Visualization: an essential tool for the analysis of Big Data

Transcription

1 Visualization: an essential tool for the analysis of Big Data Visualization White Paper SARA February 2012

2 Contents Executive summary 3 Introduction 4 Remote visualization: high-end visualization in the researcher s lab 5 Big Data requires Megapixels: tiled displays 8 Remote visualization 9 3D visualization 10 In-situ visualization 10 Implementation of workflows 11 Collaborative visualization, video conferencing and working together in the Collaboratorium 12 Facilities 13 Visualization on behalf of and by the main sectors of industry ( topsectoren ) 14 Tools and applications developed 15 Conclusion 16 References 17 Visualization White Paper SARA February 2012

3 Executive summary The scale and complexity of scientific data sets are growing at an exponential pace. A researcher needs user-friendly tools and infrastructure when exploring and analyzing Big Data. Scientific visualization is a faster and more effective means with which to acquire information and knowledge from data. Various methodologies and environments for analyzing Big Data, such as remote visualization, high-resolution visualization on multimegapixel tiled display panels, 3D visualization and in-situ visualization are discussed below in more detail. The Dutch national e-infrastructure provides researchers access to the expertise, tools and infrastructural services they need for their research. Together with other institutes SARA plays a key role in facilitating and supporting scientific research by making available expertise, user-friendly advanced remote visualization in addition to collaboration tools and services. Recently SARA has developed and launched a new service and collaborative facility for Big Data exploration and analysis together with the Netherlands escience Center (NLeSC). The new facility called Collaboratorium is based on the technology and services developed by SARA and has been designed to meet end user needs. The facility is located in SARA s ICT Synergy Hub building on the Science Park in Amsterdam. Figure 1: High-resolution visualization of simulated turbulence in atmospheric boundaries, viewed on the display inside the new Collaboratorium. Source: Professor Harm Jonker, Faculty of Applied Sciences, Delft University of Technology. Visualization White Paper SARA February

4 Introduction Scientific researchers today are having to contend with the exponential growth and complexity of data derived from observation, experiments and simulations. Data sets comprising dozens, perhaps even hundreds of Terabytes form the rule rather than the exception. Furthermore, the data acquired is becoming increasingly important within multiple disciplines, or data cannot be interpreted properly without collaborating with researchers active in other disciplines. Figure 2: Visualization of simulation of in-stent restenosis (clogging of blood vessels around a stent). Source: Project Coast, Alfons Hoekstra, University of Amsterdam, EU-FP6-IST- FET Contract ; visualization: SARA. Visualization has become an invaluable tool for exploring, analyzing and understanding Big Data. This requires usable, user-friendly and advanced tools, infrastructure and expertise. As a result of the technological advancements in high-speed networks, high-resolution displays and new interaction techniques, the needs of researchers, i.e. ease-of-use, basic interface, Big Data and collaboration, can now be met. The set-up of advanced visualization for the analysis and exploration of large and complex data sets has been coupled with facilities offering researchers the opportunity to collaborate remotely, hold presentations and communicate with their peers. New methodologies for visualizing and navigating through large data sets use advanced visualization hardware and software, and modern human computer interfaces. The combination of rendering facilities, high-resolution 2D and 3D displays, video conferencing and presentation facilities in a space ideally suited to the optimal facilitation of discussions enables researchers and groups of researchers to discuss important multi-disciplinary issues in collaboration with their remote peers from the relevant fields and e-science specialists. SARA has joined forces with NLeSC to develop the Collaboratorium, a new facility that directly tailors high-resolution displays, remote data rendering, interaction, cooperation and user-friendliness to the current and future needs of the science and business community. The Collaboratorium is based on the technology and services developed by SARA and has been designed to meet the needs of end users. SARA works closely on this development with CalIT2 (San Diego) 1, EVL (University of Illinois at Chicago) 2 and NCSA (University of Illinois at Urbana-Champaign) 3, which are the world s leading accredited institutions in this field. SARA has enjoyed an extensive and long-standing working relationship with researchers in the Netherlands and abroad, which in recent years has enabled SARA to acquire a thorough knowledge of the scientific field and the need for scientific data analysis tools based on visualization. SARA utilizes the knowledge and experience acquired for a number of facets of scientific visualization. The provision of such expertise and services, together with user support, continues SARA s long history, vast experience and expertise in creating advanced visualization for science, the government and the business community. The launch of NLeSC enables us to serve an increasing number of end users. Visualization White Paper SARA February

