Mapping Localisation Workflows on the Cloud

Today, traditional batch translation processes are no longer able to deal with the mass amount of available content. Automated management has become unavoidable. Recent advances in language technologies promise a variety of benefits and support to this automation. On the other hand, from the infrastructural perspective, Cloud computing and service-oriented modeling and architecture provide flexibility in integrating, developing, and running a variety of localisation and translation web services, which will make them readily available to translators and localisors throughout the world. However, the quality of the localisation and translation processes is an important matter and a challenging issue in this case. This depends on multiple parameters, including cost, deadlines, availability, quality of relevant linguistic assets, etc. In that aspect, we present a mechanism that is able to take into account predefined requirements and quality parameters to efficiently map an abstract workflow to the set of available service instances in the Cloud. We will also describe general issues facing automated localisation composition.