Estimating the Cost of XML Serialization of Java Objects

XML is a frequently applied data representation format to support the communication between software systems even if they are on different platforms. The process of transforming an in-memory object instance into bytes of an XML document is called XML serialization, the opposite direction is XML deserialization. Both processes introduce some cost that influences the overhead of the communication. This paper proposes a performance model that can estimate the cost of XML serialization and deserialization on Java platform based on a priori knowledge of the type and size of the XML messages. Type and size information of the XML messages is usually available in the form of an interface definition, such as a Web Services Description Language (WSDL) file. We show that in most cases the cost can be expressed as a linear function of the object size, but exceptions are discussed as well. The suggested model is validated with measurements.