5 Remote visualization: high-end visualization in the researcher s lab As a result of the growing complexity and volume of data generated by scientific models, it is proving to be a serious challenge to create the corresponding visualizations. Often large, parallel rendering capacity is required in the form of a cluster of GPUs. It is also becoming more difficult to migrate data sets due to their increasing size. An obvious solution is to leave the data in the location where it was generated. Figure 3: NetherLight in Amsterdam is a major unique European hub in the worldwide optical Research and Education network. Source: GLIF, visualization: Bob Patterson, NCSA- UIUC, data: Maxine Brown, EVL-UIC. In response to the above constraints remote visualization utilizes the options offered by the fast networks that are now becoming available locally and worldwide, across the Atlantic Ocean, such as GLIF, the Global Lamba Integrated Facility, which enables the implementation of visualization applications on a dedicated high-end rendering facility. The facility is located in the data center and features a fast data link with a central storage system. Visualization images instead of large data files are sent through the user s network connection with the facility. This means that pixels rather than files are sent. Figure 4: The principle of remote visualization: the end user sends visualization assignments via a network connection to a remote rendering cluster, which in this case is located at SARA. As soon as the images have been rendered they are returned to the user as a pixel flow. In early 2011 SARA launched a Remote Visualization Service enabling users to create visualizations on data sets stored in SARA s data center. The user only needs a remote desktop viewer and controls the visualization process from their own laptop or work station. The national e-infrastructure provides access to remote visualization facilities including an advanced rendering cluster featuring 16 nodes with high-end GPUs. It is essential for SARA to have a clear data link with the other processing and data storage facilities within the national e-infrastructure, such as Huygens, the national supercomputer; LISA, the national copute cluster and the HPC Cloud infrastructure. Visualization White Paper SARA February

6 Megapixels and gigabits Depending on requirements, a high bandwidth is often needed to transmit visual data across the network. For scientific visualizations it is vital that no data are lost and that visualizations are rendered in as much detail and as realistically as possible. The factors that come into play are as follows: the resolution of both the remote display and the data the compression factor of the images interactivity, an interactive visualization will readily need 24 frames per second depth of colour, the default is 24 bits per pixel Numerical calculation: An interactive visualization on a display having 4,096 x 2,160 pixels (a resolution of 4K) of uncompressed scientific data requires: 4096 * 2160 *24 bits per pixel * 24 fps = 5.1 gigabits per second Whereas a scientist will often not have access to a similar bandwidth, the Remote Visualization Service works with the VNC desktop protocol that compresses data. As a result the bandwidth used will remain within the bandwidth available to most scientists on the SURFnet network. A VNC session that used ParaView for viewing the CosmoGrid dataset on 1,280 x 700 used around 76 megabits per second. Figure 5: Huygens, the national supercomputer: SARA s sixth generation supercomputer. SARA s first supercomputer, a Control Data Cyber 205, was installed in The remote visualization cluster provides an extensive package of open source visualization application software for working with data. This includes: Paraview, VTK and VisIt ParaView is a VTK-based open source, scientific visualization application designed for the visualization of large-scale data sets. ParaView is actively undergoing further development by Kitware in association with the Advanced Computing Laboratory in the Los Alamos National Laboratory. 4 The Visualization Toolkit (VTK) is an open source and freely available software package for 3D computer graphics, image editing and visualization. VTK consists of a large collection of C++ classes and several interfaces for programming languages, such as Tcl/Tk, Java and Python. 5 VisIt is a point-and-click 3D scientific visualization application that supports the most widely used visualization techniques on structured and unstructured grids, such as iso-contouring and volume rendering. VisIt can interactively handle very large data sets on account of its distributed and parallel design. 6 VMD VMD is a molecular visualization program for demonstrating, animating and analyzing large bio-molecular systems, supported by 3D graphics and built-in scripting. 7 Ferret Ferret is an interactive computer visualization and analysis environment designed, for instance, to meet the needs of oceanographers and meteorologists who wish to analyze large and complex structured data sets. 8 Blender Blender is an integrated application featuring open architecture enabling the creation of a wide range of 2D and 3D scientific visualizations and other content. Blender offers a broad spectrum of functionalities for modeling, texturing, lighting, animation and video post-processing in one single package. Blender is one of the world s most popular open source graphic 3D applications. 9 Visualization White Paper SARA February

