Short name or logical name referencing?
Any piece of information modelled by COSEM can be accessed as an attribute or method of a certain COSEM interface object. To access an attribute or method, it has to be referenced.
In the metering equipment (server) the COSEM application layer provides two mechanisms to access the attributes. One is called short name (SN) referencing and the other is called logical name (LN) referencing. The data collection system (client) application layer always uses LN referencing.
When SN referencing is used, the attributes and methods of each interface object are mapped to DLMS named variables. This is done during the design of the meter. Each named variable is identified with a short name, which is a 16 bit unsigned integer. Attribute 1, the logical name of the object is mapped to a DLMS named variable identified by a base name. Except in the case of a few special objects, there are no general rules defined for assigning base names. All other attributes and methods of the object are then also mapped to DLMS named variables. The offsets between the base name and the short name identifying the other attributes and methods are defined in the definition of each interface class. The actual values of the short names thus depend on the number and kind of objects instantiated and the mapping strategy used. The base names allocated in the metering equipment can be retrieved by reading the object_list attribute of the SN Association object. When SN referencing is used, the DLMS named variables are accessed by the standard DLMS READ and WRITE services.
When LN referencing is used, attributes and methods are accessed via the logical name of the object, specifying the index(es) of the attribute(s) and/or the method(s). Logical names are defined by OBIS. When LN referencing is used, the attributes and methods are accessed by the xDLMS GET/SET and ACTION services.
As the data collection system (client) application layer always uses LN referencing, the LN service requests are mapped to SN service requests by the Short Name Mapper service element of the application layer.
SN referencing has been inherited from the original DLMS standard. LN referencing is the result of the evolution of the specification.
Which is the better? The answer is none of them. Simply put, SN referencing is more suitable for simple devices and LN referencing for complex devices.
The use of short names limits the possible number of objects within a logical device. Depending on the complexity, an interface object occupies a range of 8 (Data) to 136 (Clock) short names, starting with the base name. The whole range available is 8192. Therefore, the maximum number of objects depends on the type (complexity) of objects used and the strategy of allocating base names. On the other hand, APDU-s for SN referencing are shorter than APDU-s for LN referencing. With LN referencing, the limit is only set by the possible OBIS names (in principle 256**6). There are a few advanced features available only with LN referencing. However, this will make the software of the meter slightly more complex.
Finally, it is important to note, that the referencing style and the corresponding service set is negotiated at application association establishment and thus interoperability is fully preserved. There are known both complex and simple metering equipment using either SN or LN referencing, end even meters using both referencing styles.