BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//talks.geoinfo-lab.org//foss4g-asia-2024//8RDZK7
BEGIN:VTIMEZONE
TZID:+07
BEGIN:STANDARD
DTSTART:20000101T000000
RRULE:FREQ=YEARLY;BYMONTH=1
TZNAME:+07
TZOFFSETFROM:+0700
TZOFFSETTO:+0700
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-foss4g-asia-2024-YZRW9Y@talks.geoinfo-lab.org
DTSTART;TZID=+07:20241216T165500
DTEND;TZID=+07:20241216T171500
DESCRIPTION:Introduction\n\nWith the spread of IoT and the increasing resol
 ution of observation sensors\, the total amount of geospatial information 
 data is increasing exponentially daily. On the other hand\, the increase i
 n resolution of display devices used to analyze and visualize these data i
 s reaching its limit due to various physical constraints. The maximum reso
 lution of commercially available display devices is 8K\; 4K or 5K is consi
 dered the upper limit for desktop use.\nUsing the OS's multi-display funct
 ion or a tiled display driver provided by the GPU manufacturer\, it is pos
 sible to create a display environment with an even larger area and higher 
 resolution. However\, the middleware provided by the GPU manufacturer curr
 ently has a maximum resolution limitation of 16K [1]\, which is the maximu
 m resolution that can be achieved on a single PC.\nHowever\, even if these
  mechanisms are used to create an ultra-high-resolution display environmen
 t\, it is only possible to render data within the web browser's heap memor
 y limit in the case of WebGIS applications. For example\, the 3DWebGIS vie
 wer provided by the Tokyo Digital Twin Project [2] cannot render 3DTiles [
 3] building data for all 23 wards of Tokyo at once (textured building data
  is used for areas provided with texture).\nIn this paper\, we introduce C
 hOWDER\, a web-based tiled display driver that enables distributed renderi
 ng of 3DWebGIS content across multiple web browsers\, as a solution to the
  above problems and report the results of memory load balancing experiment
 s using ChOWDER for distributed rendering.\n\nProposal of a distributed re
 ndering method for 3DWebGIS\n\nOne possible solution to the above problems
  is to distribute the display of one WebGIS content across multiple PCs (m
 ultiple web browsers). This makes it possible to display a WebGIS at a res
 olution that exceeds the upper limit of a single PC (web browser) and dist
 ributes the memory load required to display the content across each PC (we
 b browser).\nThe scalable display system ChOWDER[4][5]\, jointly developed
  by RIKEN Center for Computational Science and Kyushu University\, is an o
 pen-source tiled display driver that can create an ultra-high-resolution p
 ixel space by arranging multiple displays that display a web browser in fu
 ll-screen mode in tiles. It also supports distributed rendering of 3DWebGI
 S.\nThis function uses iTowns[6]\, an open-source 3DWebGIS\, as middleware
 . iTowns uses Three.js as a WebGL rendering library\, and Three.js has an 
 API that can offset the view frustum[7].\nThe view frustum must be split a
 ppropriately to split and display 3D content on multiple display devices. 
 ChOWDER uses the view frustum offset API of Three.js to split a single iTo
 wns content into multiple view frustums\, enabling multiple web browsers t
 o split and render 3DWebGIS content [8].\nHowever\, at the time of the pre
 vious report [8]\, when iTowns executed a 3DTiles load command\, it loaded
  all the data without judging whether it was inside or outside the view fr
 ustum range\, so distributed rendering did not improve memory utilization 
 efficiency. Since then\, the 3DTiles load process was improved in iTowns R
 elease 2.42.0\; in this paper\, we measured the amount of heap memory cons
 umed by each browser when iTowns content was distributed and rendered usin
 g ChOWDER on multiple web browsers and confirmed the memory load distribut
 ion achieved by this method.\n\nExperimental procedures and results\n\nThe
  experimental data used was the textured building data for Chiyoda\, Minat
 o\, and Chuo wards in Tokyo\, from the 3DTiles data distributed by the PLA
 TEAU project [9] of the Ministry of Land\, Infrastructure\, Transport and 