7 Figure 6: High-resolution tiled panel displays for scientific visualization, which SARA demonstrated at various international congresses. Top right: parameter study of data from the Essence project during the Supercomputing 07 conference in Reno, Nevada; data were sent from Amsterdam as pixels. Source: Professor Henk Dijkstra, IMAU, Utrecht University, visualization IMAU. Top left and below: ultra-high, (4K) resolution media streaming of videos from Amsterdam to Supercomputing 08 in Phoenix, Arizona. Source: CineGrid Amsterdam; streaming: SARA. Aside from the above, domain-specific visualization software may perhaps also be made available. Users also have the option to install their own visualization applications for remote use. The national e-infrastructure provides access to the following display environments (see also next section): High-resolution, MultiTouch tiled display panel (7,680*2,160 pixels, 17 megapixels, 8x HD, 2x 4K). This set-up is present in the Collaboratorium. High-resolution tiled display panel (12,800*4,800 pixels, 61 megapixels). This mobile setup can be used on location, but is not equipped with thin-bezel displays. Both set-ups can be controlled by the Scalable Adaptive Graphics Environment (SAGE) middleware for high-resolution distributed visualization environments 10; 11; 12. SAGE is graphic streaming architecture for collaborative scientific visualization environments supporting resolutions of up to hundreds of megapixels. Figure 7: SAGE middleware for visualization and remote collaboration; SAGE provides a mashup of various data input sources on large tiled screens. Source: Electronic Visualization Laboratory, University of Illinois at Chicago (EVL-UIC). Collaborating researchers can run different applications on SAGE simultaneously, such as 3D rendering, remote desktop and video streaming on local or remote equipment. The images can be shared, for instance, by sending the pixels of each application to tiled display panels while SAGE also supports the export of pixels to a single screen or laptop computer. The climate survey conducted at the University of Utrecht by Professor Henk Dijkstra is a prime example of high-resolution remote visualization. Huge volumes of climate data generated by and stored on Huygens, the national supercomputer in Amsterdam, were visualized and the images were subsequently transmitted via a dedicated SURnet6 10 gigabit 13 lightpath to the local high-resolution tiled display panel for analysis. Professor Dijkstra was honored with one of the SURFnet Enlighten Your Research awards in recognition of the project. Visualization White Paper SARA February

8 Big Data requires megapixels: tiled displays Large scientific data sets are often too large to visualize on a standard desktop LCD screen. Tiled displays featuring a multiple of the resolution of a single screen can be created by connecting a number of LCD screens to each other. The resolution of a 2x2 tiled display logically achieves a resolution four times higher than that of a single screen while a 5x3 tiled display achieves a resolution 15 times higher. A tiled display featuring 8,000 x 2,700 pixels (21.6 megapixels) can consequently be created with several standard displays consisting of 1,600 x 900 pixels. Figure 8: Visualization of ultra high resolution images during GLIF at CERN in Geneva (2010). CosmoGrid simulation images of dark matter interaction were sent from SARA to CERN streamed across a SURFnet 40 gigabit per second (Gbps) optical connection. A bandwidth of more than 32 gigabits per second (sustained) was used to render adequate high-frequency images having a resolution of 13,200 * 5,200 pixels (around 68 megapixels per frame). Source: Professor Simon Portegies Zwart, Leiden Observatory, Leiden University. Visualization: SARA; streaming application: SARA with the support of SURFnet; end-to-end network connection: SURFnet, SARA and Extreme Networks. Tiled displays can be used for various types of visual data analysis as follows: 1. The visualization of ultra high resolution images (such as high-resolution geo-data, high-resolution satellite data or high-resolution astrophysics simulations). 2. The exploration and comparison of a large number of very similar data sets (such as parameter studies in climate modeling or time exposures in medical spectroscopy). Figure 9: Example of the visualization of parameter studies: comparison of various parameters on several tiled displays in a single time sequence (synchronous videos) to determine correlation. Source: SAMOC project, Professor Henk Dijkstra, IMAU, Utrecht University. Visualization using a tiled display: IMAU. Visualization White Paper SARA February

