The Network Infrastructures for Big Data Analytics

The Network Infrastructures for Big Data Analytics

DOI: 10.4018/978-1-4666-5864-6.ch007
(Individual Chapters)
No Current Special Offers


The most delectable factor here is that the stability and maturity of networking and communication technologies enable the seamless and spontaneous interconnectivity of diverse and distributed consumer electronics, electrical, mechanical, and manufacturing devices at ground level and a bevy of services (Web, enterprise, cloud, embedded, analytical, etc.) at cyber level. Any tangible artefact and article gets connected with another to get the right and relevant empowerment, which in turn facilitates more data generation and transmission. Regulated interactions amongst digitalized entities have put a stimulating foundation for hitherto unforeseen and creative new capabilities and competencies. In short, data has grandly acquired the status of an asset not only in business organizations but also in personal lives, and hence, the data gathering, storage, and leverage tasks are fast-growing. With the data explosion happening feverishly, the discipline of big data computing and analytics has become a much-discoursed and deliberated domain of study and research. In this chapter, the authors discuss the emerging and evolving network infrastructures and architectures for big data analytics.
Chapter Preview


In the days that have passed by, worldwide businesses have mostly focussed on productivity-related activities such as faster transactions, higher throughput, etc. with the limited set of business-centric data (transaction, operations, sales and marketing, customer, product, region, etc.). Nowadays, the quantity of data getting generated, captured, and capitalised is growing up exponentially and expediently. The data sources are also varying and many. With the Internet being established as the world’s largest information superhighway and open, public, and affordable communication infrastructure, the focus is on faster delivery of data-driven insights to executives, and other decision-makers for ensuring customer delight, taking informed decisions, embarking on infrastructure optimization, developing next-generation services and applications, to close down the gap between business and IT, etc. The real challenge for organizations in the forthcoming era of knowledge will be to find proven ways to better collect, analyze, monetize, and capitalize on all data heaps emanating from different and distributed sources in multiple formats. In a nutshell, big data computing represents a growing collection of robust and resilient technologies, methodologies, architectures, and tools to economically and elegantly extract value from very large volumes of a wide variety of data by enabling high-velocity capture, smart processing, and cognitive analysis towards real-time knowledge discovery and dissemination. It is expected that organizations that are best able to make real-time business decisions using derived intelligence will gain a distinct competitive advantage over those that are unable to embrace it appropriately.

Big data analytics however mandates newer software frameworks, versatile platforms, and optimized infrastructures. It requires new system designs, administrative skillsets and data management capabilities. The most pragmatic approach for most organizations is to jump-start their efforts with smaller deployments built on existing IT resources to gain the much-needed confidence and clarity. As the scope and business value of the big data discipline in transitioning enterprises to be smarter in their deeds and decisions, the importance of innovative networking technologies, topologies, and tools cannot be taken lightly. This chapter is allocated for expounding network architectures, infrastructures, platforms and practices for simplifying and streamlining big data analytics. One area where big data will have a direct impact on enterprise networks is in the area of network intelligence. The ability to automatically reconfigure the network for changing network loads and failed links (requiring zero administration for the addition of switch infrastructure) makes the network more agile. The result is a significantly reduced administration burden on the operations team, which minimizes the risk of errors and improves resiliency. The aspects on which big data differ from traditional data are shown in Table 1.

Table 1.
Component wise comparison between traditional data and big data
ComponentsTraditional DataBig Data
Data VolumeTerabytesPetabytes, Exabytes and to Zettabytes
Data TypesStructuredMulti-Structured
Data AffinityKnown RelationshipsUnknown & Complex Relationships
Data ModelSchema-centricSchema-less

Complete Chapter List

Search this Book: