The TOREADOR project is aimed at overcoming some major hurdles that until now have prevented many European companies from reaping the full benefits of Big Data analytics (BDA). Many companies and organisations in Europe have become aware of the potential competitive advantage they could get by timely and accurate Big Data analytics, but lack the IT expertise and budget to fully exploit BDA.
To overcome this hurdle, TOREADOR takes a model-based BDA-as-a-service (MBDAaaS) approach, providing models of the entire Big Data analysis process and of its artefacts. TOREADOR open, suitable-for-standardisation models will support substantial automation and commoditisation of Big Data analytics, while enabling it to be easily tailored to domain-specific customer requirements.
Besides models for representing all aspects of BDA, TOREADOR will deliver an architectural framework and a set of components for model-driven set-up and management of Big Data analytics processes. Once TOREADOR MBDAaaS will become widespread, price competition on Big Data services will ensue, driving costs of Big Data analytics well within reach of EU organizations (including SMEs) that do not have either in-house Big Data expertise or budget for expensive data consultancy.
Activities supported and largely automatised by TOREADOR will include (i) planning Big Data sources preparation (ii) negotiating machine-readable Service Level Agreements (SLAs) for BDA detailing privacy, timing, and accuracy needs (iii) choosing data management and algorithm parallelisation strategies (iv) ensuring auditing and assessment of legal compliance (for example, to privacy regulations) of BDA enactment.
From the data management point of view, the TOREADOR framework will address automatically all major problems of on-demand data preparation, including handling Big Data opacity, diversity, security, and privacy compliance. From the computational point of view, the TOREADOR framework will support abstract modelling of the BDA life cycle from distributed data acquisition/storage to the design and parallel deployment of analytics and presentation of results.
Within TOREADOR, model transformations will be used to support the transition from declarative models expressing BDA goals to procedural models describing how analytics should be parallelized and executed. The latter will be seamlessly mapped into the existing computational facilities held by TOREADOR consortium members, enabling the major IT integrators and providers in the TOREADOR consortium to deliver turn-key MBDaaS services to their customers. Also, new analytics techniques will be specified, where needed, to increase the impact of TOREADOR and make TOREADOR approach suitable for a wider set of customers and domains.
TOREADOR framework will be put to work in four pilots, covering key areas of the EU marketplace that until now have been only marginally involved in BDA: (i) on-demand security analysis of application data, (ii) on-demand analysis of energy production and delivery, (iii) aerospace products manufacturing/maintenance and (iv) clickstream analysis of Web e-commerce applications. Each pilot is based on huge data sources made available by TOREADOR user partners, who will directly and measurably benefit from integrating MBDAaaS in their decision making.
From the innovation point of view, we expect a big innovation boost when TOREADOR models will be made available to third parties wishing to develop new BDAs in accordance to emerging market opportunities. Additional exploitation opportunities will also come from adding new services to the TOREADOR framework, like adding support for immersive or augmented reality reporting, interfacing with external open data sources and with customers’ on-premises systems and sensors. To take advantage of these opportunities the TOREADOR project will strongly support early start-up creation by internal and external stakeholders, and run yearly calls for innovation ideas based on TOREADOR vision.