CdcFileFormat

• Base type Enumeration

The file formats supported are: cdcBulkASCII ASCII format similar to the 'bulkASCII' format of the CISCO-BULK-FILE-MIB. cdcBulkBinary Binary format similar to the 'bulkBinary' format of the CISCO-BULK-FILE-MIB. cdcSchemaASCII Again an ASCII format, but contains parser friendly hints for parsing data values. Format descriptions : --------------------- The below description applies to cdcBulkASCII & cdcBulkBinary and is extracted from CISCO-BULK-FILE-MIB. Some differences are highlighted below: 1) Data for a single data group can be collected more than once into the same VFile (due to periodic polling). Each such instance of a data group, can be treated much like different 'table' types in the CISCO-BULK-FILE-MIB. 2) Every object & table tag contains an additional sysUpTime field. Similarly each row tag contains the value of the sysUpTime when the data for that row was collected. The rest of the description is taken from the CISCO-BULK-FILE-MIB. The file contains two types of fields: tags and data. Tags identify portions of the file and are discussed in detail below. All other information is in data fields. Note: For efficiency and compactness data fields are not tagged with a type. The interpreter of the data must thus know or have access to appropriate MIB syntax descriptions to understand the file. All data fields are positional relative to a tag and every data field has a length prefix. All initial length prefixes are one byte. For any data type the distinguished length value 255 indicates that the data content is null, that is, no data content value was available and there are no additional bytes in the data field. INTEGER data fields include all data that maps to ASN.1 INTEGER, regardless of length and whether signed or unsigned. They have a length prefix value of zero to eight, followed by that many bytes of data, high-order byte first. High order bytes that are all zero are omitted, thus a length of zero indicates a value of zero. For signed numbers, leading bytes of all ones (hex FF) are omitted if the next remaining byte has the high bit on. This implies that the file parser must know the difference between signed and unsigned integers. OCTET STRING values have a length prefix value of zero to two for a subsequent unsigned byte count for the number of bytes in the OCTET STRING itself, which immediately follows the byte count. The byte count can thus range from zero to 65,535. OBJECT IDENTIFIER values have a length of zero to 128, for the number of sub-identifiers. Each subsequent sub-identifier is encoded as an unsigned INTEGER of 0-4 bytes. The bulk binary file layout directly reflects the contents of the cbfDefineFileObjectTable. It has tagged sections corresponding to cbfDefineObjectClass with a few additional tags for utility purposes. A tag is one byte with one of the following values: -2 row -1 prefix 0 reserved 2 table The prefix tag changes the default OID prefix that is assumed to precede all OIDs that are not MIB object data values. The prefix tag may appear anywhere another tag could appear. A prefix tag is followed by one OID data field. The default prefix is 1.3.6.1. A file need not set the prefix to the default value. Note that when changing the prefix, the default portion must be included at the beginning of the new prefix. Typically the prefix will change for each table or group of scalar objects. A table tag is followed by one INTEGER data field whose value is the number of columns in the table (including the sysUpTime column), as implemented by the agent. This is followed by an OID field containing the sysUpTime OID. This is followed by one OID data field for each column. This is the OID for the column minus the prefix and the instance (typically one subidentifier). The OIDs are then followed by one row for each row in the table. A row starts with a row tag, one OID data field containing only the instance portion of the OIDs for the objects in that row and one data field indicating the sysUpTime when the row was sampled. Following this is one data field of appropriate type for each column. The bulk ASCII form mechanically translates bulk binary into human-readable text. The indicator for a null value is '~'. An INTEGER becomes the integer value with a preceding '-' for negative values and no leading zeros. An OCTET STRING becomes the byte values in hexadecimal, lower case, two characters per byte (that is, with leading zeros), no delimiters between bytes. An OBJECT IDENTIFIER becomes the usual dotted decimal form. A tag becomes the tag's name, spelled out fully in lower case, followed by one blank and the data field(s) for the tag, separated by spaces and ending with a carriage return/line feed. All tags are at the beginning of a 'line' that is terminated with a carriage return/line feed that immediately precedes the next tag or the end of file.