Page 220 - Big data - Concept and application for telecommunications
P. 220
5 Big data - Concept and application for telecommunications
The IoT platform is required to associate time stamps to collected IoT data if data have no associated time
stamps.
The IoT platform is recommended to have the ability to associate semantic description to collected IoT data
if there is no associated semantic description for the IoT data, assuming that the IoT platform can understand
the meaning of the IoT data (e.g., according to the associated data model).
7.5.3 Data storage
The IoT platform is required to support secure and high-volume data storage for the IoT data.
The IoT platform is also recommended to support secure and high-volume data storage for the IoT data with
semantic annotation.
7.5.4 Data transfer
The IoT platform is required to support buffer mechanisms for data transfer in order to reduce network load.
NOTE – If the network is busy, the IoT platform can buffer IoT data within a tolerable time before
transmission.
The IoT platform is recommended to have the ability to compress IoT data before their transfer in order to
reduce the transfer load.
7.5.5 Data analysis
The IoT platform is recommended to have the ability to perform some pre-defined universal data analysis
(e.g., universal data statistics and data mining algorithms) according to application requests and send the
data analysis results to the requesting applications.
The IoT platform, for its internal use, is also recommended to have the ability to perform data analysis for
classifying which service type the IoT data is associated with, so that the IoT platform can efficiently deal with
data collection, data access and data transfer according to the service type.
7.5.6 Data query
The IoT platform is required to support data query on the stored historical IoT data and raw IoT data.
The IoT platform is recommended to support data query on real-time processed results of collected IoT data.
The IoT platform is recommended to support semantic data query.
The IoT platform is required to support data query with access control enabling query scope restriction.
7.6 Requirements of the IoT application server for big data
7.6.1 Data collection
The IoT application server is required to have the ability to collect IoT data from devices, gateways, IoT
platforms and other IoT application servers.
7.6.2 Data pre-processing
The IoT application server is required to have the ability to pre-process IoT data. Examples of data pre-
processing include validation, cleaning of invalid and irrelevant IoT data, insertion of default values for
incomplete IoT data, compression of IoT data and format transformation of IoT data.
Data received by the IoT application server from devices, gateways, IoT platforms or other IoT application
servers may be invalid or irrelevant. The IoT application server is required to identify invalid and irrelevant
IoT data by certain methods (e.g., data correlation) and then remove them.
Required data fields of IoT data received by the IoT application server from devices, gateways, IoT platforms
or other IoT application servers may be incomplete. The IoT application server is required to supplement
missing data fields with default values.
212 Network and infrastructure