SLIDE 1
A Scalable Server for 3D Metaverses
Ewen Cheslack-Postava, Tahir Azim, Behram F.T. Mistree, Daniel Reiter Horn, Jeff Terrace, Philip Levis, and Michael J. Freedman
sirikata.com
Metaverses
2
A Scalable Server for 3D Metaverses Ewen Cheslack-Postava, Tahir - - PowerPoint PPT Presentation
A Scalable Server for 3D Metaverses Ewen Cheslack-Postava, Tahir - - PowerPoint PPT Presentation
A Scalable Server for 3D Metaverses Ewen Cheslack-Postava, Tahir Azim, Behram F.T. Mistree, Daniel Reiter Horn, Je ff Terrace, Philip Levis, and Michael J. Freedman sirikata.com Metaverses 2 Metaverses 3 Metaverses 3 Metaverses 3
SLIDE 2
SLIDE 3
Metaverses
3
SLIDE 4
Metaverses
3
SLIDE 5
Metaverses
3
SLIDE 6
Metaverses
- Games
- Augmented reality
- Historical recreations
- Collaborative visualization
- ... what will users create?
4
Applications:
SLIDE 7
5
SLIDE 8
6
SLIDE 9
7
SLIDE 10
8
SLIDE 11
9
These are systems problems.
SLIDE 12
10
Object Discovery
SLIDE 13
10
Object Discovery
SLIDE 14
10
Object Discovery
SLIDE 15
11
How do we scale up the world without limiting the scope of interaction?
SLIDE 16
Sirikata
12
Seamless, scalable, and federated metaverses
SLIDE 17
Insight
13
The real world scales.
SLIDE 18
Design Principle
14
Scale by applying real-world constraints to the system.
SLIDE 19
15
Object Discovery
SLIDE 20
15
Object Discovery
SLIDE 21
Solid Angle Queries
16
Insight: Limited display resolution Solid angle: how large an object appears
SLIDE 22
17
Ideal
SLIDE 23
18
Distance, 3000 Objects
SLIDE 24
19
Solid Angle, 3000 Objects
SLIDE 25
20
Solid Angle & Aggregates, 3000 Objects
SLIDE 26
21
Ideal
SLIDE 27
22
Object Discovery
Solid angle queries are global. How do we efficiently and scalably evaluate solid angle queries?
SLIDE 28
Data Structure - BVH
23
SLIDE 29
Data Structure - BVH
23
A B C D A B C D
SLIDE 30
Data Structure - BVH
23
A B C D X A B C D X
SLIDE 31
Data Structure - BVH
23
A B C D X Y A B C D X Y
SLIDE 32
Data Structure - BVH
23
A B C D X Y Z A B C D X Y Z
SLIDE 33
Data Structure - BVH
24
A C D Y A B C D X Y Z Q B X Z
SLIDE 34
Data Structure - BVH
25
A C D Y A B C D X Y Z Q B X Z
SLIDE 35
Data Structure - BVH
26
A C D Y A B C D X Y Z Q B X Z
SLIDE 36
New Data Structure - LBVH
27
A C D Y A B C D X (A) Y (C) Z (A) B X Z Q
SLIDE 37
New Data Structure - LBVH
27
A C D Y A B C D X (A) Y (C) Z (A) B X Z Q A
SLIDE 38
LBVH
28
75 - 90% fewer nodes tested than with BVH
SLIDE 39
Dynamic Objects
29
Moving objects make the LBVH inefficient over time A B X
SLIDE 40
Dynamic Objects
29
Moving objects make the LBVH inefficient over time A B X
SLIDE 41
Dynamic Objects
30
SLIDE 42
Dynamic Objects
30
- Split between static and dynamic objects
SLIDE 43
Dynamic Objects
31
10 - 15% less expensive during short, 100 second experiment Benefit improves over time
SLIDE 44
Standing Queries
A B C D X (A) Y (C) Z (A) Cuts avoid redundant work
SLIDE 45
Standing Queries
A B C D X (A) Y (C) Z (A) Cuts avoid redundant work
SLIDE 46
Standing Queries
A B C D X (A) Y (C) Z (A) Cuts avoid redundant work
SLIDE 47
Standing Queries
A B C D X (A) Y (C) Z (A) Cuts avoid redundant work
SLIDE 48
Standing Queries
A B C D X (A) Y (C) Z (A) Cuts avoid redundant work
SLIDE 49
Standing Queries
A B C D X (A) Y (C) Z (A) Cuts avoid redundant work
SLIDE 50
Standing Queries
A B C D X (A) Y (C) Z (A) Cuts avoid redundant work
SLIDE 51
Standing Queries
38
20 - 56% increase in query evaluation rate
SLIDE 52
Aggregation
39
A B C D X (A) Y (C) Z (A)
SLIDE 53
Aggregation
39
A B C D X (A) Y (C) Z (A)
SLIDE 54
Aggregation
39
A B C D X (A) Y (C) Z (A)
SLIDE 55
Aggregation
39
A B C D X (A) Y (C) Z (A)
SLIDE 56
Aggregation
39
A B C D X (A) Y (C) Z (A)
SLIDE 57
Aggregate Queries
- Queries on a server are all similar
- Aggregate queries to reduce inter-
server querying load
- Filter results further before returning
results to querier
40
SLIDE 58
Server Discovery
41
10
10
SLIDE 59
Server Discovery
- Geometric server discovery
- Determine which other servers need to
be queried
- Built on same LBVH data structure
42
SLIDE 60
43
SLIDE 61
44
SLIDE 62
Also in the Paper
- Globally consistent distributed data
structure mapping regions to servers
- Global routing table enabling all-pairs
communication
- Forwarder with intuitive, physically-
based weighting emphasizing local traffic
45
SLIDE 63
Wiki World
46
Automatically find information about
- bjects on Wikipedia
SLIDE 64
But wait, there’s more...
- Audio: distant siren, roar of a crowd
- Efficient property updates
47
There are many more systems challenges at the intersection of systems, graphics, PL, databases, ... A few examples:
SLIDE 65
Thank You
Download and code at
sirikata.com
Questions?
48