Cloud Computing has become an important paradigm with most enterprises either having moved or having intentions to move to this architecture. The advantages of virtualised access to elastic resources combined with a ‘pay as you go’ economic model is attractive both for outsourced Cloud requirements and for within-enterprise management control of ICT. Most of the concepts of Cloud computing are not novel but utilising these concepts together in novel ways is causing one of the greatest revolutions in ICT.

However, many challenges remain. Public Cloud providers encourage proprietary lock-in; the internal Cloud architectures are not yet mature; the utilisation of a SOA (Service Oriented Architecture) paradigm in a Cloud environment is not fully worked out with many variants and appropriate programming languages providing declarative specification of intent while allowing subtle control of dynamic resource allocation are lacking.

Above all there is the challenge of quality: quality of the service provision (quality of service, service level agreement management); quality in modelling business requirements in a Cloud environment; quality in the languages used to express those requirements and to execute them; quality in the services in both what they provide and how they provide it; quality in service composition including dynamicity; quality in scheduling execution including partitioning and parallelism; quality in monitoring execution and triggering actions to maintain quality of service. There is also the quality of business continuity provision and the quality of virtualisation – including seamless execution transfers to assure timeliness and the quality of the Cloud infrastructure and platform provision in terms of ‘green ICT’. Unlike other areas of ICT there is not yet for Clouds a quality of conformance since there is no agreed standard (unlike, for example, WWW).

Quality is related to the satisfaction of functional and non-functional requirements. Key questions include how to define all these aspects of quality, how to measure them, how to monitor their value during the whole process from requirements to execution completion and how to utilise the information generated to improve quality end-to-end.