OpenGeo

1. The Value of
Open Source Software
Open Analytics Summit
Eddie Pickle
March 25, 2013

2. who is this guy?
I’m a geo guy working for a geo
software company

5. As we know,
There are known knowns.
There are things we know we know.
We also know
There are known unknowns.
That is to say
We know there are some things
We do not know.
But there are also unknown unknowns,
The ones we don’t know
We don’t know.

6. S#!* I know S#!* I don’t know
S#!*
I don’t know
I don’t know!

7. S#!* I know S#!* I don’t know
S#!*
I don’t know
S#
! I don’t know!
*I
sh
ou
d l
kn
ow

8. S#!* I know S#!* I don’t know
S#!*
I don’t know
I don’t know
S# ou
ab
!* t o
I s pe
ho n
ul sou
d
k n rc e
ow

10. development search
big data nosql

11. not a fad
not trend y

12. trend

13. etc!...

14. “how do you make a living
writing free software?”

15. etc!...

18. ma
rke
tp
ow
er
sta
dev ff
lop e
me
nt
flexibility
licens
e liabi
lities
clo
ud
rea
di n
es
s

19. #0: technical
superiority

20. • fewer bugs
• more modularity
• better security
• faster release cycles
• better performance

22. more developers and testers
than any one company
could possibly field

23. #1: cloud readiness
scaling
rapid deployment

24. $0 capital cost

26. your
server

27. $N per core
X cores
Ouch!

28. $N / instance ∗ X instances
Ouch!

29. #2: license liability
(lack of)

30. 30 30

31. 10 5

32. proprietary software licenses
are a legal liability
that must be managed

33. c trl- ctr
rl-C c
l- V
ct lC trl-Ctrl-C
r l- V
-V
trl-
rl-V
ctr
ctr
trl
ctr ctr
r l- C
C trl-Vtrlctrl-Vtrlr- c cV
r l- C
-C cct
l- V l- C
c
c
c c V trl- tlr-lC trl-C
l- c cr -V
ctr
-Cct
ct
V
l- C ctrl-V ctr
l- C
ct
ct
t c l-V
rtrl-
tcl-C ctrl-C c
l-Vtrtlr-Ctrl-C
trl
C trctrl-C ctr cc cV
trl-Vc-tVl-C c l-Vtrl-C
trl ctrl-
r
c
c
ctr
c trl- ctrl-Vtrl-C
-C
l- V
ctrl-V l-V ctrl-VV
c
rl- c
rl-V ct -V
r t
ctctrl-C cctrl-C
ctr ctr cctct
ct rl
l-V
rl-C
l-V l-C rl-rl-
ctr
ctrl
rl-V ctrl -V
ctr ctrl- ctc ctrc V
ctrl-Vctrl-C ctrl-C
ctr l-C V rltrl- l-trl-
ctrl-V
ct rl
ctrl- ctrl- ctrc -V VC C
-C
l- V
-C
V cC l-trl-
ctrl-V
r l- V ct - V
ctrl-Vctrl-C
ctr
rl-C
l-C ctrl- c trl C C l-
c tr l
V ctrl ctrl-
C
-V
trl-V trl-C
ctrl- cC tr -V tr
c c
trl-
-Cc Cc
V trl l- -V c t
ctrl
ctr ctrr
r l- C
ct rl C c
ctrl- ctrl- ctr -V C
l-V c tl--C trl-C
lV ctrl-V rl-C
ct rl-
c
V C l-
trl- ctrl- c ct
-Cct
ctrl- C
ct -V ct
V trl- rl C
C
V V ct rl- ctrl-Vc
C
l- V
trl
C
l-
ct
tr t

34. #3: flexibility
heterogeneity

35. • “flexibility” is an attribute
of components
– flexible components are easy
to connect and adapt

36. • “heterogeneity” is an
attribute of the system
– heterogeneous systems use
parts from many sources

37. or ?

38. internal services external wms
corporate
data warehouse
ArcSDE

39. internal services external wms
corporate
data warehouse

40. temporary
data warehouse
external wms
corporate
data warehouse
internal services

41. corporate warehouse
offline for 28 days

42. Web Framework
GIS Server Data Management
Tool
Database

43. ArcSDE

44. open source is a tool
that needs to be made useful
within the context
of existing systems

45. homogeneous systems
and single-vendor strategies
can be convenient

46. but,
lack flexibility
for fast adaptation

47. and,
homogeneous systems
can be susceptible to
“population catastrophes”

48. #4: staff development
and recruiting

49. “unobvious motivations
for adoption”

50. “Contrast an open-
source implementation
position with a ‘defined
skill set’ job…”

51. “...where the first
diagnostic action is to
reboot the server...”

52. “...and the second is to
call the vendor and wait
in a telephone hold
queue...”

