1 / 27

A Quality Certification Model for Grid Research Projects the ETICS feasibility Study

A Quality Certification Model for Grid Research Projects the ETICS feasibility Study. Adriano Rippa ( adriano.rippa@eng.it ) - Engineering Ingegneria Informatica s.p.a. On behalf of the ETICS Project. Summary. Introduction to QA concepts The starting point of the study

oki
Download Presentation

A Quality Certification Model for Grid Research Projects the ETICS feasibility Study

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. A Quality Certification Model for Grid Research Projectsthe ETICS feasibility Study Adriano Rippa (adriano.rippa@eng.it) - Engineering Ingegneria Informatica s.p.a. On behalf of the ETICS Project

  2. Summary • Introduction to QA concepts • The starting point of the study • The proposed Grid Quality Certification Model (GQCM) • Comparing GQCM and other QA standards • Timeline and publications • Conclusions ETICS 2nd Review - CERN 15/02/2008

  3. Quality Assurance refers to several concepts Could be removed. • Quality of the implementation process • High level steps of the software production cycle suggesting what the organization must do (not how) to have effective development processes that “may lead” to good software. • Quality of the requirements management • Correct collection/management of requirements and relation with the customer and stakeholders, to reduce the percentage of failures due to misinterpreted requirements. • Quality of the service • Performances and correctness of the service • Finally quality of the software… QUALITY Quality of Product Quality of Process CMM GQCM ITIL ISO – 900x ETICS 2nd Review - CERN 15/02/2008

  4. Some QA Terminology • According to ISO 9126 documentation we can define: • Measure: the number or category assigned to an attribute of an entity by making a measurement (sometimes used as synonymous of metric) • Metric: The defined method to measure an attribute and the scale • Measurement: The use of a metric to assign a value (which may be a number or category) from a scale to an attribute of an entity) ETICS 2nd Review - CERN 15/02/2008

  5. Quality is a matter of measure! • State-of-the-art provide hundreds of metrics • The most common • Cyclomatic complexity • Lines of Code • Function Points • Mean Time Between(to) failure • Bugs density • … • Other approaches (Goal Question Metric - GQM) promote user defined metrics • Anomalies distribution • Effort used to solve anomalies • Cost of not founded anomalies • … ETICS 2nd Review - CERN 15/02/2008

  6. Quality Standards • There are many standards to asses the quality of the processes of an organization • CMM • ISO family • ITIL • AQAP • But QA means initial investments and managing QA means devote resources to it! • Several studies show that lot of companies (e.g. many Small/Medium Enterprises) can’t afford the initial effort and don’t recognise the promised increase of value. • Only ~70 companies in the world are certified at level 5 • 50 of them are in India • Only 25% of the companies in the world are level 2 or above ETICS 2nd Review - CERN 15/02/2008

  7. QA in research project is missing! • Current quality assurance standards are useful but • They need lot of time to be applied. The organisation need to be structured and certified. What for the short-live consortia? • They provide only theoretical guidelines which need to be adapted and realised. What for the objectivity? • It’s hard to systematically verify goodness of results: managing tools needed • They need resources to be devoted to • People need training and certification needs inspections and time to be achieved ETICS 2nd Review - CERN 15/02/2008

  8. QA in grids: ourproposal • GQCM is amodelforquality assurance that is • fully automatable in measuring and verifying activities to reduce investments and management effort, • not subjective, to certify the object not the process nor the organization, • product oriented, not process oriented, • …easily adoptable within (Grid) Research Projects ETICS 2nd Review - CERN 15/02/2008

  9. GQCM: Preliminary Remarks (1/2) • GQCM is independent from the ETICS tools • Any GQCM implementation needs just “automation” • GQCM has been developed within a Grid project and to asses the quality of grid software research projects but it can be used for any software applications • GQCM has been rounded off according to the gotten feedbacks from expert people and potential users who confirmed the usefulness of the model to reduce costs and time spend and to have something of more practical. ETICS 2nd Review - CERN 15/02/2008

  10. GQCM: Preliminary Remarks (2/2) • GQCM has been developed according to several standards: • GQCM has been described according to several ISO standards (e.g. ISO/IEC 25000, 14598) • GQCM has been restructured according to ISO 25041 • Quality attributes has been named using the same terminology as ISO 9126 ETICS 2nd Review - CERN 15/02/2008

  11. GQCM: Structure • GQCM is structured in Evaluation Modules (EM). • The set of evaluation techniques are grouped in families. Every family is an EM • 5 EMs: • EM: Static analysis • EM: Coding style • EM: Structural testing • EM: Functional testing • EM: Standards compliance ETICS 2nd Review - CERN 15/02/2008

  12. Evaluation Modules 1/5 This and the next 4 slides could became 3 instead than 5 removing the applicability and the input susections • EM Static analysis • Quality characteristics: • Reliability – maturity • Maintainability – analysability • Maintainability – changeability • Maintainability – testability • Technique: • Static analysis of classes. Statistics on measures are used as predictor of quality characteristics. • Applicability: • Object oriented programming languages. • Input: • source code, each class of the code is analysed ETICS 2nd Review - CERN 15/02/2008

