Peer-to-Peer 3D Streaming Dissertation Oral Exam Shun-Yun Hu Department of Computer Science and...

Preview:

Citation preview

Peer-to-Peer 3D Streaming

Dissertation Oral Exam

Shun-Yun Hu

Department of Computer Science and Information EngineeringNational Central University

Dissertation Advisor: Prof. Jehn-Ruey Jiang

2009/11/17

IEEE INFOCOM 2008

IEEE INFOCOM 2008

IEEE INFOCOM 2008

Motivation

Two trends in virtual environments (VEs)Larger and more dynamic contentMore worlds

Content streaming is needed80% - 90% content is 3D (e.g., 3D streaming)

How to support millions of concurrent users?

5/

Imagine you start with a globe

Zoom in…

To Chung-Li

and NCU

Right now it’s flat…

But in the near future…

Outline

Introduction Background A Model for P2P 3D Streaming The Design and Evaluation of FLoD FLoD Extensions Discussions Conclusion

12/

13/

What is 3D streaming?

Continuous and real-time delivery of 3D contentover network connections to allow user interactions without a full download.

14/

Object streaming

Hoppe 1996Progressive Meshes

15/

Scene streaming Multiple

objects Object

selection & transmission

Teler &Lischinski2001

16/

Visualization streaming Large volume Time-varying Resource

intensive

Olbrich & Pralle 1999

17/

Image-based streaming

Server-rendered

Thin clients Less

responsive

Cohen-Or et. al.2002

18/

3D streaming vs. media streaming

Video / audio media streaming is very matured

User access patterns are different for 3D content Highly interactive Latency-sensitive Behaviour-dependent Non-sequential

Analogy Constant & frequent switching of multiple channels

The scalability problem

Client-server: has inherent resource limit

Resource limit

[Funkhouser95]

19/

A potential solution

Peer-to-Peer: Use the clients’ resources

Resource limit

[Keller & Simon 2003] 20/

Outline

Introduction Background A Model for P2P 3D Streaming The Design and Evaluation of FLoD FLoD Extensions Discussions Conclusion

21/

22/

World model & area of interest (AOI)

23/

Model and assumptions

For a given object (mesh or texture)

All content is initially stored at a server

State managementSmall & updatable (~ KB)May require security / anti-cheatingEx. Avatar positions, health points, equipments

Content managementLarge & relatively static (~ MB)May authenticate via hashingEx. 3D polygonal models & textures

State vs. content management

24/

25/

3D streaming requirements

Streaming qualityUser's perspective“how much?” & “how fast?”Speed

ScalabilityServer's perspectiveHow to offload?Concurrent users

Challenges for P2P 3D streaming

Distributed visibility determination Minimize server involvement Efficient determination without global knowledge

Dynamic group management Discovery of data sources Continuous avatar movements and real-time constrain

Peer & piece selection Optimal visual quality Content availability and bandwidth constrain

26/

A conceptual model

Pre-install: movement, rendering (client) 3D streaming: partition + fragmentation (server)

prefetching + prioritization (client) P2P: selection (client)

27/

P2P 3D streaming issues

Object discovery

Source discovery

State exchange

Content exchange

P2P video/file sharing

28/

Outline

Introduction Background A Model for P2P 3D Streaming The Design and Evaluation of FLoD FLoD Extensions Discussions Conclusion

29/

30/

Observation

Users tend to cluster at hotspots Overlapped visibility = shared content

31/

Object discovery via scene descriptions

star: self triangles: neighborscircle: AOI rectangles: objects

32/

Source (neighbor) discovery via VON

Boundary neighbors

New neighbors

Non-overlapped neighbors

[Hu et al., IEEE Network, 2006]

Voronoi diagrams identify boundary neighbors for neighbor discovery

Flowing Level-of-Details (FLoD)

Object discovery: scene descriptions

Source discovery: VON

State exchange: query-response (pull)

Content exchange: random peer selection sequential piece

selection

33/

34/

System architecture

Data flows

(A): scene request list (B): scene descriptions

(C): piece request list (D): object pieces

35/

Prototype experiment

Progressive models in a scene (by NTU) Peer-to-peer AOI neighbor requests (by NCU)

36/

Prototype experiment

Data 3D scene from a game demo (total ~50 MB)

Setup 100 Mbps LAN 10 participants, 48 logins captured in 40 min.

Results Found matching client upload & download Avg. server request ratio (SRR): 36%

37/

Simulation setup

Environment 1000x1000 world, 100ms / step, 3000 steps client: 1 Mbps / 256 Kbps, server: 10 Mbps (both)

Objects Random object placement (500 objects) Object size based on prototype (~ 15 KB / object)

User behavior Random & clustering movement (1.5 * ln(n) hotspots)

38/

Simulation metrics

Scalability Bandwidth usage (Kbytes / sec) Server request ratio (% obtained from server)

Streaming quality Base latency (delay to obtain 1st piece) Fill ratio (obtained / visible data)

39/

40/

Server bandwidth usage

41/

Client bandwidth usage (random)

42/

Client bandwidth usage (cluster)

43/

Effect of user density

44/

Fill ratio

45/

Base latency

46/

Effect of upload bandwidth

Outline

Introduction Background A Model for P2P 3D Streaming The Design and Evaluation of FLoD FLoD Extensions Discussions Conclusion

47/

Problems with basic FLoD

Source discovery: too few sources State exchange: pull may be slow Content exchange: better than random?

