Within the big data and cloud computing paradigms, traditional enterprise decision support systems (which are based on data warehouses and enterprise business intelligence) may not be as agile and as scalable as required. In this paper, the authors propose a service-oriented decision support system (SODSS) solution, where data, information, and analytics are modeled as reusable services in the cloud that can be dynamically composed for scalable, agile decision and process support.
The proposed SODSS consists of data as a service (DaaS), information as a service (IaaS), and analytics as a service (AaaS) platforms, in the cloud. The DaaS layer provides metadata management, customer data integration, and data access as services, regardless of where the data may be stored or how it is formatted. The IaaS layer provides data integration as a service, allowing a single coherent view by flexible integration of data from different sources across the business. These services include ad hoc querying, reporting, online analytical processing (OLAP), dashboards, content search, and data mashups. The AaaS component provides front-end analytic tools as services, such as optimization, data mining, and text mining.
Although AaaS can eliminate the need for physical data marts and can achieve economies of scale, it faces challenges such as the ability “to process very large amount[s] of structured and unstructured data [in the cloud] in a very short time to produce accurate and actionable results.” Cloud-based analytics should consider in-house enterprise data as well as real-time events. The current industry trend is to move to this next generation of big data analytics and decision support systems in the cloud. Thus, cloud-based decision support system architects, data scientists, enterprise IT project managers, and business managers who plan to migrate core business analytics to the cloud may find this paper useful.