Authentication Mechanism

The security authentication of all APIs utilizes the Access Key and request signature mechanism. Access Key consists of Access Key ID and Secret Access Key, both of which are strings. For each HTTP request, the algorithm described below is utilized to generate one authentication string. The authentication string is submitted in the Authorization header. The server verifies the correctness of authentication string based on the generating algorithm. The format of the authentication string is bce-auth-v{version}/{accessKeyId}/{timestamp}/{expirationPeriodInSeconds}/{signedHeaders}/{signature}.

• The version stands for a positive integer.
• The timestamp stands for the UTC time when the signature is generated.
• The expirationPeriodInSeconds stands for the expiration period of signature. The signedHeaders stands for the header list involved in the signature algorithm. The headers are separated by a semicolon (;), e.g. host; x-bce-date. The list is arranged in lexicographic order. (This API signature only utilizes two headers of host and x-bce-date.)
• The signature is the 256-bit signed hexadecimal notation, and composed of 64 lower-case letters.

After Baidu AI Cloud receives a request of users, the system will use the same SK and authentication mechanism to generate an authentication string, and compare it with the authentication string contained in the user request. If the two authentication strings are the same, the system considers that the user has the designated operation permission, and executes the related operations; if the two authentication strings are different, the system will ignore the operation and return an error code. For details of authentication mechanism, please see Authentication Mechanism.

Communication Protocol

Support HTTP and HTTPS call methods. HTTPS calling is recommended to improve data security.

Description of Request Structure

The data exchange format is JSON, and all request/response body contents are encoded in UTF-8.

The request parameters include the 4 kinds below:

Parameter Type	Description
URI	It is used to indicate the operating entity, such as PUT /v1/user/{userName}
Query parameter	Request parameters carried in URL.
HEADER	Incoming through HTTP header field, such as: x-bce-date
RequestBody	Request data body organized in JSON format.

Description of Response Structure

Response values are divided into two formats:

Response content	Description
HTTP STATUS CODE	E.g. 200,400,403,404, etc.
ResponseBody	Response data body organized in JSON format.

Version Number

Parameter	Type	Parameter position	Description	Required or not
version	String	URI parameters	API version number	Required

Date and Time

There are various methods to express date and time. For the purpose of unification, unless there are conventions or corresponding specifications, all fields expressing date and time specified in the HTTP standard utilize GMT, and UTC time for other places expressing date and time. The users should follow ISO 8601, and the following restrictions are made:

• Fields expressing the date all utilize the YYYY-MM-DD format, e.g.2014-06-01 which means June 1, 2014.
• Fields expressing time all utilize the hh:mm:ss format, with the capital letter Z added at the end, which means UTC time. E.g. 23:00:10Z means UTC time: 23:00:10.
• When the date and time is combined, the capital letter T is added between the two items, e.g. 2014-06-01T23:00:10Z means UTC time: 23:00:10 on June 1, 2014.

Normalized String

Generally, one string can include any Unicode character. In programming, this kind of flexibility will bring about numerous troubles. Therefore, the concept of "normalized string" is introduced. One normalized string only includes the percent-encoding characters and URI (Uniform Resource Identifier) unreserved characters. RFC 3986 stipulates that "URI non-reserved characters" include the following characters: Letters (A-Z, a-z), numbers (0-9), hyphen (-), dot mark (.), underline (_), tilde (~). The method to change any string into a normalized string is:

• To convert strings into UTF-8 encoded byte stream.
• To keep all "URI non-reserved characters" unchanged.
• To make the Percent-encoding specified in RFC 3986 once for the rest bytes, namely, two hexadecimal strings representing the byte value are behind one %. All the letters are in upper case.

Example: Original string: This is an example for test, Corresponding specification string: this%20is%20an%20example%20for%20%E6%B5%8B%E8%AF%95.