Real environment considerationsPeer heterogeneityBandwidth utilization

48/

FLoD enhancements

Enhanced peer & piece selection Wei-Lun Sung (ACM NOSSDAV’08)

Bandwidth-aware streamingChien-Hao Chien (ACM NetGames’09)

49/

50

Enhanced Selection

Proactive notification of availability (push) Periodic incremental exchange of content

availability information with neighbors.

Msg_Type Obj_ID Max_PID Obj_ID Max_PID ‧‧‧‧

incremental content information

50/

51

Multi-Level AOI Request

Localized requests may prevent contentions Peers request from closer neighbors/levels first

51/

Compare enhanced strategy with FLoD

Simulation Environment

52/

53

Base Latency

53/

54

Fill ratio

54/

Bandwidth-aware Peer Selection

Region-based Peer List to increase sources Pre-allocation of connection channels

Multi-source peer selection Channel neighbors (bandwidth reservation) AOI neighbors (no response guarantee) Server (no response guarantee)

Tit-for-Tat peer selection (from BitTorrent) Channel-neighbor first Higher contributor first

55/

Simulation environment

World Size 1000 x 1000 (units)

Cell Size 100 x 100 (units)

AOI Radius 100 (units)

Time steps 1500 (steps/ sec)

Object Data Size Range

100 – 300 (KB)

% of Base Piece 10%

Refinement Piece Size

5 (KB)

Server Bandwidth Download/Upload

1000/ 1000 (KB/sec)

User Bandwidth Distribution

Downlink (KB/sec)

Uplink (KB/sec)

Fraction of

nodes96 10 0.05

187 30 0.45375 100 0.40

1250 625 0.10

[Bharambe et al, 2006]

56/

Streaming quality (= BW utilization)

100 to 500 objects, fixed at 100 peers

57/

System scalability

50 to 450 peers, fixed 300 objects

58/

Fill ratio time-series (QoS)

original FLoD Enhanced

59/

Outline

Introduction Background A Model for P2P 3D Streaming The Design and Evaluation of FLoD FLoD Extensions Discussions Conclusion

60/

LODDT (Cavagna et al. 2006)

‧ ‧‧

Object Tree Node Aura

U

61/

HyperVerse (Botev et al, 2008)

Backbone + overlay architecture

62/

Comparisons

63/

Outline

Introduction Background A Model for P2P 3D Streaming The Design and Evaluation of FLoD FLoD Extensions Discussions Conclusion

64/

Summary

P2P 3D streaming has four main issues Object discovery Source discovery State exchange Content exchange

FLoD demonstrates that P2P allows Much lower server resource usage Better performance in crowding

FLoD’s performance can be enhanced with Pushed-based state exchange Pre-allocated fixed-size bandwidth channels

65/

Conclusion

3D streaming could become an important net traffic Non-sequential access Latency-sensitive

Peer-to-peer streaming is promising Reduce server resource usage Dynamic interest groups

New area with many interesting issues Graphics: progressive encoding / decoding, compression Networking:group discovery, prefetching, topology, versioning

66/

Future works

Practical AdoptionsDynamic content updateTopology-aware P2P 3D streamingSecure P2P 3D streaming

Open questionsMany small worlds vs. one large worldHigh-definition (HD) contentIncentives & killer apps

67/

FLoD publications

1. Shun-Yun Hu, "A Case for 3D Streaming on Peer-to-Peer Networks," in Proc. ACM Web3D, Apr. 2006, pp. 57-63.

2. Shun-Yun Hu, Ting-Hao Huang, Shao-Chen Chang, Wei-Lun Sung, Jehn-Ruey Jiang, and Bing-Yu Chen, "FLoD: A Framework for Peer-to-Peer 3D Streaming," in Proc. IEEE INFOCOM, pp. 1373-1381, Apr. 2008.

3. Wei-Lun Sung, Shun-Yun Hu, and Jehn-Ruey Jiang, "Selection Strategies for Peer-to-Peer 3D Streaming," in Proc. NOSSDAV, May. 2008.

4. Chang-Hua Wu, Shun-Yun Hu, and Li-Ming Tseng, "Discovery of Physical Neighbors for P2P 3D Streaming," in Proc. ICUMT, Oct. 2009.

5. Mo-Che Chan, Shun-Yun Hu, and Jehn-Ruey Jiang, "Secure Peer-to-Peer 3D Streaming," Multimedia Tools and Applications, vol. 45, no. 1-3, Oct. 2009, pp. 369-384.

6. Chien-Hao Chien, Shun-Yun Hu, and Jehn-Ruey Jiang, "Bandwidth-Aware Peer-to-Peer 3D Streaming," in Proc. NetGames, Nov. 2009.

7. Shun-Yun Hu, Jehn-Ruey Jiang, and Bing-Yu Chen, "Peer-to-Peer 3D Streaming," IEEE Internet Computing, to appear, 2009.

68/

69/

Q & A

Thank you!

http://ascend.sourceforge.net

70/

Related work 3D streaming

Progressive meshes [Hoppe 96] Geometry image [Gu et al. 02] Scene streaming [Teler and Lischinski 2001]

P2P media streaming Zigzag, oStream, Coolstreaming, Prime

Nonlinear media streaming Channel Set Adaptation (CSA) [Gotz, 2006]

P2P 3D streaming LOD-DT [Cavagna et al. 2006]

Secure P2P 3D streaming

How to authenticate content from untrusted peers?

Four types of contentWhole model (digital signature)Linear stream (hash chain)Independent stream (Rabin-based)Partially linear stream (hash DAG)

72/

Cache utilization

Experimental results

74

Extended Candidate Buffer Non-AOI neighbors may still possess data Maintain extra list of non-AOI neighbors

R

S Obj

74/