Surajit A. Bose

1
Rethinking the Computer Cluster

• Surajit A. Bose
• Cluster Operations Manager
• Residential Computing
• Stanford University

2
Credits
3
Credits

• Howard Spivak, Brooklyn College, CUNY

4
Credits

• Howard Spivak, Brooklyn College, CUNY
• Larry Leifer, Mechanical Engineering, Stanford

5
Credits

• Howard Spivak, Brooklyn College, CUNY
• Larry Leifer, Mechanical Engineering, Stanford
• ME 013 N Team Buck-it oApples

6
Credits

• Howard Spivak, Brooklyn College, CUNY
• Larry Leifer, Mechanical Engineering, Stanford
• ME 013 N Team Buck-it oApples
• Eric Bennett

7
Credits

• Howard Spivak, Brooklyn College, CUNY
• Larry Leifer, Mechanical Engineering, Stanford
• ME 013 N Team Buck-it oApples
• Eric Bennett
• Will Bishop

8
Credits

• Howard Spivak, Brooklyn College, CUNY
• Larry Leifer, Mechanical Engineering, Stanford
• ME 013 N Team Buck-it oApples
• Eric Bennett
• Will Bishop
• Tristan Harris

9
Credits

• Howard Spivak, Brooklyn College, CUNY
• Larry Leifer, Mechanical Engineering, Stanford
• ME 013 N Team Buck-it oApples
• Eric Bennett
• Will Bishop
• Tristan Harris
• Jen Lau

10
Credits

• Howard Spivak, Brooklyn College, CUNY
• Larry Leifer, Mechanical Engineering, Stanford
• ME 013 N Team Buck-it oApples
• Eric Bennett
• Will Bishop
• Tristan Harris
• Jen Lau
• Alex Li

11
Credits

• Howard Spivak, Brooklyn College, CUNY
• Larry Leifer, Mechanical Engineering, Stanford
• ME 013 N Team Buck-it oApples
• Eric Bennett
• Will Bishop
• Tristan Harris
• Jen Lau
• Alex Li
• Jaireh Tecarro

12
Credits

• Howard Spivak, Brooklyn College, CUNY
• Larry Leifer, Mechanical Engineering, Stanford
• ME 013 N Team Buck-it oApples
• Eric Bennett
• Will Bishop
• Tristan Harris
• Jen Lau
• Alex Li
• Jaireh Tecarro
• Sindy Lee, Residential Computing, Stanford

13
About this Presentation
14
About this Presentation

• Stanfords cluster setup

15
About this Presentation

• Stanfords cluster setup
• The perceived problems

16
About this Presentation

• Stanfords cluster setup
• The perceived problems
• Student survey details

17
About this Presentation

• Stanfords cluster setup
• The perceived problems
• Student survey details
• Product design solutions from ME 013 N

18
About this Presentation

• Stanfords cluster setup
• The perceived problems
• Student survey details
• Product design solutions from ME 013 N
• Plan of action

19
About Stanford

• 12,000 undergrad and grad students, 10,000
  residing on campus

20
About Stanford

• 12,000 undergrad and grad students, 10,000
  residing on campus
• Widely differing living environments

21
About Stanford

• 12,000 undergrad and grad students, 10,000
  residing on campus
• Widely differing living environments
• Four large dorm complexes, 58 residences each

22
About Stanford

• 12,000 undergrad and grad students, 10,000
  residing on campus
• Widely differing living environments
• Four large dorm complexes, 58 residences each
• Eight undergrad and three grad stand-alone dorms

23
About Stanford

• 12,000 undergrad and grad students, 10,000
  residing on campus
• Widely differing living environments
• Four large dorm complexes, 58 residences each
• Eight undergrad and three grad stand-alone dorms
• Undergraduate Suites

24
About Stanford

• 12,000 undergrad and grad students, 10,000
  residing on campus
• Widely differing living environments
• Four large dorm complexes, 58 residences each
• Eight undergrad and three grad stand-alone dorms
• Undergraduate Suites
• Two grad residential complexes

25
About Stanford

• 12,000 undergrad and grad students, 10,000
  residing on campus
• Widely differing living environments
• Four large dorm complexes, 58 residences each
• Eight undergrad and three grad stand-alone dorms
• Undergraduate Suites
• Two grad residential complexes
• 36 small row houses

