[PPT] - The Natjonal Integrated Cyberinfrastructure System as an open PowerPoint Presentation

SLIDE 1

The Natjonal Integrated Cyberinfrastructure System as an

pen commons in South Africa

Bruce Becker, SANREN Competency Area, Meraka Instjtute, CSIR

bbecker@csir.co.za | @brusisceddu

EGI Technical Forum Lisbon, 2015

15/05/19 EGI Technical Forum, Lisbon, 2015 1

SLIDE 2

Outline

What is this “Open” you speak of ?
What is Open ?
What is an Open Commons ?
Systems Thinking for Cyberinfrastructure in

South Africa : NICIS

Implicatjons
For researchers, developers and operators
For the economy
For the region

SLIDE 3

Wherefore art thou ?

What does it mean to be a commons ?

SLIDE 4

SLIDE 5

Everybody comes from somewhere

Jenny Hallward

SLIDE 6

Everybody uses the Commons in their own way

Jenny Hallward

SLIDE 7

What is a Commons Anyway ?

… belonging to or afg fgectj tjng the whole of a community...

...anyone can contribute...

Share

SLIDE 8

The Commons, Openness and Science

“The more we understand about science and its complexitjes, the more important it is for scientjfjc data to be shared openly” htup:/ /creatjvecommons.org/science

“Science == Openness” ?

SLIDE 9

NRF Statement on Open Access

htu tup:/ /www.nrf.ac.za/media-room/news/statement-open-access-research-publicatj tjons-natj tjonal-research- foundatj tjon-nrf-funded

From 01 March 2015, authors of research papers generated from

research [..] funded by NRF, should deposit their fjnal peer-reviewed manuscripts […] to the administering Instjtutjonal Repository with an embargo period of no more than 12 months.

Earlier Open Access may be provided should this be allowed by the

publisher.

If the paper is published in an Open Access journal [..this paper..] should

be deposited into the administering Instjtutjonal Repository and Open Access should be provided as soon as possible.

In additjon, the data supportjng the publicatjon should be deposited in an

accredited Open Access repository, with the provision of a Digital Object Identjfjer for future citatjon and referencing.

SLIDE 10

Social Aspects of an Open Commons

Openness is a social evolutjon, distributjng power and increasing transparency,

with many aspects:

Open Source
Makes it easier to share and contribute
Open Science
Making the scientjfjc method more transparent and reproducible
acknowledging contributjons from various places
Open Access
To content (access to knowledge)
To infrastructures (in order to generate content)
Open Standards
Ensure interoperability and avoid lock-in
Provide means to contribute to the development of the commons

SLIDE 11

How can this possibly work ?

We seem to be in a dilemma: this cannot be

achieved through entjrely top-down or botuom-up actjvitjes:

Is strict hierarchy (Tier1, Tier2, instjtutjonal, etc) the

wrong model ?

Implies rigidity, predictability, control
Perhaps the correct model is a fmatuer, open,

complex-adaptjve system with organising agents ?

Implies fmexibility, non-linearity, complex interactjons

SLIDE 12

Systems Thinking

ORDER ENABLES CREATIVITY

SLIDE 13

(Complex Adaptjve)

Systems Thinking

The number of elements is suffj

ffjciently large that conventjonal descriptjons (e.g. a system of difgerentjal equatjons) are not only impractjcal, but cease to assist in understanding the system. Moreover, the elements interact dynamically, and the interactjons can be physical or involve the exchange of informatjon

Such interactj

tjons are rich, i.e. any element or sub-system in the system is afgected by and afgects several

ther elements or sub-systems
The interactj

tjons are non-linear: small changes in inputs, physical interactjons or stjmuli can cause large efgects or very signifjcant changes in outputs

Interactjons are primarily but not exclusively with immediate neighbours and the nature of the infmuence is

modulated

Any interactj

tjon can feed back onto itself directly or afuer a number of intervening stages. Such feedback can vary in quality. This is known as recurrency

