Page 75 - Big data - Concept and application for telecommunications

2 Big data - Concept and application for telecommunications 2

6.1 Challenges and benefits of big data exchange

The following challenges are important to be considered for big data exchange:
– various sources, types and formats of data: big data service providers have to handle diverse
aspects of data and data sources during data collection, storage and integration;
– schema-on-read: usually big data is stored in a raw format, but after data is discovered and
captured, it is transformed to fulfil application's requirements;
– unawareness of suitable data/Unconstrained usage of data: sometimes, the big data service
customer does not recognize what kinds of data are really needed caused by unconstraint usage of
data in big data ecosystem. Therefore, data provider or big data service provider should offer data
with their usage for big data service customer to choose the helpful data for resolving their problem.
Exchanging data in a big data ecosystem is expected to provide the following benefits:
– mitigation of silos in the ecosystem through a better sharing of high-variety data between involved
parties;
– monetization of data enabling better revenues to be made by parties from high volume of data
exchanged in the ecosystem;
– openness of the publicly available data contributing to human society and economic activities;
– facilitation of the appearance about new and effective business models;
– interconnection of valuable, high-variety, and high-volume data contributing more to human
society and economic activities.

6.2 General concepts of big data exchange

Figure 6-1 illustrates the primitive model for a big data exchange. In this model, the following principles apply:
– a data source and a data target communicate with each other through a "data exchange". Through
this (illustrated by the black arrow in Figure 6-1), data is exchanged from the data source to the data
target. During this exchange of data, data processing may be performed;

– the data source is an entity which collects datasets (including raw data or already processed data)
or big data service's output and exports them to the data target;
– in the relationship, the data target is an entity which imports dataset (including raw data or
processed data) or big data service's output from the data source. Data export from the data source
is triggered either by the big data source itself or by an initiation request received from the data
target.

Figure 6-1 – Primitive model of big data exchange

Big data exchange patterns result from a composition of the primitive model Figure 6-2 illustrates two major
patterns:

– direct exchange pattern: a direct exchange of data from a peer data source to a peer data target
(see clause 7.1.1);
– intermediary exchange pattern: an indirect exchange of data through an intermediary control and
data processing agent (e.g., an intermediator) (see clause 7.1.2).

Moving data – Data exchange and data flow 67