26
About Stanford (continued)

• Different residential identities

27
About Stanford (continued)

• Different residential identities
• Academic theme houses

28
About Stanford (continued)

• Different residential identities
• Academic theme houses
• Cultural or ethnic theme houses

29
About Stanford (continued)

• Different residential identities
• Academic theme houses
• Cultural or ethnic theme houses
• Co-ops

30
About Stanford (continued)

• Different residential identities
• Academic theme houses
• Cultural or ethnic theme houses
• Co-ops
• Dorms by Class Year

31
About Stanford (continued)

• Different residential identities
• Academic theme houses
• Cultural or ethnic theme houses
• Co-ops
• Dorms by Class Year
• Greek houses

32
The Clusters
33
The Clusters

• Each house has its own computer cluster

34
The Clusters

• Each house has its own computer cluster
• 77 clusters all over campus

35
The Clusters

• Each house has its own computer cluster
• 77 clusters all over campus
• Range in size from 215 computers

36
The Clusters

• Each house has its own computer cluster
• 77 clusters all over campus
• Range in size from 215 computers
• Public clusters are also available

37
The Clusters

• Each house has its own computer cluster
• 77 clusters all over campus
• Range in size from 215 computers
• Public clusters are also available
• Two general purpose clusters

38
The Clusters

• Each house has its own computer cluster
• 77 clusters all over campus
• Range in size from 215 computers
• Public clusters are also available
• Two general purpose clusters
• One iLounge

39
The Clusters

• Each house has its own computer cluster
• 77 clusters all over campus
• Range in size from 215 computers
• Public clusters are also available
• Two general purpose clusters
• One iLounge
• One language lab

40
The Clusters

• Each house has its own computer cluster
• 77 clusters all over campus
• Range in size from 215 computers
• Public clusters are also available
• Two general purpose clusters
• One iLounge
• One language lab
• One multimedia studio

41
The Clusters

• Each house has its own computer cluster
• 77 clusters all over campus
• Range in size from 215 computers
• Public clusters are also available
• Two general purpose clusters
• One iLounge
• One language lab
• One multimedia studio
• Non-AComp clusters

42
The Clusters

• Each house has its own computer cluster
• 77 clusters all over campus
• Range in size from 215 computers
• Public clusters are also available
• Two general purpose clusters
• One iLounge
• One language lab
• One multimedia studio
• Non-AComp clusters
• ITSS, academic departments, etc.

43
The Realities
44
The Realities

• More and more residents are bringing their own
  computers to campus

45
The Realities

• More and more residents are bringing their own
  computers to campus
• 98 of students have their own computer

46
The Realities

• More and more residents are bringing their own
  computers to campus
• 98 of students have their own computer
• Several have more than one

47
The Realities

• More and more residents are bringing their own
  computers to campus
• 98 of students have their own computer
• Several have more than one
• The prevalence of laptops is increasing

48
The Realities

• More and more residents are bringing their own
  computers to campus
• 98 of students have their own computer
• Several have more than one
• The prevalence of laptops is increasing
• Ownership runs to 66 of undergraduates and 75
  of grad students

49
The Realities

• More and more residents are bringing their own
  computers to campus
• 98 of students have their own computer
• Several have more than one
• The prevalence of laptops is increasing
• Ownership runs to 66 of undergraduates and 75
  of grad students
• Wireless connectivity is also increasing

50
The Realities

• More and more residents are bringing their own
  computers to campus
• 98 of students have their own computer
• Several have more than one
• The prevalence of laptops is increasing
• Ownership runs to 66 of undergraduates and 75
  of grad students
• Wireless connectivity is also increasing
• The main quad, many academic buildings, all the
  libraries, and some student residences offer this

51
The 1,000,000 Question

• Do we still need computer clusters, particularly
  in the residences?

