T +31 20 314 0950 info@sig.eu www.sig.eu

Green Software and Sustainability Joost Visser, Software Improvement Group Knowledge Network Green Software #KNGS

January 18, 2011

I 11

Software Development 1964 – resource aware

2

© 2012 Software Improvement Group

I 11

Software development Now – resource agnostic

3

© 2012 Software Improvement Group

•  Large scale. Global distribution. Failure rates estimated at 30%. •  Graphical development environments perform continuous compilation. •  Nightly regression testing. Test environments duplicate production.

I 11 4

© 2012 Software Improvement Group

“The most strategic aspect of energy efficient computing will be the evolution of application software to facilitate system-wide energy efficiency.” Communications of the ACM , March 2010, vol. 53, No. 3 Towards Energy-Efficient Computing by David J. Brown (Sun Microsystems) and Charles Reams (Cambridge University)

I 11

Power Loss Chain Extended version …

5

© 2012 Software Improvement Group

75% 25% Computational efficiency

Functional necessity

35% 65% Useful work Frivolous features

Optimal cycles and storage Algorithmic inefficiency

Percentages are indicative only

Image source: British Computer Society Data Centre Specialist Group Performing the right work?

Performing the work right?

I 11 6

© 2012 Software Improvement Group

Application-level energy-efficiency

How?

design build test require operate

Estimate energy consequences of requirements Eliminate frivolous requirements

Estimate energy consequences of design decisions Consider less wasteful alternatives

Monitor energy consumption Provide feedback to development

Energy-ware coding of algorithms, data structures, communication, concurrency.

High-quality coding to minimize build, test, re-work, and maintenance effort

I 10

Taxonomy

7

© 2012 Software Improvement Group

Green IT

Greening by IT

Greening of IT

Green Grid

Green Data Center

Green Hardware

Green Software

Functional necessity

Computational efficiency

Algorithmic Data structures Protocols

Data retention Responsiveness Graphical presentation

Green Software Development Rework

Maintainability Testability

Environment

I 10

Knowledge Network Green Software kngs.wikidot.com

8

© 2012 Software Improvement Group

Network Knowledge

… …

Sharing knowledge Creating awareness Joint research

I 11 9

© 2012 Software Improvement Group

Sustainable Application Scan Example knowledge sharing KNGS

Service developed by SIG •  Identify opportunities for energy savings

by review of architecture, coding, deployment

Case “Rabo Rekening Delen” •  Application developed by Rabobank, Ordina, Service2Media

•  Opportunities for saving energy •  Removing proxy server to reduce data translation steps and

communication volume may lead to 30% savings •  Reducing image size in mobile browser application may lead

to 10% savings •  Eliminating some synchronization moments to reduce radio

traffic may lead to 5% savings

I 11 10

© 2012 Software Improvement Group

Software Energy Footprint Lab Example knowledge sharing KNGS

Lab created by Hogeschool van Amsterdam (CleanTech) •  Controlled measurement of energy profiles

of application software •  First students sponsored by SIG •  Hardware donations by e.g. Schuberg Philis

Questions •  How do different database management systems

compare in terms of energy footprint? •  How do different programming languages compare? •  How do different browsers, word processors compare? •  How do different data formats and communication protocols compare? •  How accurate are software energy profiling tools?

…

I 11 11

© 2012 Software Improvement Group

Software Energy Label The • on the horizon

Software Energy Label •  Standard criteria •  Meaningful rating •  Certification •  Allow consumers / procurement officers to recognize Green Software

Challenges •  What should be the target of evaluation?

application – specific release – specific installation – development and test process •  What characteristics should be captured?

energy consumption – scalability – co-existence •  On what indicators can software applications be compared?

consumption per user – per transaction – per feature •  How can acceptable thresholds for these indicators be determined?

I 11 12

© 2012 Software Improvement Group

Dr. ir. Joost Visser j.visser@sig.eu @jstvssr www.sig.eu +31 20 314 0950

Knowledge Network Green Software @KNGreenSoftware #KNGS kngs.wikidot.com