Page 217 - Big data - Concept and application for telecommunications

Big data - Concept and application for telecommunications 5

NOTE – Data format transformation after data collection and data de-noising before the data analysis can be
seen as two typical examples of data pre-processing scenarios.

7.1.5 Data analysis
Data analysis is the IoT data operation related to the computing process on IoT data aiming to derive
additional information according to application requirements, e.g., for making decisions, predicting future
results, etc.
7.1.6 Data visualization

Data visualization is the IoT data operation related to the visual representation of IoT data.
7.1.7 Data query

Data query is the IoT data operation related to querying IoT data, including raw IoT data and processed IoT
data.

NOTE – Data query on historical IoT data and real-time IoT data can be seen as two typical examples of data
query scenarios.

7.2 Requirements of the device for big data

7.2.1 Data collection

The device is required to have enough precision to sense IoT data from monitored things in order to measure
and record the characteristics of the monitored things.

The device is recommended to mark the sensed IoT data with labels to indicate the types (related to the
monitored things' characteristics) of the sensed IoT data, so that other IoT components can understand the
types of the sensed IoT data (e.g., physiological data for healthcare).

7.2.2 Data pre-processing
The device is recommended to organize sensed IoT data according to pre-defined information models and
represent them according to the pre-defined formats.
The device is recommended to associate time stamps to the collected IoT data, in order to keep track of
chronology of sensed data.
The device is recommended to associate semantic description to the collected IoT data, so that other
technical components of the IoT, i.e., gateway, IoT platform and IoT application server, can understand the
semantics of the sensed IoT data.
NOTE 1 – Examples of associated semantic description may include provenance, location, etc.
The device is recommended to have the ability to clean IoT data, e.g., removing redundant IoT sensed data
in order, for example, to minimize the amount of transferred IoT data.
NOTE 2 – The specific data cleaning capabilities of a device depend on its computing resources.

7.2.3 Data storage
The device is recommended to store the sensed IoT data according to rules (including e.g., temporally or
permanently).

7.2.4 Data analysis
The device is recommended to make appropriate data analysis for sensed IoT data and operational status of
the device itself (e.g., battery status, transmission bandwidth) to help its decision making process, including
e.g., when to send out what signaling data to other IoT components.

Network and infrastructure 209