Page 328 - Big data - Concept and application for telecommunications

5 Big data - Concept and application for telecommunications

In the data extraction stage, it is required to extract data from heterogeneous sources, using professional
tools or application programming interfaces (APIs).

In the data transformation stage, it is required to judge the validity and integrity of data, as well as error
correction aiming at certain fields, standardize data structures according to the demands of services and
merge and summarize data in records and files.
In the data load stage, the data which has been pre-processed is loaded to target data warehouses.

7.5 Requirements for unified traffic management items

When the collected data has been pre-processed and stored in a big data platform, it is required to analyse
a set of items for unified traffic management. There are two computing patterns; real-time steam processing
and off-line batch processing. Based on the result data of these computing patterns, we can sense the quality
of service (QoS) in the mobile network.

7.5.1 Real-time stream processing items
Real-time stream processing data items are required to compute all of the time for multiple dimensions for
as long as the data from the mobile network is collected.
It is required, to but not limited to, process items as follows:

– Number of users
It is required to calculate the number of users registered to the mobile network and the number of users
using a service in real time for multiple dimensions, such as numbers in different cells and in different base
stations, numbers in different network types, numbers in different administrative regions and numbers in
different network types.
– Traffic

It is required to calculate the total traffic, uplink traffic and downlink traffic in a mobile network in real-time
for multiple dimensions, such as traffic in different cells, traffic in different base stations, traffic in different
administrative regions, traffic in different network types, traffic of different applications and traffic of
different users.

– Internet speed
It is required to calculate Internet speed in a mobile network in real-time for different dimensions, such as
Internet speed of different cells, speed of different base stations, Internet speed of different administrative
regions, Internet speed of different network types, speed of different applications and speed of different
users.

7.5.2 Off-line batch processing items
Off-line batch processing items are required to analyse for multiple dimensions at regular intervals, such as
daily, weekly, monthly and annually.
It is required to, but not limited to, process items as follows:

– Average number of users
It is required to analyse the average number of users in a mobile network offline at different intervals.
i) Average number of users registered to the mobile network
It is required to calculate the average number of mobile network registered users at different
intervals in different dimensions, such as the average number of network registered users in
different cells, the average number of network registered users in different base stations, the
average number of network registered users in different administrative regions, the average
number of network registered users of different network types and the average number of network
registered users of different applications.
ii) Average number of services being used

320 Network and infrastructure