Presentation is loading. Please wait.

Presentation is loading. Please wait.

Architecture & Performance Community Place case study Presented by u4068413 Jin Hyung, SEO.

Similar presentations


Presentation on theme: "Architecture & Performance Community Place case study Presented by u4068413 Jin Hyung, SEO."— Presentation transcript:

1 Architecture & Performance Community Place case study Presented by u4068413 Jin Hyung, SEO

2 Contents Introduction Architectural Possibilities Basic CP System Server Scalability Architectural features of current CP Performance Issue Distributed Architecture Conclusion

3 Introduction Aim : infrastructure that will allow many users to participate in a shared, interactive 3D world Large scale system capable of supporting many geographically dispersed users, interconnected through low bandwidth, high latency communication links

4 Possible System Architecture Client–Server Architectures : simplest method of building a distributed system Peer to Peer Systems : each client maintains a copy of data Hybrid Systems : merge client-server and peer to peer model

5 CP System Architecture Ⅰ Server Manager Dynamic data WWW browser CP browser Client System (PC) WWW browser CP browser Client System (PC) WWW server (httpd) Application Object (AO) VSAP VSCP HTTP Basic CP Architecture

6 CP System Architecture Ⅱ Browser  Works in conjunction with a HTML browser  Loads the 3D data file (in VRML 2.0 format)  Contacts the server via the VSCP that runs above IP  CP browser supports the VRML2.0 standard and uses Java as its scripting language Browser-server communication VSCP (Virtual Society Communication Protocol) :  Efficient communication of 3D scene transformation  Open-ended support for script specific message

7 CP System Architecture Ⅲ Server  Position tracker and message forwarder  Role :  Manage the state of Object on behalf of connected users  Carry out a similar function for any script level messages that are generated by a browser as a result of user interaction

8 Application Programming Models Ⅰ SSS (Simple Shared Script) Model  Small shared applications in the 3D world  Master browser AO (Application Object) Model  More complicated application  VSAP (Virtual Society Application Protocol)  Dynamic addition of VRML data and associated scripts to an existing scene  Commercial environment

9 Application Programming Models Ⅱ CP Browser Server Aura Manager Dynamic data Simple Shared ScriptApplication Object 4 1 1 2 2 3 3 4 5 5 5 6 6 Application Object (AO)

10 Scaleability For scalability in large application…. Static scene data is downloaded initially as part of the VRML file while dynamic data can be managed using local scripts plus message passing Offload some processing into the client browser using the local scripting facility Sophisticated applications can be managed by external processes and can use the local script to manage local updates in individual browsers Limit the number of messages needed between browsers

11 Spatial Areas of Interest Area of application can be divided into several groups Aura manager is responsible for defining groups of spatially co-located objects Each objects in the same group are maintained consistent together

12 Feature of CP Architecture Total state of the shared VRML scene is split between VRML data managed at the clients and data managed by the server since server is responsible for maintaining the location and some attribute data belong to client (ex> avatar data) By pushing data out to clients :  Reduce server load  Increase performance

13 Server Performance Ⅰ Sony workstation Sun Sparc SGI Indigo2Sony NEWS Client Server Server Performance Experiment Environment

14 Server Performance Ⅱ CPU load with the number of connection  Sony and SGI are comparable and Sun 10% more powerful  At 30% CPU idle time, Sony - 400 connections, SGI - 425 connections while Sun - 520 connections Network Traffic  Sony & SGI support a maximum 500 connections before network performance tails off while Sun does almost 630 connection AURASIZE & MAXINAURA  Increasing the MAXINAURA will lead to a growth in traffic  The interplay with the AURASIZE is less clear

15 Browser Performance Event processing cost within the browser  cost of this execution path is 0.99 ms (averaged over 3000 runs) Cost of basic networking function  average cost of 6.48 ms (repeated 3000 times) Application message cost  SSS model is more costly and sensitive to server load

16 Distributed Architecture Ⅰ WWW browser CP browser Client System (PC) WWW browser CP browser WWW browser CP browser User 1 User 2 User 3 VSCP Multicast based Group comms. Replica of Data item Master copy Of data item

17 Distributed Architecture Ⅱ Spatial model is used exactly as in the single server case It partitions the database into groups of spatially co-located objects who manage their consistency using a group communication model

18 Conclusion Most important goal of the CP project has been to provide an infrastructure that allows easy creation of such spaces, within a familiar framework, the WWW Scaling and performance issue should be always considered to support large numbers of users and more complicated scenes

19 Related Works WWW and using full 3D shared space Cybergate system (Blacksun Inc.) Moondo Pueblo Project ( Chaco Communication)

20 Q & A


Download ppt "Architecture & Performance Community Place case study Presented by u4068413 Jin Hyung, SEO."

Similar presentations


Ads by Google