Such systems may be open and it may be diffj

ffjcult or impossible to defj fjne system boundaries

Complex systems operate under far from equilibrium conditjons. There has to be a constant fm

fmow of energy to maintain the organizatj tjon of the system

Complex systems have a history. They evolve and their past is co-responsible for their present behaviour
Elements in the system may be ignorant of the behaviour of the system as a whole, responding only

to the informatjon or physical stjmuli available to them locally

SLIDE 14

The Natjonal Integrated Cyberinfrastructure System

SLIDE 15

VISION: natjonal leadership in the provision of a comprehensive Cyber-Infrastructure, essentjal to 21st century advances for South African research, educatjon and innovatjon. MISSION: increase knowledge creatjon through provision of a natjonal platgorm of essentjal Cyber-Infrastructure.

15 15/03/31

PRINCIPLES

Joint planning and budgetjng
Good governance
Visibility of CI services
Sustainability
Constructjve stakeholder

engagement NICIS: TIER 1 Advanced Services Governance

Strategy Advisory Board

Senior Management User and Community Inputs

NICIS: A System for Cyberinfrastructure

SLIDE 16

+

NICIS Components: Network

SLIDE 17

NICIS Recommendatjon: CHPC in essentj tjally its current form should take on the role of the Computj tjng Services area, with some changes to its mandate.

NICIS Components: Compute

SLIDE 18

NICIS Components: Data

SLIDE 19

Nice kit ! ... What about the people ?

SLIDE 20

“Nous sommes tous de jeunes barbares que nos jouets neufs émerveillent encore.”

A. de Saint-Exupery

“Terre des Hommes”

Une Terre Pour Tous

htup:/ /aaroc.github.io/blog/2014/11/26/Terre-des-hackers

SLIDE 21

La verité pour l’un, fût de bâtjr, […] elle est, pour l’autre, de l’habiter.

Pour le colonial qui fonde un empire, le sens de la vie est de conquérir. Le soldat meprise le colon, mais le but de cetue conquete n’etait-il pas l’etablissement de ce colon ? […]

Le Colon et le Soldat

SLIDE 22

Towards a Commons: What's missing ?

Several central components have been put place for the creatjon
f an e-Infrastructure commons in South Africa:
Signifjcant computjng and data hardware investment
Network infrastructure connectjng almost everyone
Open Access Statement and policies on central funding
Culture stjll needs to catch up to technology (again)
A culture of appreciatjng the common good, of contributjng and

collectjve actjon is desirable

Small and niche contributjons can be just as impactgul as large, shiny
nes.
Are we ready to give up control and allow true partjcipatjon by all ?

SLIDE 23

Open Commons as Exchanges

Interoperable infrastructures
Standard interfaces to common services
Coordinatjon of OLAs and SLAs to provide coordinated operatjons
Independent Core Services to encourage 3rd party and community

contributjon

Monitoring and Accountjng
Identjty and Service Federatjon
Discoverability and re-usability (publishing, preserving and licensing)
Specialisatj

tjon and Skills

Re-Use : Develop “Building Blocks” and donate them to the commons
Reward and Recognitj

tjon : Make it worth the while to do things for the long-run, not the annual performance evaluatjon

SLIDE 24

‘Government Housing’ vs ‘Community Projects’

To answer the questjon “who owns the commons”, we need to

address constjtuency:

Access to e-Infrastructure (like other basic necessitjes such as

housing), could be provided:

via a central scientjfjc bureaucracy, according to natjonal standards
or a framework could be in place which allow communitjes to build

their “housing”.

Who gets to live in the housing ? Ie, who gets to use, develop,

and exploit the e-Infrastructure ?

To put it difgerently – what will we do with all the kit and tools

that WLCG, SKA, H3A etc are building ?