52
Lets get rid of them!
53
Lets get rid of them!

• Theyre expensive and difficult to maintain

54
Lets get rid of them!

• Theyre expensive and difficult to maintain
• Equipment costs

55
Lets get rid of them!

• Theyre expensive and difficult to maintain
• Equipment costs
• Imaging and maintenance software costs

56
Lets get rid of them!

• Theyre expensive and difficult to maintain
• Equipment costs
• Imaging and maintenance software costs
• Staff time

57
Lets get rid of them!

• Theyre expensive and difficult to maintain
• Equipment costs
• Imaging and maintenance software costs
• Staff time
• Staff and student salaries

58
Lets get rid of them!

• Theyre expensive and difficult to maintain
• Equipment costs
• Imaging and maintenance software costs
• Staff time
• Staff and student salaries
• Better ways to spend money

59
Lets get rid of them!

• Theyre expensive and difficult to maintain
• Equipment costs
• Imaging and maintenance software costs
• Staff time
• Staff and student salaries
• Better ways to spend money
• Software agreements for student computers

60
Lets get rid of them!

• Theyre expensive and difficult to maintain
• Equipment costs
• Imaging and maintenance software costs
• Staff time
• Staff and student salaries
• Better ways to spend money
• Software agreements for student computers
• Wireless infrastructure

61
Lets get rid of them!

• Theyre expensive and difficult to maintain
• Equipment costs
• Imaging and maintenance software costs
• Staff time
• Staff and student salaries
• Better ways to spend money
• Software agreements for student computers
• Wireless infrastructure
• In-room network connection support

62
So we looked at actual usage
63
So we looked at actual usage

• We asked students about their use of the computer
  clusters

64
So we looked at actual usage

• We asked students about their use of the computer
  clusters
• Part of Residence Life survey for undergrads

65
So we looked at actual usage

• We asked students about their use of the computer
  clusters
• Part of Residence Life survey for undergrads
• Stand-alone survey for graduate students

66
So we looked at actual usage

• We asked students about their use of the computer
  clusters
• Part of Residence Life survey for undergrads
• Stand-alone survey for graduate students
• 3,263 undergraduates (53) and 1,089 graduate
  students (17) responded

67
So we looked at actual usage

• We asked students about their use of the computer
  clusters
• Part of Residence Life survey for undergrads
• Stand-alone survey for graduate students
• 3,263 undergraduates (53) and 1,089 graduate
  students (17) responded
• We also asked students for suggestions

68
So we looked at actual usage

• We asked students about their use of the computer
  clusters
• Part of Residence Life survey for undergrads
• Stand-alone survey for graduate students
• 3,263 undergraduates (53) and 1,089 graduate
  students (17) responded
• We also asked students for suggestions
• Direct questions on the survey

69
So we looked at actual usage

• We asked students about their use of the computer
  clusters
• Part of Residence Life survey for undergrads
• Stand-alone survey for graduate students
• 3,263 undergraduates (53) and 1,089 graduate
  students (17) responded
• We also asked students for suggestions
• Direct questions on the survey
• Project for a Product Design class to design a
  useful and pleasing computer cluster experience

70
Survey Results
71
Survey Results

• 51 of graduate and 83 of undergraduate
  respondents have used the computer cluster

72
Survey Results

• 51 of graduate and 83 of undergraduate
  respondents have used the computer cluster
• 69 of graduate and 87 of undergraduate
  respondents rate the availability of computers as
  good or excellent

73
Survey Results

• 51 of graduate and 83 of undergraduate
  respondents have used the computer cluster
• 69 of graduate and 87 of undergraduate
  respondents rate the availability of computers as
  good or excellent
• The vast majority of graduates as well as
  undergraduates average less than an hour a day in
  the computer clusters

74
Average Number of Hours Spent in Cluster
75
Reasons for Cluster Use Undergraduate Students
76
Reasons for Cluster Use Undergraduate Students

• Printing (63)

77
Reasons for Cluster Use Undergraduate Students

• Printing (63)
• Study space away from room (38)

78
Reasons for Cluster Use Undergraduate Students

• Printing (63)
• Study space away from room (38)
• Available software or courseware (35)

79
Reasons for Cluster Use Undergraduate Students

• Printing (63)
• Study space away from room (38)
• Available software or courseware (35)
• Convenient Location (28)

80
Reasons for Cluster Use Undergraduate Students

• Printing (63)
• Study space away from room (38)
• Available software or courseware (35)
• Convenient Location (28)
• Group or partner work space (18)

81
Reasons for Cluster Use Graduate Students
82
Reasons for Cluster Use Graduate Students

• Printing (37)

83
Reasons for Cluster Use Graduate Students

• Printing (37)
• Convenient Location (31)

84
Reasons for Cluster Use Graduate Students

• Printing (37)
• Convenient Location (31)
• Other (30)

85
Reasons for Cluster Use Graduate Students

• Printing (37)
• Convenient Location (31)
• Other (30)
• Own computer not working