 Tourism of Japan.\nThe experiment first displayed the 3DTiles building dat
 a for the above three wards in full screen on a single 4K resolution displ
 ay using iTowns on ChOWDER. The heap memory size of the web browser at thi
 s time was 268MB.\nNext\, the same content was displayed on a ChOWDER dist
 ributed display consisting of four 4K displays arranged in two horizontal 
 and two vertical rows. Each display had a full-screen web browser. The hea
 p memory sizes of each web browser were 133MB\, 188MB\, 68.3MB\, and 37.7M
 B.\nFinally\, we conducted an experiment using nine 4K displays arranged i
 n three rows and three columns. The heap memory sizes of each web browser 
 were 66.8MB\, 122MB\, 140MB\, 84.3MB\, 87.2MB\, 56.9MB\, 41.2MB\, 38.4MB\,
  and 33.6MB.\nFrom these experimental results\, it can be said that distri
 buted rendering of 3DWebGIS using ChOWDER achieves memory load balancing.\
 nDuring distributed rendering\, the heap memory size of each web browser i
 s different because the amount of 3DTiles data contained in each responsib
 le drawing area is different. Also\, the total heap memory size of all bro
 wsers is larger than when rendering in a single browser because iTowns loa
 ds 3DTiles data that is wider than its view frustum\, and data loading in 
 overlapping areas occurs during distributed rendering.\n\nFuture work and 
 conclusion\n\nIn this experiment\, we measured the web browser's heap memo
 ry\, but did not measure GPU memory consumption. However\, because 3DWebGI
 S uses WebGL for rendering\, we believe that a more precise evaluation can
  be made by measuring GPU memory consumption as well.\nIn addition\, since
  distributing rendering across more web browsers is expected to further di
 stribute memory load\, we plan to conduct experiments by increasing the nu
 mber of distributed displays.\nIn this paper\, we have shown the limitatio
 ns of current 3DWebGIS when the data to be displayed increases\, and propo
 sed a distributed rendering method as a means to solve this problem\, and 
 introduced the view frustum offset API of Three.js\, iTowns\, a 3DWebGIS t
 o which it can be applied\, and ChOWDER\, a web-based tiled display driver
  that incorporates them\, as a means to realize this method. Furthermore\,
  we have presented the results of an experiment that shows that memory loa
 d distribution is achieved by distributed rendering using these and demons
 trated that this method is one solution to the increase in data to be disp
 layed in 3DWebGIS.\n\nReferences\n\n[1] Limitations. About NVIDIA Mosaic. 
 https://www.nvidia.com/content/Control-Panel-Help/vLatest/en-us/mergedProj
 ects/nvwks/SLI_Mosaic_Mode.htm Accessed July 29\, 2024.\n\n[2] Tokyo Digit
 al Twin Project. https://info.tokyo-digitaltwin.metro.tokyo.lg.jp/ Accesse
 d July 29\, 2024.\n\n[3] 3DTiles. The open specification for 3D data. http
 s://cesium.com/why-cesium/3d-tiles/ Accessed July 29\, 2024.\n\n[4] Kawana
 be\, T.\, Nonaka\, J.\, Hatta\, K.\, & Ono\, K. (2018\, September). ChOWDE
 R: an adaptive tiled display wall driver for dynamic remote collaboration.
  In International Conference on Cooperative Design\, Visualization and Eng
 ineering (pp. 11-15). Cham: Springer International Publishing.\n\n[5] ChOW
 DER GitHub repository. https://github.com/SIPupstreamDesign/ChOWDER Access
 ed July 29\, 2024.\n\n[6] iTowns (in French). https://www.itowns-project.o
 rg/ Accessed July 29\, 2024.\n\n[7] three.js API Reference. https://threej
 s.org/docs/#api/en/cameras/PerspectiveCamera.setViewOffset Accessed July 2
 9\, 2024.\n\n[8] Kawanabe\, T.\, Hatta\, K.\, & Ono\, K. (2020\, September
 ). ChOWDER: A New Approach for Viewing 3D Web GIS on Ultra-High-Resolution
  Scalable Display. In 2020 IEEE International Conference on Cluster Comput
 ing (CLUSTER) (pp. 412-413). IEEE.\n\n[9] Project PLATEAU portal site (in 
 Japanese). https://www.geospatial.jp/ckan/dataset/plateau Accessed July 29
 \, 2024.
DTSTAMP:20260511T143958Z
LOCATION:Room34-1104
SUMMARY:On the performance of distributed rendering system for 3DWebGIS app
 lication on ultra-high-resolution display - Tomohiro KAWANABE
URL:https://talks.geoinfo-lab.org/foss4g-asia-2024/talk/YZRW9Y/
END:VEVENT
END:VCALENDAR