SLIDE 25

Implicatjons for … “Commoners”

Educators Researchers Developers Operators

htup:/ /brucellino.github.io/blog/2015/02/24/ECommonsStrategy/

SLIDE 26

Implicatjons: Teaching and Research

Teaching and Research in the same environment
Path to scale from university or group resources

to natjonal or regional infrastructure

Skills developed throughout, fed back into the

commons

The use of the commons becomes part of standard

practjce in teaching and research

The commons is assured of new ideas, skills and

stjmulus.

SLIDE 27

Fork my Science

Science faces a reproducibility crisis:

htup:/ /www.nature.com/nature/focus/reproducibility/

Scientjfjc output is the result of Data, Algorithms,

Applicatjons, Workfmows – and interpretatjon

Reproducible science:
Discoverable
Citable
Executable
Accessible

SLIDE 28

Fork my Infrastructure

Big science ofuen requires and builds it's own

infrastructure

LHC, SKA, CORDIS, H3ABioNet
Ofuen developed entjrely in parallel to “natjonal” infrastructure
Ofuen new techniques and technologies are invented

in these large projects

Are they shared, improved and supplemented with
ther players in the commons ?

SLIDE 29

Fork my Infrastructure

Operatjons, best practjces encoded
Open Infrastructure means Open Code
htups:/

/github.com/AAROC/DevOps

Everything = Code
You can test, validate, reproduce research

infrastructure

build build passing passing

SLIDE 30

Implicatjons for the economy

Where will all the Ph.D.s go ?
An Open Commons means expertjse can cross

boundaries

Technical (tools, applicatjons, methodologies)
Instjtutjonal (cross-cuttjng, collaboratjve actjvitjes)
Private and public enterprise using the same

infrastructure

standardisatjon around a common set of tools
Innovatjon of new services with-reusable components
People develop transferrable skills

Research Software Engineer Infrastructure Architect (DevOps)

SLIDE 31

Implicatjons for the South Africa

Is NICIS a Natj

tjonal e-Infrastructure Commons ?

No. It can become aspects of one, partjcularly bringing order to

the system

“Easiest” to start with Data and Identjty infrastructures
Hypothesis : There is no consensus regarding the need for or

development of a broader commons

SAGrid → SA e-Science Commons
An updated strategy to implement aspects of an Open Commons
Quietly re-tooling the existjng infrastructure and resources to

allow them to become a commons, engaging with “the web” to have more voices and contributjons to the system.

SLIDE 32

Implicatjons for the Region

Is there a Regional Commons ?
No. There is an embryonic regional infrastructure

(stjmulated by CHAIN-REDS)

The Africa-Arabia Regional Operatjons Centre can

catalyse

Sci-GaIA and TANDEM projects will be focussing on

aspects during the next 2 years

What's taking you so long ?
The lack of order and coordinatjon ← this we can fjx
The erosion of the commons ← this: not so easy to fjx

SLIDE 33

Final word

SLIDE 34

Conclusions: e-Science Commons

A Commons allows sharing, and community development
Specialisatjon and markets can emerge
Cyberinfrastructure is a complex undertaking
It needs to be addressed as a system
Placing the right constraints on it can make it work
Cyberinfrastructure is expensive – but also cross-cuttjng
Open exchanges create lively ecosystems
Cyberinfrastructure is public
Openness, accountability, transparency and constjtuency are

crucial to its success

SLIDE 35

Conclusions: South Africa

NICIS has the opportunity to stjmulate the adoptjon of a

commons in South Africa

The NREN can bring connectjvity to all, but cannot change

the culture – this is a societal issue.

Expose “commons” services to South Africans –
Data and document publishing (PIDs and repos)
Interoperable Infrastructure
Develop new services as common resources
Collaboratjve development platgorms
Executable Infrastructure

SLIDE 36

Thanks

Bruce Becker (SANREN|SAGrid|AAROC)
Twituer: @brusisceddu
Github: @brucellino
htup:/

/brucellino.github.io

Africa-Arabia Regional Operatjons Centre:
htup:/

/aaroc.github.io

Github: @AAROC