UKON's Visual Computing Center for Cyber Security



Functional Components Description

The infrastructure provided by the University of Konstanz consists of several components which together enable the visual analysis of cyber security information. The individual components allow to efficiently analyze data using the machine learning infrastructure and afterward visualize it on either the Powerwall or in the Control Room setup. The interactive infrastructure aims to involve the user in the analysis process of security relevant information to semi-automatically generate and verify knowledge. The individual components and their interfaces are described in detail in the following sections. The visual computing center for cyber security consists of the following components: 

  • Powerwall: The Powerwall consists of a small cluster computer connected to the 5x2.1m² Powerwall with 11 megapixels, which was updated in 2016 for stereo projection and improved resolution. Through its wall-sized display, it is capable of displaying large-scale visualizations and animations of large datasets allowing for the development and evaluation of new modes of data analysis.

  • Control Room: The Control Room features a combination of three 4k displays and an HD-resolution capacitive touch table for data visualization and interaction technology research. The combination of high-resolution screens with a touch-table allows the design of state-of-the-art and application-specific interaction techniques.

  • Machine Learning Infrastructure: The Visual Computing Center for Cyber Security features a GPU-Cluster which consists of five NVIDIA Tesla P100 HPC graphic cards. Until now, these are among the fastest GPUs for high-performance computing. The compute-cluster is primarily used for the analysis of large and high-dimensional datasets with data mining techniques, including state-of-the-art neural networks.


Services provided

Research and Innovation

The Visual Computing Center for Cyber Security may be used for the research and development of innovative technologies in research projects. These projects should be mainly conducted by researchers in the domains Visual Analytics (VA) and Human-Computer Interaction (HCI) and may be focused on Cyber Security.  

Teaching and Training Activities

Individual components of the Visual Computing Center for Cyber Security can be used as assets in teaching activity. For instance, the Powerwall can be used in the teaching and presentation of visualization techniques requiring a high-resolution screen.


Keywords

Powerwall, Control Room, Machine Learning Infrastructure


Services

  • Powerwall: Small cluster computer connected to the 5x2.1m² Powerwall with 11 megapixels. Displays large-scale visualizations and animations of large datasets allowing for the development and evaluation of new modes of data analysis.

  • Control Room: Tthree 4k displays and an HD-resolution capacitive touch table for data visualization and interaction technology research.

  • Machine Learning Infrastructure: GPU-Cluster which consists of five NVIDIA Tesla P100 HPC graphic cards. Primarily used for the analysis of large and high-dimensional datasets with data mining techniques, including state-of-the-art neural networks.


Technical equipment

  • Assets:

    • Powerwall:

      • Display

        • Powerwall Display (Quantity:1)

          • Size: 5m x 2.1m

          • Resolution: 5224px x 2160px

          • Optical Output: 12,000 Lumen

      • Projectors

        • Galaxy 4k-12 projectors (Quantity:2)

          • Luminous system: D-ILA system

          • Pixel pitch: 6.8 millimicrons

          • Picture prime: 4,096 dots X 2,400 dots

          • Projection lens: 1.22 time electromotive zoom

          • Projection distance: Approx. 2.5m - 12m

          • Optical Output: 3,5000 Lumen

          • Iluminant Lamp: 825W Xenon Lamp

          • Contrast Ratio: 10,000:1

      • Workstation for running the Powerwall

        • Intel Xeon E5-2687W0 3.1 GHz (Quantity:2x8core)

        • Samsung M393B2G70BH0-CK0 16 GB DDR4 (Quantity:8x)

        • NVIDIA Quadro M6000 (Quantity:1)

        • Samsung 850 Pro 512GB SSD (Quantity:1)

      • Workstation for running the Powerwall with 3D

        • Intel Xeon E5-2687W0 3.1 GHz (Quantity:2x8core)

        • Samsung M393B2G70BH0-CK0 16 GB DDR4 (Quantity:8x)

        • NVIDIA Quadro M6000 (Quantity:1)

        • Samsung 850 Pro 512GB SSD (Quantity:1)

      • Streaming-Server

        • Intel Core i7-5930K (Quantity:1x6cores)

        • Kingston HyperX DDR4 2133 C14 8 GB RAM (Quantity:8x8GB)

        • NVIDIA GeForce GTX 960 (Quantity:1)

        • Samsung 850 Pro 512GB SSD (Quantity:1)

        • Samsung 850 Pro 2TB SSD (Quantity:2)

    • Control Room:

      • Displays

        • Samsung Digital Signane QH65H (Quantity:1)

          • Resolution: 3840x2160 (4k UHD)

          • Contrast: 6000:1

          • Diagonal Size: 65”

          • Type: LED

        • Eyevis EYE-LCD-6000-QHD-LD (Quantity:2)

          • Resolution: 3840x2160 (4k UHD)

          • Contrast: 5000:1

          • Diagonal Size: 60”

          • Type: LED

      • Workstation for controlling the Displays

        • Intel(R) Xeon(R) CPU X5690 3.47GHz (Quantity:1x6cores)

        • Kingston 9965433-034.A00LF 12 GB DDR3 RAM (Quantity:3)

        • NVIDIA Quadro K5000 (Quantity:3)

        • Samsung SSD 830 120 GB (Quantity:1)

        • Seagate Constellation ES.3 ST2000NM0011 2TB HDD (Quantity:2)

      • Touch-Displays

        • 3M Multi-Touch Display C4667PW (Quantity:1)

          • Resolution: 1920 x 1080 (FullHD)

          • Display Colors: 1.07 billion

          • Contrast Ration: 4000:1

          • LCD Technology: SPVA

          • Number of Touch Points: 60 points with palm rejection

          • Touch Point Speed: <12 milliseconds

          • Input Type: Finger, thin glove

          • Touch Communication: USB

          • Operating System Support: Win 8/7/Vista/XP, Linux, Mac

    • Machine Learning Infrastructure

      • GPU Server 1 with Multi-User Environment

        • Intel(R) Xeon(R) CPU E5-2637 v4 @ 3.50GHz (Quantity:2x4cores)

        • Kingston 9965604-001.D00G 16GB DDR4 RAM (Quantity:6)

        • NVIDIA Quadro M6000 GPU (Quantity:2)

        • 10 GB/s Intel Ethernet Controller 10G X550T (Quantity:1)

        • Samsung SSD 850 EVO 250GB (Quantity:2)

        • SEAGATE ST2000NM0034 HDD 2TB (Quantity:2)

      • GPU Server 2 with Multi-User Environment

        • Intel(R) Xeon(R) CPU E5-2640 v4 @ 2.40GHz (Quantity:2x10cores)

        • Samsung M393A2G40EB2-CTD 16GB DDR4 RAM (Quantity:8)

        • NVIDIA Tesla P100 GPU (Quantity:5)

        • 10 GB/s Intel Ethernet Controller 10-Gigabit X540-AT2 (Quantity:1)

        • Samsung SSD 850 500 GB (Quantity:2)

        • Western Digital WDC WD10JPLX-00M HDD 1TB (Quantity:6)


Use request

Non-profit