86
Reasons for Cluster Use Graduate Students

• Printing (37)
• Convenient Location (31)
• Other (30)
• Own computer not working
• In-room network connection not yet active

87
Reasons for Cluster Use Graduate Students

• Printing (37)
• Convenient Location (31)
• Other (30)
• Own computer not working
• In-room network connection not yet active
• Scanning (21)

88
Reasons for Cluster Use Graduate Students

• Printing (37)
• Convenient Location (31)
• Other (30)
• Own computer not working
• In-room network connection not yet active
• Scanning (21)
• Available software or courseware (18)

89
Reasons for Cluster Use Graduate Students

• Printing (37)
• Convenient Location (31)
• Other (30)
• Own computer not working
• In-room network connection not yet active
• Scanning (21)
• Available software or courseware (18)
• Study space away from room (15)

90
Desired Improvements
91
Desired Improvements

• Color Printing

92
Desired Improvements

• Color Printing
• More Windows computers

93
Desired Improvements

• Color Printing
• More Windows computers
• More quiet individual study spaces

94
Desired Improvements

• Color Printing
• More Windows computers
• More quiet individual study spaces
• Better black-and-white printing

95
Desired Improvements

• Color Printing
• More Windows computers
• More quiet individual study spaces
• Better black-and-white printing
• Better group study spaces

96
Desired Improvements

• Color Printing
• More Windows computers
• More quiet individual study spaces
• Better black-and-white printing
• Better group study spaces
• Large displays for collaborative work

97
Desired Improvements

• Color Printing
• More Windows computers
• More quiet individual study spaces
• Better black-and-white printing
• Better group study spaces
• Large displays for collaborative work
• Multimedia facilities

98
Desired Improvements

• Color Printing
• More Windows computers
• More quiet individual study spaces
• Better black-and-white printing
• Better group study spaces
• Large displays for collaborative work
• Multimedia facilities
• Scanners

99
Desired Improvements

• Color Printing
• More Windows computers
• More quiet individual study spaces
• Better black-and-white printing
• Better group study spaces
• Large displays for collaborative work
• Multimedia facilities
• Scanners
• Video editing software

100
Designing the Cluster Experience
101
Designing the Cluster Experience

• ME 013 N Designing the Human Experience

102
Designing the Cluster Experience

• ME 013 N Designing the Human Experience
• Introductory Product Design class taught in
  Winter 2003

103
Designing the Cluster Experience

• ME 013 N Designing the Human Experience
• Introductory Product Design class taught in
  Winter 2003
• ResComp was approached by faculty member and
  asked to guide a final project

104
Designing the Cluster Experience

• ME 013 N Designing the Human Experience
• Introductory Product Design class taught in
  Winter 2003
• ResComp was approached by faculty member and
  asked to guide a final project
• Students chose to focus on reengineering the
  cluster users experience

105
Observations
106
Observations

• Computer clusters are old and outdated

107
Observations

• Computer clusters are old and outdated
• Spaces not designed specifically for that purpose

108
Observations

• Computer clusters are old and outdated
• Spaces not designed specifically for that purpose
• Dont scale to new technology

109
Observations

• Computer clusters are old and outdated
• Spaces not designed specifically for that purpose
• Dont scale to new technology
• They resemble prison cells tiny, windowless,
  airless

110
Observations

• Computer clusters are old and outdated
• Spaces not designed specifically for that purpose
• Dont scale to new technology
• They resemble prison cells tiny, windowless,
  airless
• Confining

111
Observations

• Computer clusters are old and outdated
• Spaces not designed specifically for that purpose
• Dont scale to new technology
• They resemble prison cells tiny, windowless,
  airless
• Confining

Basically, theyre not fun places
112
Observations (continued)
113
Observations (continued)

• Generally, students go to the cluster for
  services they dont have or cant afford

114
Observations (continued)

• Generally, students go to the cluster for
  services they dont have or cant afford
• No computer of their own

115
Observations (continued)

• Generally, students go to the cluster for
  services they dont have or cant afford
• No computer of their own
• Need higher-performance technology

116
Observations (continued)

• Generally, students go to the cluster for
  services they dont have or cant afford
• No computer of their own
• Need higher-performance technology
• Need specific software

117
The Solution
118
The Solution

• Give students access to technology they want but
  cannot afford

119
The Solution

• Give students access to technology they want but
  cannot afford
• Thermal-dye photo printers

