Avatar Mobility in 1 Wei Tsang Ooi Mehul Motani Huiguang Liang - - PowerPoint PPT Presentation

avatar mobility in
SMART_READER_LITE
LIVE PREVIEW

Avatar Mobility in 1 Wei Tsang Ooi Mehul Motani Huiguang Liang - - PowerPoint PPT Presentation

Avatar Mobility in 1 Wei Tsang Ooi Mehul Motani Huiguang Liang Ian Tay Ming Feng Neo National University of Singapore 2 My life is so great that I literally wanted a second one! - Dwight Schrute, The


slide-1
SLIDE 1

Avatar Mobility in

黃瑋璨 新加坡國立大學

1
slide-2
SLIDE 2

Wei Tsang Ooi Mehul Motani Huiguang Liang Ian Tay Ming Feng Neo National University of Singapore

2
slide-3
SLIDE 3

“My life is so great that I literally wanted a second one!”

  • Dwight Schrute,

The Office

3
slide-4
SLIDE 4

256x256 m regions.

4
slide-5
SLIDE 5

avatar mobility: who is where, when

5
slide-6
SLIDE 6

why do we care?

6
slide-7
SLIDE 7

research in systems support for NVE

7
slide-8
SLIDE 8 8
slide-9
SLIDE 9

How to partition a world into regions and assign regions to servers considering

  • communication cost
  • hand-over rate
  • balancing server load

:

9
slide-10
SLIDE 10 10
slide-11
SLIDE 11

How to predict avatar movement (end therefore what a user will see next)?

11
slide-12
SLIDE 12 12
slide-13
SLIDE 13

AoI-based scheme

13
slide-14
SLIDE 14

How many connections? How stable are the connections?

14
slide-15
SLIDE 15

supernode-based scheme

15
slide-16
SLIDE 16

How to pick supernodes? How stable are the supernodes?

16
slide-17
SLIDE 17

how to simulate avatar mobility?

17
slide-18
SLIDE 18

random walk random waypoint clustered movement :

18
slide-19
SLIDE 19
  • r,

small-scale implementation

19
slide-20
SLIDE 20

no large-scale NVE available until recently

20
slide-21
SLIDE 21

482,594

residents logged in between 2-9 June 2008

21
slide-22
SLIDE 22

secondlife.com/whatis/economy-graphs.php

22
slide-23
SLIDE 23
  • collect mobility traces of

avatars in Second Life

  • what it means w.r.t. systems

design for NVEs?

23
slide-24
SLIDE 24

collecting traces

24
slide-25
SLIDE 25

how do avatars move inside a distributed virtual environment?

25
slide-26
SLIDE 26

how are avatars distributed within a region?

26
slide-27
SLIDE 27

how long do they stay at a location?

27
slide-28
SLIDE 28

do they move in groups?

28
slide-29
SLIDE 29

etc.

29
slide-30
SLIDE 30

FPS MMORPG NVE

30
slide-31
SLIDE 31

Linden, can we get access to the server traces?

No.

31
slide-32
SLIDE 32
  • Wrote our own client
  • Parses packets using libsecondlife
  • Insert bots into regions
  • Log positions of avatars every 10s
32
slide-33
SLIDE 33

difficulties

33
slide-34
SLIDE 34

running out of memory

34
slide-35
SLIDE 35

anti-bots policy

35
slide-36
SLIDE 36
  • ver crowded region
36
slide-37
SLIDE 37

inter-region tracking

37
slide-38
SLIDE 38
  • Wrote our own client
  • Parses packets using libsecondlife
  • Insert bots into regions
  • Log positions of avatars every 10s
38
slide-39
SLIDE 39

who is where, when (doing what)

39
slide-40
SLIDE 40 40
slide-41
SLIDE 41

Freebies

41
slide-42
SLIDE 42

The Pharm

42
slide-43
SLIDE 43

Isis

43
slide-44
SLIDE 44

Ross

44
slide-45
SLIDE 45

Mobility Patterns

45
slide-46
SLIDE 46

Freebies: number of visits to a cell

46
slide-47
SLIDE 47

Freebies: average pause time in a cell

47
slide-48
SLIDE 48

Freebies: average speed in a cell

48
slide-49
SLIDE 49

Isis: number of visits to a cell

49
slide-50
SLIDE 50

caching/prefetching based on popularity of locations?