9 3. Multi-modal visualization, accumulating a large number of various sources of data on a single screen to explore and discuss relationships and their correlation (e.g. various sources of biological data, such as DNA sequences, a genetic map and micro array data). Figure 10: Multi-modal visualization: assembling multiple output flows on a tiled display for the purpose of consulting various data sources simultaneously. Source: Han Rauwerda, Swammerdam Institute for Life Sciences, University of Amsterdam. Remote visualization The infinite possibilities of this basic technology combined with data streaming across ultra high bandwidth networks were already demonstrated several years ago in 2006 amongst others by SARA, when large data sets were streamed from Amsterdam to a 100 megapixel screen in the USA across a transatlantic 20 gigabit network connection 16. Figure 11: High-resolution images were created using the remote visualization environment. Data from the CosmoGrid project were rendered on SARA s remote visualization cluster with the Paraview application. The user interface (left) runs on the scientist s workstation or laptop. The images rendered are returned as a pixel flow to a display set-up, in this case a 5*3 tiled panel at SARA (right). Source: CosmoGrid, Professor Simon Portegies Zwart, Leiden Observatory, Leiden University. Visualization: SARA. A further example of a high-resolution remote visualization of a Big Data set is the CosmoGrid data set, simulating the creation of the universe (Professor Simon Portegies Zwart, et al, Leiden University). One single time step in this data set contains 2,048^3 or 8.6 billion particles. It is virtually impossible to study the data properly on a single HD screen. For a demonstration at CERN, SARA created a visualization with a resolution of 13,600 x 5,200 pixels, approximately 30 times that of the HD resolution. The visualization was rendered and streamed from Amsterdam and shown at CERN in Geneva. During the demonstration bandwidths of more than 37 gigabits were achieved. Visualization White Paper SARA February

10 3D visualization Some more complex data sets are inherently multi-dimensional. In certain cases stereo visualization may provide added value when analyzing these data sets. It has frequently been proven that objects rendered in 3D are easier to recognize and interpret. Various visualization packages are capable of generating 3D images. The advent of consumer-grade stereo (3D) displays makes 3D visualization accessible to low-cost visualization applications. Streaming high-resolution 3D stereo images across networks is still subject to research (in the CineGrid project, for instance) 14; 15, but will shortly be available for large-scale use. Figure 12: Desktop 3D visualization: a 3D desktop environment has become accessible for scientifc visualization with the advent of the relatively inexpensive 3D HD televisions. As a result it is now feasible to stream stereo images from a remote rendering environment or pre-renderd 3D videos to the scientist s workstation. Source: CosmoGrid, Professor Simon Portegies Zwart, Leiden Observatory, Leiden University. 3D visualization: SARA. Areas where 3D visualization can provide added value include molecular visualization, the visualization of medical images (such as MRI), gaming, astrophysics, climate modeling and cultural heritage (archeology and architecture). In-situ visualization In-situ visualization involves the visualization of data while producing the data. An example is presenting the full or partial results of a simulation while the computations are still being made. The developments in the simulation can therefore be monitored and provides a better picture of the strengths and/or weaknesses of a model and the applied model parameters. And if a model appears to be producing incorrect results barring a few iterations, the simulation can be discontinued, thus saving costly computer resources. It is also possible to debug a simulation code using an in-situ visualization because the internal status of the simulation can be viewed. In addition, in-situ post-processing can be carried out on the simulation data. By editing and/or selecting data based on the in-situ visualization, the resultant data can be efficiently stored for future use in cases where the full set of simulation data might be too large to enable comprehensive storage. Visualization White Paper SARA February

11 Computational steering expands on in-situ visualization. While studying the in-situ visualization the user may decide to change the parameters in order to control the simulation. The real-time interaction between the user and the simulation is therefore based on the in-situ visualizations. Figure 13: Parameter study of SAMOC climate simulation data: comparison of various parameters (such as flow and temperature) displayed next to each other in a daily workflow and on different scales in synchronous time steps collaboration between IMAU (Utrecht University) and SARA. Source: SAMOC project, Professor Henk Dijkstra, IMAU, Utrecht University. Visualization for a tiled display: IMAU. Workflow development: IMAU and SARA. Implementation of workflows If visualizations of a complete data set or part of a data set are regularly created based on variable parameters there may be a need for a work flow. This may involve zooming in on a particularly exciting part of the research in greater detail or comparing various parameter settings. SARA provides support for developing workflows, such as creating a web application for the end user in which he or she specifies parameters and then commissions the next visualization. SARA will subsequently create the visualization and stream the result to the user s display. Such a workflow is currently being used by the IMAU at Utrecht University for the analysis of a large set of climate data. Visualization White Paper SARA February