120
The Solution

• Give students access to technology they want but
  cannot afford
• Thermal-dye photo printers
• Scanners, digital camera, camcorder

121
The Solution

• Give students access to technology they want but
  cannot afford
• Thermal-dye photo printers
• Scanners, digital camera, camcorder
• Digital photo and video editing workstations with
  powerful processors

122
The Solution

• Give students access to technology they want but
  cannot afford
• Thermal-dye photo printers
• Scanners, digital camera, camcorder
• Digital photo and video editing workstations with
  powerful processors
• Powerful laptops that can be checked out

123
The Solution

• Give students access to technology they want but
  cannot afford
• Thermal-dye photo printers
• Scanners, digital camera, camcorder
• Digital photo and video editing workstations with
  powerful processors
• Powerful laptops that can be checked out
• Large flat screen monitors that you can control
  with your own laptop

124
The work environment
125
The work environment

• Students should want to work there

126
The work environment

• Students should want to work there
• Abundance of natural light with large windows
  along one side of the wall

127
The work environment

• Students should want to work there
• Abundance of natural light with large windows
  along one side of the wall
• Comfortable furniture (i.e. sofas and nice
  chairs )

128
The work environment

• Students should want to work there
• Abundance of natural light with large windows
  along one side of the wall
• Comfortable furniture (i.e. sofas and nice
  chairs )
• Quiet environment superior to dorm/lounge

129
The work environment

• Students should want to work there
• Abundance of natural light with large windows
  along one side of the wall
• Comfortable furniture (i.e. sofas and nice
  chairs )
• Quiet environment superior to dorm/lounge
• Small loft with padded, slanted walls, comfy
  chairs, and large windows which would serve as a
  reading room

130
The work environment

• Students should want to work there
• Abundance of natural light with large windows
  along one side of the wall
• Comfortable furniture (i.e. sofas and nice
  chairs )
• Quiet environment superior to dorm/lounge
• Small loft with padded, slanted walls, comfy
  chairs, and large windows which would serve as a
  reading room
• A group corner for collaborative work

131
Design models
132
Workstations
Printer
Other digital equipment
Movable Desk
133
Group Display
Whiteboards
Storage area for technology
Group Desk
Group Area
Put laptops here
134
Practicalities?
135
Practicalities?

• Some ideas are impractical

136
Practicalities?

• Some ideas are impractical
• Thermal-dye printers

137
Practicalities?

• Some ideas are impractical
• Thermal-dye printers
• Unstaffed but secure laptop checkout

138
Practicalities?

• Some ideas are impractical
• Thermal-dye printers
• Unstaffed but secure laptop checkout
• But the design principles are sound

139
Practicalities?

• Some ideas are impractical
• Thermal-dye printers
• Unstaffed but secure laptop checkout
• But the design principles are sound
• Attractive, functional, and flexible spaces

140
Practicalities?

• Some ideas are impractical
• Thermal-dye printers
• Unstaffed but secure laptop checkout
• But the design principles are sound
• Attractive, functional, and flexible spaces
• Built to account for collaboration, ergonomics,
  and scalability

141
Practicalities?

• Some ideas are impractical
• Thermal-dye printers
• Unstaffed but secure laptop checkout
• But the design principles are sound
• Attractive, functional, and flexible spaces
• Built to account for collaboration, ergonomics,
  and scalability
• Much of the technology exists already

142
Practicalities?

• Some ideas are impractical
• Thermal-dye printers
• Unstaffed but secure laptop checkout
• But the design principles are sound
• Attractive, functional, and flexible spaces
• Built to account for collaboration, ergonomics,
  and scalability
• Much of the technology exists already
• Event Sharing

143
What is Stanford doing?
144
What is Stanford doing?

• Buying scanners like crazy

145
What is Stanford doing?

• Buying scanners like crazy
• Looking at the possibility of installing group
  corners with large displays in some areas

146
What is Stanford doing?

• Buying scanners like crazy
• Looking at the possibility of installing group
  corners with large displays in some areas
• Coming up with a multi-year cluster furnishings
  design project master plan

147
What is Stanford doing?

• Buying scanners like crazy
• Looking at the possibility of installing group
  corners with large displays in some areas
• Coming up with a multi-year cluster furnishings
  design project master plan
• Keeping our eyes open and our fingers crossed

148
Questions?
149
Thanks!