The huge and dynamic amount of bioinformatic resources (e.g., data and tools) available nowadays in Internet represents a big challenge for biologists –for what concerns their management and visualization– and for bioinformaticians –for what concerns the possibility of rapidly creating and executing in-silico experiments involving resources and activities spread over the WWW hyperspace. Any framework aiming at integrating such resources as in a physical laboratory has imperatively to tackle –and possibly to handle in a transparent and uniform way– aspects concerning physical distribution, semantic heterogeneity, co-existence of different computational paradigms and, as a consequence, of different invocation interfaces (i.e., OGSA for Grid nodes, SOAP for Web Services, Java RMI for Java objects, etc.). The framework UBioLab has been just designed and developed as a prototype following the above objective. Several architectural features –as those ones of being fully Web-based and of combining domain ontologies, Semantic Web and workflow techniques– give evidence of an effort in such a direction. The integration of a semantic knowledge management system for distributed (bioinformatic) resources, a semantic-driven graphic environment for defining and monitoring ubiquitous workflows and an intelligent agent-based technology for their distributed execution allows UBioLab to be a semantic guide for bioinformaticians and biologists providing (i) a flexible environment for visualizing, organizing and inferring any (semantics and computational) "type" of domain knowledge (e.g., resources and activities, expressed in a declarative form), (ii) a powerful engine for defining and storing semantic-driven ubiquitous in-silico experiments on the domain hyperspace, as well as (iii) a transparent, automatic and distributed environment for correct experiment executions.