Visualization in distributed systems. Overview.

1
Visualization in distributed systems. Overview.

• Remote visualization means interactive viewing of
  three dimensional scientific data sets over the
  global network.
• Scientists use remote parallel computer recourses
  in many scientific simulations.
• Scientific data sets are in the gigabyte or even
  terabyte size range.
• It is impossible or unreasonable to send the
  entire data set over the network.
• Moreover, the client usually has a limited amount
  of memory and CPU power for viewing and analyzing
  the data, and scientific data are too large to be
  processed by a single computer.
• So, we need powerful visualization tools to
  analyze massive data sets.
• Parallel visualization is a solution.

M.Iakobovski. P.Krinov, S.Muraviov
2
Visualization in distributed systems. Goals.

• RemoteViewer is aimed to quickly and easily
  process and visualize massive data sets (3D CFD
  simulations results).
• Scalar data such as pressure or temperature may
  be viewed as a series of iso-surfaces and/or as a
  series of slices
• Vector data such as the velocity field can be
  interactively explored using trajectories

Mesh Flow over plane Iso-surfaces
M.Iakobovski. P.Krinov, S.Muraviov
3
Visualization in distributed systems. Problem
statement.

• RemoteViewer
• Scalar fields visualization
• Vector fields visualization
• Data compression
• Parallel and grid realization

Structure
Visualization of Scalar and Vector
fields Iso-surfaces for scalar fields Trajectories
for vector field Cubic tetrahedral meshes
Computational Server Data Server Visualization Server ? Client Workspace
RemoteViewer
Tecplot
M.Iakobovski. P.Krinov, S.Muraviov
4
Visualization in distributed systems. Issues.
Scalar data visualization methods Syntesis
Reduction
Synthesis
a)
Reduction

• a) Edge removal
• b) Node removal
• c) Topology refinement

b)
c)
Vector data visualization General visualization
technique geometric visualization. Main
approaches Experimental analog to the flow
visualization Path line calculations (individua
l trajectory) Streakline calculation (fog or
smoke) Timelines calculations (coloration) Basic
principle - computation of massless particle
trajectories
Data animation
M.Iakobovski. P.Krinov, S.Muraviov
5
Visualization in distributed systems. Status.

• Operations on visualization server
• Data processing
• Iso-surfaces compression
• Data transferring to the client
• Operations on Client
• Setting of boundaries of the visual area (zoom)
  and resolution
• Setting of basic image characteristics (number of
  iso-surfaces, represented on the screen,
  corresponding function values number of
  trajectories of particles and coordinates of
  their starting points)
• The 3D image is displayed on the client computer
  screen and can be explored using rotation and
  zooming without referring back to the server
• If the closer examination of a smaller object
  fragment is required, the demand for image of
  this fragment is sent to the server
• The new image can approximate the object with the
  higher accuracy due to the reduction in data size

M.Iakobovski. P.Krinov, S.Muraviov