53. “It is easy to understand
why open-source jobs
are prized.”
City of Northglenn, CO

54. #5: market power

55. Many organizations are in a hole
• More and better web services are required
• Need to serve more users
• Need to store and process more data
• Have shrinking budgets
55

56. The first law of holes
When you find yourself in one…
Stop digging!
56

57. Relative Value in Software Systems
Functionality and power
Costs
Control
57

58. Relative Value in Software Systems
The value of software can be expressed with a
simple equation, where
V = Value
F = Functionality and Power
O = Operational Costs
C = Control
58

59. Relative Value in Software Systems
V = (F/O) * C
59

60. Value
How much a desired object or condition is worth
relative to other objects or conditions
Value increases…
As software functionality increases relative to costs
As enterprises achieve more control (market power!)
60

61. Functionality and Power
Overall software utility
Includes…
– Reliability (system up-time and “buggy-ness”)
– Scale (numbers of users or amounts of data)
– Power (operations enabled, data or collaborators
incorporated into applications
– Etc.
61

62. Organizational Costs
Software costs include…
– License
– Maintenance
– Hardware and infrastructure
– Personnel costs to evaluate, test, deploy, operate and
maintain systems
– Etc.
62

63. Control
The control an enterprise has in its options to
access, use and maintain a given software
system.
– Legal rights of access to the source code
– Number of available suppliers and ability to influence
– Access to the software over the desired length of use
• And the probability that access might be limited or revoked
63

64. Relation of C (Control) to V (Value)*
Binary log function is postulated as model
*No Control = ? Value
64

65. Harms of Closed Source
• (Un)reliability
• Un-hackability
• Agency
• Lock-in
• Network Externalities
65

66. Evolution of Open Source
Give users control
66

67. 1. Build out functionality
V = F ( /O) * C
67

68. 2. Lower costs
V = /O
(F ) * C
68

69. 3. Increase Control
V = (F/O) * C
69

70. Elements of Control
• Ability to choose
– Alternative suppliers
– Bring in-house
• IP rights
– Right to modify
– Right to determine end of life
• Right to work on your schedule
• If you don’t have these, you have no Control
– And that will cost you
70

71. How does changing Control
affect Value?
It’s just math…
71

72. Control can affect Functional Power
V = (F * C) * 1/O
V = (F * 1) *
1/O
Assume C = 1 for two alternatives
72

73. Control can affect Functional Power
V = (F * C) * 1/O
V = (F * 1/2) *
1/O C = 1/2 for one alternative
Assume
73

74. Control can affect Costs
V = (1/O * C) * F
V = (1/O * 1) *
Ffor two alternatives
Assume C = 1
74

75. Control can affect Costs
V = ((1/O)*C) * F
V = ((1/O)*(1/2)) *
F
Assume C = 1/2 for one alternative
75

76. Value Comparisons
Commercial Open
Unsupported Open
Closed Source Source (OpenGeo
Source
Suite, etc.)
Functional Medium high Medium Medium high
Power, F Strong Functionality offset Reliable and scalable. High reliability and
by lower reliability and scalability. Expanding
scalability functionality
Operational High Medium Low
Costs, O Software license costs + No software license No software license costs.
Maintenance costs + high costs, but all Expert support lowers
Infrastructure costs due to maintenance costs are Maintenance costs, and
poor scalability. left to users. Good strong scalability lowers
Scalability resulting in infrastructure costs.
lower scalability costs.
Control, C Low Medium High
Closed source leads to Offers a solid Adds to control and choice
asymmetrical relations with alternative to closed by creating an excellent
customers and lowers source. Do it yourself, third option for enterprises,
customer control and choice or hire as needed. the ability to modify on your
own and/or have access to
OpenGeo resources.
76

77. Comparing Value for Software Alternatives
Commercial Open
Closed Unsupported
Source (OpenGeo
Source Open Source Suite)
Functional Power, Medium high Medium Medium high
F
Operational High Medium Low
Costs, O
Control, C Low Medium High
Relative Position Medium Low Medium High
on V = F/O * C
77

78. Questions and Observations
• Over time measures of relative value will shift
• The primary fields of contention would seem to
be Functional Power and Organizational Costs
• The role of Control is central
78

79. Questions and Observations
• Both open and closed source software may
continue to increase in Functional Power, but
the lead held by closed source is likely to shrink
• It is likely that closed source providers will
simply lower price
– They will be “borrowing” against future functionality (i.e.
today’s discount = tomorrow’s reduced capability)
– They will look to maintain Control as that is the long term
lever towards shifting Value their way
79

80. ma
rke
tp
ow
er
sta
dev ff
lop e
me
nt
flexibility
licens
e liabi
lities
clo
ud
rea
di n
es
s

81. Keep In Touch

Check out our blog at:
http://opengeo.org

Email me at:
Eddie Pickle
epickle@opengeo.org
81