12 Collaborative visualization, video conferencing and working together in the Collaboratorium Collaboratorium : a new visualization and presentation facility for scientific research An advanced presentation and visualization space has been designed and developed within the national e-infrastructure and is located in SARA s ICT Synergy Hub building at Science Park Amsterdam. The visualization facility has been made available to support science and the business community. SARA, the newly established Netherlands escience Center (NLeSC) and other parties deploy the Collaboratorium to support scientific research in the Netherlands. Figure 14: Artist s impression of the Collaboratorium: a space designed for collaborative purposes using presentations, video conferencing, remote collaboration and highresolution visualization. The facility, which is operated by SARA and the Netherlands escience Center, is available from April 2012 for scientific research in the Netherlands. The Collaboratorium is financed by SURF with a grant from the Ministry of Economic Affairs, Agriculture and Innovation. Visualization: SARA. The facility offers excellent opportunities for collaboration and technology transfer between science, the government and the business community and thus contributes to government policy in the main sectors of industry. The space invites collaboration and reciprocal communication between local and remote users in an informal setting. Local support is available from visualization experts when using the facility and visualizing scientific or other data. The facilities are designed for both low-threshold use for day-to-day tasks and advanced use for users with specialist needs. A wide range of presentation equipment is supported, such as the usual presentations (PowerPoint), visualizations in 3D stereo and high-resolution interactive applications. Visualization White Paper SARA February

13 Facilities Inviting and informal space The Collaboratorium features a central conference table accommodating seven people. The table provides network, imaging, sound and power connections. Laptops can be connected directly. Figure 15: Thin-bezel tiled display wall: the edges of the displays are ultra thin; the edge between the screens is only seven millimeters in total. Displayed data: Professor Harm Jonker, Faculty of Applied Sciences, Delft University of Technology. Figure 16: Multi-touch interaction on a tiled display wall: various sources of data output are rendered in several windows on a tiled display, such as video conferencing images, copies of laptop screens, visualizations and a scrapbook. Individual windows can be moved, resized and minimized using the touch interface, which is similar to an ipad interface. Source: Electronic Visualization Laboratory, University of Illinois at Chicago. High-resolution display wall The video wall forms a key feature of the space and consists of 4x2 full HD thin bezel screens providing a total resolution of 7,680 x 2,160 pixels (16.6 megapixels). The bezels or ultra-thin edges of only a few millimeters between the screens ensure minimal visual disruption of the images. The video wall offers flexibility in choosing where what image flows (laptop images, video conferencing and visualization systems output) are shown on the video wall and the number of screens on which the images are projected. Consequently, an optimal mashup can be generated together with the simultaneous visualization of input from various data flows. Multi-touch interaction Chris Johnson, University of Utah, stated in his publication entitled Top Scientific Visualization Problems that one of the ten major problems involved in scientific visualization is the Human Computer Interface (HCI). The way in which people interact with computers, however, has changed significantly in recent years. This was largely brought about by the gaming industry and the touch interfaces used in smart phones and tablet computers. These methodologies for interacting with and manipulating data in a visualization environment are now available to researchers. The entire surface of the video wall features a multi-touch overlay, which enables direct interaction with the projected images and applications. The full screen functions similar to a giant ipad 17. Video conferencing An advanced video conferencing set is available, equipped with two cameras. From the display wall the first camera is directed towards conference participants seated at the table while the other camera provides an overall view of the entire space from the rear, which means that the images on the video wall and the presenter are also visible and can be rendered remotely. 3D stereo projection Apart from using the video wall, it is possible to project 3D stereo images on a separate projection screen. The screen has been installed in such a way that part of the video wall can continue to be used for showing 3D images. The 3D projection is added to the mashup on the video wall. Visualization White Paper SARA February

14 Easy and flexible operation The entire space can be operated with a touch panel making it easy to control the different components. In addition, a number of pre-defined settings can be easilyaccessed easily with a short-cut button. Figure 17: Visualization embedded in SARA s cloud solutions: in addition to the specialized rendering cluster several large systems are available for upscaling visualizations, such as the SARA HPC Cloud, the Huygens supercomputer and the Hadoop system. Integration with national HPCV facilities To facilitate scientific visualization the space is also linked to the national High-Performance Computing and Visualization (HPCV) infrastructure via high-speed networks, while SURFnet7 links the facility to the outside world. For the visualization of large data sets a remote rendering cluster is embedded in the data and processing environment: the national supercomputer Huygens, the HPC Cloud and Grid environment. Visualization specialists are available to support visualization projects for researchers. Visualization on behalf of and by the main sectors of industry ( topsectoren ) The main sectors of industry, the so-called topsectoren, can benefit from a large number of high-quality applications, including programs enabling the examination and analysis of large heterogeneous data sets. Applications are available for areas such as water management, the creative industry, medical image processing, the design of new materials and electronic equipment, green genetics (seed enhancement) and functional foods. Figure 18: The top 9 sectors of industry: visualization can play a key role in analyzing data from simulations, observations and experiments in various industry sectors. As part of government policy on the main sectors of industry more focus is being placed on supporting the market when using the national infrastructure, including advanced and high-resolution visualization. Ministry of Economic Affairs, Agriculture and Innovation. Copyright: Schwandt Infographics. geavanceerde en hoge-resolutie visualisatie. Ministerie van EL&I copyright: Schwandt Infographics. Visualization White Paper SARA February

