Account Backend Database Installation

Figure 21. Authentication Account Provider Figure 9 allows you to select the authentication account provider. Database All accounts will be stored in the registry database. This is the recommended backend. LDAP Registry accounts integrated with LDAP server. External Registry accounts integrated with other external storage. To integrate Oracle Service Registry, with an external backend, you must implement the interface com.systinet.uddi.account.ExternalBackendApi and add it to the registry installation. For more information about LDAP, Oracle XML and External account backends, please see Section 6, External Accounts Integration

4.9. Multilingual Data

This section describes how Oracle Service Registry supports the storage of UDDI structures in the multilingual data format. There are two types of text fields in UDDI structures: Unicode fields and ASCII fields. Page 87

4.9. Multilingual Data

Unicode fields are intended for human readable information, the field length is measured in number of characters as follows: Max Length in chars Field Name 255 name of businessEntity and businessService 255 keyName 255 keyValue 255 useType 255 description 80 addressLine 255 personName ASCII fields are intended for machine processing, such as URIs. The length is measured in bytes. ASCII fields can typically hold multilingual data. Its length is limited by the number of bytes of its serialized form in UTF-8 encoding. For example, the name of a tModel can carry 85 Japanese characters, because Japanese characters are encoded into three bytes each under UTF-8 encoding 2553=85. Max Length in bytes Field Name 255 name of tModel 4096 overviewURL 4096 discoveryURL 10 sortCode 255 email 50 phone 4096 accessPoint 8192 instanceParms

4.9.1. Oracle

Oracle database supports Unicode characters in both types Unicode and ASCII of fields.

4.9.2. MSSQL

MSSQL supports Unicode characters only in Unicode fields. Unicode characters are stored successfully to ASCII fields only if they match with the server collation, otherwise are converted to question marks ?. For example, Japanese characters are stored correctly if the Japanese_Unicode_Cl_AS collation is default to the server. If the English collation is set up, Japanese characters are converted to ? characters.

4.9.3. DB2

The DB2 database supports Unicode characters in both types of fields. Maximal length of a field is measured in bytes in the default database schema despite it being a Unicode field. You can use any Unicode characters, but allowed string length is not guarantied. For example, the name of a tModel can carry 85 Japanese characters, because Japanese characters are encoded into three bytes each under UTF-8 encoding 2553=85. Note that longer strings produce a database exception. The restriction is made because the cumulative length of indexed columns is limited to 800 bytes. The default schema prefers performance to multiple language support. Page 88

4.9.3. DB2