50
slide-51
SLIDE 51

Isis: average pause time in a cell

51
slide-52
SLIDE 52

pick supernodes from sticky location?

52
slide-53
SLIDE 53

Isis: average speed in a cell

53
slide-54
SLIDE 54

mobility model: random walk + pathway ?

54
slide-55
SLIDE 55

churn rate

55
slide-56
SLIDE 56 56
slide-57
SLIDE 57 57
slide-58
SLIDE 58

Reasonably high churn (up to 6/min)

58
slide-59
SLIDE 59

1 min 10 min 1 hr 2 hr

Highly skewed. Some stay for hours.

59
slide-60
SLIDE 60

cannot pick supernodes uniformly

60
slide-61
SLIDE 61

clustering of avatars

61
slide-62
SLIDE 62

meeting: encounter between two avatars (within each other AoI)

62
slide-63
SLIDE 63

Meet many different avatars.

63
slide-64
SLIDE 64

1 min 10 min 1 hr 2 hr

Most meetings are short.

64
slide-65
SLIDE 65

Meeting size is large.

65
slide-66
SLIDE 66

high overhead in maintaining AoI neighbors

66
slide-67
SLIDE 67

meeting stability: avg meeting size

  • ver

num of avatars met

67
slide-68
SLIDE 68

Wide range of stability

68
slide-69
SLIDE 69
  • ther tidbits
69
slide-70
SLIDE 70

little temporal variations

can use historical information to predict future

70
slide-71
SLIDE 71

rotate 18% of the time

Second Life’s prefetching is wasteful

71
slide-72
SLIDE 72

25-35% revisits the same region in a day

region-based caching?

72
slide-73
SLIDE 73

proxy-based texture caching

73
slide-74
SLIDE 74

why textures?

74
slide-75
SLIDE 75

62 - 81%

  • f traffic are textures
75
slide-76
SLIDE 76

316 MB

  • f textures in Isis
76
slide-77
SLIDE 77

64m 80o

77
slide-78
SLIDE 78

64m 80o

78
slide-79
SLIDE 79

403 TB

  • f textures retrieved

in Isis in a day

79
slide-80
SLIDE 80

SL servers texture proxy clients

80
slide-81
SLIDE 81

what caching algorithm to used?

81
slide-82
SLIDE 82

2Q

82
slide-83
SLIDE 83

FIFO LRU

83
slide-84
SLIDE 84

FIFO LRU cache miss

84
slide-85
SLIDE 85

FIFO LRU cache hit

85
slide-86
SLIDE 86

FIFO LRU cache hit

86
slide-87
SLIDE 87

FIFO LRU scan resistant

87
slide-88
SLIDE 88 88
slide-89
SLIDE 89

3Q

89
slide-90
SLIDE 90

FIFO LRU Victim Buffer

90
slide-91
SLIDE 91

FIFO LRU cache hit Victim Buffer

(sorted by popularity)

91
slide-92
SLIDE 92

how to define popularity of texture?

92
slide-93
SLIDE 93

Freebies: number of visits to a cell

93
slide-94
SLIDE 94

little temporal variations

can use historical information to predict future

94
slide-95
SLIDE 95

popularity of texture = popularity of cell

95
slide-96
SLIDE 96

2Q 3Q OPT Ross 50 MB 0.58 0.62 0.70 Ross 25 MB 0.28 0.36 0.47 Freebies 50 MB 0.48 0.50 0.68 Freebies 25 MB 0.21 0.33 0.50

Per-byte Hit Rate

96
slide-97
SLIDE 97

conclusion

97
slide-98
SLIDE 98

understanding real avatar mobility is crucial to design good NVEs

98
slide-99
SLIDE 99

謝謝 歡迎發問及指教

99

99