  13. Evaluation Modules 2/5 • EM Coding style • Quality characteristics: • Maintainability – analysability • Technique: • Static analysis of the source code. • Applicability: • Most programming language • Input: • source code ETICS 2nd Review - CERN 15/02/2008

  14. Evaluation Modules 3/5 • EM Structural testing • Quality characteristics: • Functionality – accuracy • Reliability – maturity • Technique: • Structural testing. The intention of this EM is to test specific classes that are identified by static measures as being statistically more likely to have many errors. • Applicability: • Object oriented programming languages • Input: • source code ETICS 2nd Review - CERN 15/02/2008

  15. Evaluation Modules 4/5 • EM Functional testing • Quality characteristics: • Functionality – accuracy • Functionality – interoperability • Reliability – maturity • Portability – adaptability • Portability - installability • Technique: • Functional testing • The purpose of this EM is both to check platform compliance and to check to functional abilities of the software product. • Applicability: • General • Input: • compiled code • user documentation ETICS 2nd Review - CERN 15/02/2008

  16. Evaluation Modules 5/5 • EM Standards compliance • Quality characteristics: • Functionality – standards compliance • Technique: • Functional testing • This EM has a good potential for automating the test. • Applicability: • Standard specific • Input: • compiled code ETICS 2nd Review - CERN 15/02/2008

  17. GQCM: Final Score • Final score should be provided according to the following schema. The items which should be available for the users are: • A table summarizing the results • A list of passed and non passed tests • All the important information as: • Software product (e.g. name, version, executable code, documentation..) • Platform (name, version, date) • Quality characteristics (name, evaluation result, evaluation module identification) • Standard compliance (for each standard: name, version, date) • Identification of evaluation report (organization, report number, date) • Identification of certification body (organization, contact information) • Certification data (dates, certification number) • Electronic signature of certification record ETICS 2nd Review - CERN 15/02/2008

  18. GQCM: Summarizing Table example G = Good M = Medium P = Poor ETICS 2nd Review - CERN 15/02/2008

  19. GQCM: FAQ • I can’t add any overhead to my project • This model (and the capability of automate) will reduce the effort in performing continuous build and test activities (e.g. coverage tests) on different releases • How much costs adopting it? • Nothing, the model will be discussed publicly and the final version will be released under open license (e.g. Creative Commons) • The ETICS framework is provided as a service running on a dedicated infrastructure, free of charge for Research Project • My organisation is certified ISO/CMMi so I…? • See next slides… ETICS 2nd Review - CERN 15/02/2008

  20. GQCM integration in an ISO 9126 certified organization • Facts: • ISO 9126 define quality attributes • GQCM support the measure and evaluation of quality metrics • Why are they compatible each other? • Many ISO9126 aspects can be evaluated by GQCM GQCM apply what ISO 9126 asks • Any tool implementing GQCM and can be integrated as supporting ISO9126 adoption • So • GQCM can be both an independent quality model and a “tool” to be easily compliant with ISO 9126 ETICS 2nd Review - CERN 15/02/2008

  21. GQCM integration in a CMMi certified organization • Facts • CMM is process oriented (it gives guidelines). • GQCM is product oriented. • Hypothesis: • CMM2 • asks to “control and value” • CMM3 • asks to “use internal standards and have a reference model” • CMM4 • asks to “have measured and controlled processes using quantitative and statistical techniques” • CMM5 • asks to have an “improvement process and quality based on measurement”. • Our Thesis • GQCM is not in contrast with CMMI and can be integrated in organization as a tool to support CMMI adoption ETICS 2nd Review - CERN 15/02/2008

  22. Current ETICS calculable metrics and GQCM ETICS 2nd Review - CERN 15/02/2008

  23. The timeline EGEE ’07 1-5 Oct Budapest (Hun) OGF 21 15-19 Oct Seattle (USA) OGF 20/EGEE UF 7-11 May Manchester (UK) Now EELA 3° Conference 3-5 Dec 2007 Catania (IT) OCTOBER DECEMBER FEBRUARY MAY QUALIPSO Conference 16-17 Jan 2008 Rome (IT) ECHOGRID/EUChinagrid Conference 24-25 April Beijing (CHINA) Belief Conference 25-28 June Rio de Janeiro (BRA) ETICS 2nd Review - CERN 15/02/2008

  24. GQCM: publications • 2 papers submitted and accepted • When/Where • EELA third conference • 3-5 December 2007 • Catania/Italy • QUALIPSO first conference • 16-17 January 2008 • Rome/Italy • A quality oriented conference ETICS 2nd Review - CERN 15/02/2008

  25. GQCM: feedbacks • The model has got good feedbacks and acknowledgements • People are asking for more information • Some issues about the name of the model (solved) • One issue about ISO9000 compliance (outside the scope?) • Requests for metrics about the process  outside the scope? ETICS 2nd Review - CERN 15/02/2008

  26. Conclusions • GQCM is a certification model • Automatable • Free • Not in contrast with classical standards • Ready to interact with classical standards • Not limited to the ETICS build and test tool • Not limited to research projects • Not limited to grid software • GQCM require less human effort to be used because it is almost fully automatable • ETICS tool is ready to implement GQCM ETICS 2nd Review - CERN 15/02/2008

  27. Q&A http://www.eu-etics.org ETICS 2nd Review - CERN 15/02/2008

More Related