UNESCO Institute for Statistics Factors influencing open source software adoption Meeting on the Management of Statistical Information Systems (MSIS 2014)
Post on 24-Dec-2015
214 Views
Preview:
Transcript
UN
ESCO
Insti
tute
for S
tatis
tics
Factors influencing open source software adoption
Meeting on the Management of Statistical Information Systems (MSIS 2014)
Dublin, Ireland and Manila, Philippines 14-16 April 2014
Brian Buffett, Head of Statistical Services & Technology
UN
ESCO
Insti
tute
for S
tatis
tics
Overview
2
Context Innovation as a process Open Source Software adoption The Statistical Industry Summary
UN
ESCO
Insti
tute
for S
tatis
tics
Context Perspective of software adopters – not producers Explore statistical agency adoption of OSS Messages for ‘industry built’ software sharing
initiatives?
3
UN
ESCO
Insti
tute
for S
tatis
tics
Stages of Innovation Adoption Stages of <Open Source Software> adoption
1. Unaware of Open Source Software (OSS)2. Aware of OSS but not in the specific area (business or technical)3. Seeking information about OSS in the area4. Aware of OSS in the area, considering (possible exit point)5. Piloting OSS (possible exit point)6. Using OSS in production
Moving through the stages relies on the accumulation of knowledge
From external sources From internal experiences & projects
Individuals are fundamental to the process5
UN
ESCO
Insti
tute
for S
tatis
tics
Evolution The majority of organisations who consider OSS end up
adopting it Once organisations start using OSS, usage expands from
simple applications to mission-critical applications Once organisations have adopted OSS for a wide range
of applications, it is no longer viewed the same – it begins to be treated the same as commercial software
Once that point is reached, OSS is evaluated against the same criteria as commercial (COTS) s/w and must meet the same requirements and standards
6
UN
ESCO
Insti
tute
for S
tatis
tics
Towards successful OSS adoptionManaging Change
Build support for the initiative Identify & address internal barriers Identify risks Impart skills and knowledge on teams which will be
affected Introduce governance Adapt purchasing policies NOTE: OSS has a specific culture and there are
champions, detractors and neutral persons regardless of a specific business case
7
UN
ESCO
Insti
tute
for S
tatis
tics
Procurement example Government (or Agency) procurement, asset
management & requirements specification will need to evolve to support OSS and industry sharing models
Example from UK Government ICT Procurement Guide Evaluation is to be based on best value and consider
commercial and OSS solutions equally. In case of ties, OSS wins Example of requirements to be considered:
Security Scalability Minimum and essential functionality Transferability Manageability Maintenance & support requirements Total cost of ownership, including exit and transition costs 8
UN
ESCO
Insti
tute
for S
tatis
tics
Where are organisations focussing their efforts – top ranked goals
Improve integration between applications Increase process-oriented statistical production (reduce
silos) Reduce IT costs Increase industrialisation of statistics Increase collaboration with other statistical agencies Adopt or increase use of DDI and/or SDMX standards Use IT to increase innovation
10
UN
ESCO
Insti
tute
for S
tatis
tics
Policy Eleven (out of 43) organisations have organizational
policies regarding OSS adoption and use. Policy orientation:
Ten policies are oriented in favour of OSS adoption One policy is neutral
11
UN
ESCO
Insti
tute
for S
tatis
tics
Statistical OrganisationsOpenness to OSS
12
12345
Very resistant to OSS adoptionResistant to OSS adoptionNeutral to OSS adoptionOpen to OSS adoptionVery open to OSS adoption
05
101520253035
1 2 3 4 5
Freq
uenc
y
Bin
UN
ESCO
Insti
tute
for S
tatis
tics
Statistical Organisations and the Innovation Adoption Stages
13
123456
Unaware of OSSAware of OSS but uninformedSeeking information regarding OSSInformed about OSS, have not implementedPiloting OSSUsing OSS In Production
05
101520253035
1 2 3 4 5 6
Freq
uenc
y
Bin
UN
ESCO
Insti
tute
for S
tatis
tics
The official statistics industry Has shared goals and this was confirmed by the study Is widely researching, evaluating, and using OSS
consistent with high rate of adoption in entire public sector Is using OSS in all areas of statistical operations
Is planning to increase OSS usage in the next 48 months Views best practices as important but has not
implemented them / is not implementing them Identified OSS having lower operating costs than COTS Identified the same major concerns as other industries
Availability of service and support Security
14
UN
ESCO
Insti
tute
for S
tatis
tics
Which characteristics of OSS are most important?
Highest level of agreement across staff roles Ability to use OSS without restrictions Support in OSS of open standards Ability to reduce dependencies on vendors
Highest variance between staff roles Ability to view and modify source code Ability to create and distribute derivative works Ability to participate in & contribute to OSS development Ability to participate in & contribute to OSS communities
15
UN
ESCO
Insti
tute
for S
tatis
tics
Good Practices when Deploying OSS Managing maintenance and support costs Having rules governing OSS development, maintenance,
security, and support Having guidelines & approval processes for choosing OSS
products prior to adoption Rating & ranking OSS risks and mitigating those risks Monitoring the OSS community to ensure adequate
development & support of components used by the agency Ensuring relevant internal skills in OSS development or
operations Establishing a support model for the OSS product/solution Establishing internal development, testing, and change
management processes for OSS16
UN
ESCO
Insti
tute
for S
tatis
tics
Greatest Concerns regarding OSS Availability of service and support Security of the software Lack of relevant internal skills and knowledge of OSS
(development and operations) Product Immaturity Inability of OSS to help meet business goals
17
UN
ESCO
Insti
tute
for S
tatis
tics
Services from external companies re. OSS
In highest demand Consulting Integrating multiple OSS components Training Lifecycle support Integrating OSS with other software
Lowest demand Legal support OSS operations Certification of OSS OSS selection
18
UN
ESCO
Insti
tute
for S
tatis
tics
Security Consistently ranked as a top concern across all industries Security evaluations by specialist firms found security to
be comparable between OSS and COTS s/w OSS is not more or less secure than COTS
Heartbleed openssl backdoor notwithstanding!
19
UN
ESCO
Insti
tute
for S
tatis
tics
Summary Innovation adoption is a process – barriers, risks,
elements to manage all change as organisation changes People are critical to the process Major concerns are maintenance/support and security
& security of OSS is equivalent to COTS Statistics agencies need to carefully evaluate the true
lifecycle costs of OSS vs other models OSS adoption could provide good lessons for industry
sharing initiatives Efforts to increase software sharing could benefit from
targeting assistance at the various process stages20
UN
ESCO
Insti
tute
for S
tatis
tics
Thank You!
Brian BuffettHead of Statistical Services & Technology
Email: b.buffett@unesco.org
21
top related