Mobile application data stream detection in application aware networks
Johansson, Axel (2019-06-20)
Mobile application data stream detection in application aware networks
(20.06.2019)
Julkaisu on tekijänoikeussäännösten alainen. Teosta voi lukea ja tulostaa henkilökohtaista käyttöä varten. Käyttö kaupallisiin tarkoituksiin on kielletty.
suljettu
Julkaisun pysyvä osoite on:
https://urn.fi/URN:NBN:fi-fe2019092329505
https://urn.fi/URN:NBN:fi-fe2019092329505
Tiivistelmä
With the huge expectations for the upcoming 5th generation cellular network technology
(5G) and with the ambition towards highly connected societies, the demands for future
networks are ever-increasing. 5G is an umbrella term for upcoming network technologies
coming in 2020’s. That includes new antenna technologies and introduces new band-
widths. One of the most intriguing and revolutionary technologies is network slicing.
The concept is to slice down mobile broadband networks to create number of virtual
mobile networks with optimised profiling for each usage scenario.
For personal mobile devices alone, there is need for sophisticated network slicing.
Mobile devices rely on the network very differently depending on what application is in
use. E.g. a video streaming services rely heavily on downstream bandwidth but do not
care much about upstream bandwidth or latency. Mobile games, on the other hand need
low latency in both upstream and downstream but not much bandwidth. Map services
and navigation applications poll the service frequently but do not require continuous data
stream in either direction and health monitoring applications use only little background
data infrequently. Mobile devices can change quickly from application to another, which
requires the slices to be dynamically generated and allocated.
This thesis is about network slicing in application aware networks and implementation of
traffic inspector for Android mobile devices to monitor network behaviour. The collected
network data is refined to indicate networks stream sizes for each process running on
the end user equipment. Based on the analysis of this data, feasible network profiles are
introduced.
The results of this study suggest that network efficiency could be improved by
application-based network slicing for mobile services. The individual data streams for
each process are distinct. Additionally, it was found that in different applications, these
data streams have their own characteristics. Based on these findings, the feasibility of
application aware network slicing for these applications seems probable and it could
result in improvement in network efficiency.
(5G) and with the ambition towards highly connected societies, the demands for future
networks are ever-increasing. 5G is an umbrella term for upcoming network technologies
coming in 2020’s. That includes new antenna technologies and introduces new band-
widths. One of the most intriguing and revolutionary technologies is network slicing.
The concept is to slice down mobile broadband networks to create number of virtual
mobile networks with optimised profiling for each usage scenario.
For personal mobile devices alone, there is need for sophisticated network slicing.
Mobile devices rely on the network very differently depending on what application is in
use. E.g. a video streaming services rely heavily on downstream bandwidth but do not
care much about upstream bandwidth or latency. Mobile games, on the other hand need
low latency in both upstream and downstream but not much bandwidth. Map services
and navigation applications poll the service frequently but do not require continuous data
stream in either direction and health monitoring applications use only little background
data infrequently. Mobile devices can change quickly from application to another, which
requires the slices to be dynamically generated and allocated.
This thesis is about network slicing in application aware networks and implementation of
traffic inspector for Android mobile devices to monitor network behaviour. The collected
network data is refined to indicate networks stream sizes for each process running on
the end user equipment. Based on the analysis of this data, feasible network profiles are
introduced.
The results of this study suggest that network efficiency could be improved by
application-based network slicing for mobile services. The individual data streams for
each process are distinct. Additionally, it was found that in different applications, these
data streams have their own characteristics. Based on these findings, the feasibility of
application aware network slicing for these applications seems probable and it could
result in improvement in network efficiency.