Big data architecture evolution: 2014 and beyond

Reviewer: De Wang

Big data architecture is rapidly evolving due to the recent dramatic growth of information. People search for useful information and at the same time generate more information. Efficiently retrieving information and effectively managing it becomes a big challenge. We need a comprehensive, high-performance architecture to support big data. This paper investigates the evolution of big data architecture in terms of big data governance and a data ingestion strategy. Furthermore, it lists the challenges of big data, including security and privacy, usability, high performance, information management, and business models. The authors provide a solution to those challenges with big data business intelligence management (BDBIM). In the end, Mohammad et al. conclude that the environment of an organization affects how the end results will be produced. They also mention the central concept of big data architecture: “data is either streaming in or some [extract, transform, and load, ETL] processes are in progress with an organizational environment with which they have some sort of relationship.” Overall, the paper points out the problems in the current architecture and the trend of big data architecture. I agree that this area still needs more effort from the research community. Online Computing Reviews Service

Reviewer: Salvatore F. Pileggi

This paper discusses big data in a wide context, including the development of effective and sustainable architectures to support analytics, as well as cloud services on a large scale and business intelligence. The enormous expectations around big data point out a great number of challenges that require novel approaches and solutions to efficiently address significant concerns from most stakeholders involved in common business processes. The most interesting aspect of this work is the topic itself. Big data has such momentum to make any perspective of analysis and discussion definitely welcome. Indeed, this work looks well suited in the current literature. On the other hand, the authors probably try to address and discuss too many (complex) topics in a short document. Consequently, the reader feels a general lack of depth. That feeling is further reinforced by more than one questionable issue related to the presentation. First of all, the introductory part is composed of a couple of examples, while a common extension of the topics introduced in the abstract would likely have been much more effective. In addition, while the figures probably deal with key concepts, they are weakly explained or sometimes never even cited in the paper. A similar problem is related to the list of challenges that objectively needs some structure and some abstraction. Despite the interesting focus and premise, I don't feel that this paper catches the current state of the art from the intended perspective. Online Computing Reviews Service