTEGER 2, which is version 3; then an INTEGER of 20 bytes, the serial number. See certificates for the rest.