This paper presents an extension of the single-user scientific visualization system ZoomIn to a collaborative system. We discuss the principal challenges of remote collaboration in the case of scientific visualization and present the solutions realized for ZoomIn. The main issues identified include the support of slow/fast network connections, portability, ease-of-use, privacy, configuration of the working mode and permissions. We describe why the main concepts of the original ZoomIn are well suited for a collaborative extension and how collaboration is introduced. In particular, the volume of data for scientific visualization may be very large. Therefore, the transfer of data has to be optimized. We present three different methods of data transfer for collaborative visualization and compare their efficiency with respect to particular application scenarios.