Parallel visualization is assuming an increasing role in the deployment of Web and Grid based systems for scientific applications. The visualization process consists of a set of filters or components that are executed in a pipelined assembly that should be adaptively configured on the basis of user requirements, processing architecture and network characteristics. In this paper we focus our attention on the visualization of 3D irregular meshes produced by the interrogation of volumetric data using an isosurface extraction algorithm. We consider a simplified pipeline consisting of two components: isosurface extraction, and mesh simplification. We show that also in this simple case an in-deep analysis is necessary in order to optimize the whole pipeline. In fact different implementation and load balancing strategies are possible for each single component, but the whole pipeline optimization could be achieved combining non-optimal implementation of individual stages. Moreover the quality of the produced mesh should be considered in the selection of an adequate component implementation. The proposed analysis permits to point out trade-offs and algorithmic requirements that should be considered in the design of a complete visualization system for advanced Grid applications.