15 Experience and expertise in advanced visualization are available for research purposes as well as the business community. Business sectors and government bodies that have used advanced visualization include the pharmaceutical industry (Johnson&Johnson Pharmaceutical Research & Development) 18, oil and gas exploration (Shell, Norsk Hydro), food (Unilever), media (NOB Cross Media Facilities), urban planning and design (the Directorate-General for Public Works and Water Management, the municipalities of Amsterdam, Almere, Rijswijk, and the aerospace industry (Fokker) and numerous other parties. Support for the visualization and analysis of complex data is available to the science and business communities. Tools and applications developed Over the years SARA has developed visualization tools in diverse fields in collaboration with various partners. Examples include: Various tools for the University of Amsterdam s Institute for Biodiversity and Ecosystem Dynamics (IBED), based on Google Maps and Google Earth for visualizing bird migration and influential factors The visualization of complex networks (mathematical graph structures) for the section Computational Science, University of Amsterdam A high-resolution visualization of Cosmogrid astrophysics simulation data for Leiden University Visualization of climate modeling for IMAU, Utrecht University A CAVE environment with moving avatars for the treatment of phobia for the University of Amsterdam, Department of Psychology SARAgene, 3D data analysis VR application for genomics research for Johnson&Johnson Pharmaceutics Virtual Supermarkt, a 3D serious game supporting research into purchasing behavior for VU University Amsterdam, Prevention and Public Health Department Visualization White Paper SARA February

16 Conclusion Visualization plays an invaluable role in the analysis of Big Data. The growth and complexity of data, coupled with the importance of the data for multiple disciplines, requires advanced visualization environments for data analysis and extensive specialist knowledge and expertise. The Dutch national e-infrastructure provides researchers access to such expertise and advanced remote visualization hardware and software. SARA thus plays a key role in association with the Netherlands escience Centre and other institutes in facilitating and supporting scientific research, not least because of the availability of the new Collaboratorium for data analysis, 3D and other visualizations, video conferencing and collaboration, which is operated by these institutes in the shared ICT Synergy Hub in Amsterdam. Figure 19: ICT Synergy Hub at Science Park Amsterdam, which houses SARA and its sister company Vancis, as well as the Netherlands escience Center (NLeSc) and the European Grid Initiative (EGI). Visualization White Paper SARA February

17 References 1. [Online] 2. [Online] 3. [Online] 4. [Online] 5. [Online] 6. [Online] 7. [Online] 8. ferret.wrc.noaa.gov/ferret/ [Online] 9. [Online] 10. SAGE: the Scalable Adaptive Graphics Environment. L. Renambot, A. Rao, R. Singh, B. Jeong, N. Krishnaprasad, V. Vishwanath, V. Chandrasekhar, N. Schwarz, A. Spale, C. Zhang, Goldman, J. Leigh, A. Johnson,. sl : Proceedings of WACE 2004, Nice, France. 11. Leigh, J. et al. The global lambda visualization facility: An international ultra-high-definition wide-area visualization collaboratory. Future Generation Computer Systems Volume: 22, Issue: 8, pp Enabling multi-user interaction in large high-resolution distributed environments. R. Jagodic, et al. 2011, Future Generation Computer Systems vol. 27 issue 7, July 2011, pp [Online] [Online] [Online] 16. Top Scientific Visualization Research Problems. Johnson, Chris.R. 2004, IEEE Computer Graphics and Applications, vol. 24, no. 4, pp Jagodic, R. [Online] 18. Mining the human genome using virtual reality. Stolk, Wielinga et al. sl : EGPGV 02 Proceedings of the Fourth Eurographics Workshop on Parallel Graphics and Visualization, Visualization White Paper SARA February