**2. On-line visualization as a sequence application in grid environment**

Grid-based applications that generate visualization outputs on-the-fly require a unified application framework. With regards to the Grid computational environment, it is advantageous to divide the performation of Grid applications into three separate phases.


### **2.1 Fetching the data from simulations**

The first, computational phase is submmision the application to the HP, Grid or Cloud environment. The whole application consists of a large number of consecutive series of jobs. The output from a given series becomes the input for the next series.

The user must prepare the data on the Storage Element and provide the configuration file before launching the job controller. Data can be added even while the application is running (providing the relevant rules are kept). During its run the user can observe the contents of directories with input and output data on the Storage Element. The simulation outputs are treated as input data by the visualization process. The job controller starts their processing by launching the visualization tool. This is the subject of visualization phase.

### **2.2 Transformation the data into a displayable form**

The aim of this phase is the transformation of the data into a displayable form and the creation of virtual scenes according to the requirements of the client. This phase is carried out by a visualization tool, which is together with its controlling agents, is the crux of this article. Apparently, the best approach is to create a separate visualization tool for each type of simulation. Only closely related applications, for instance different simulations of environmental phenomena, enable an usage of a shared control script. The user is free to decide in which phase of the computation the visualization tool is activated. Alternatively, the tool can be run after the whole computation is completed. While the computations are performed, the user can observe the contents of directories with input and output data on the Storage Element. A simulation output is in a certain phase considered to be the input for the visualization process. In the next cycle, the user prepares a new configuration file and runs the visualization tool as a new application.
