Dynamic Resource Management On Top of SDN and NFV platforms by applying Machine Learning
- Abstract
- The upcoming 5th generation network is not only restricted to mobile devices and high
bandwidth provision but also will support multiple services with their dynamic resource demands
while enabling ultra-reliability and low latency. As a matter of fact, this increased in the domain
made it very complex to design, configure and control the next generation platforms.
Considering the fact, SDN (Software Defined Networking) and NFV (Network Function
Virtualization) enabled programmability and elasticity in network infrastructure and also
provided freedom from the dedicated hardware was a major achievement which broke the
monopoly of the major networking industry. They are capable to accomplish the dynamic
demands of upcoming applications through the programmable capabilities and network slicing
i.e. provisioning of multiple virtual resources over a single physical infrastructure. The SDN and
NFV based platforms enable slicing and optimized resource provisioning but they still require
complex configurations for setting up each slice and it requires several proficient experts for any
additional updates. Also, the 5G support of many services resulted in a lot of dynamicity in terms
of user demands e.g. number of cars in C-V2X and also in events like Olympic games the traffic
stream through the network can raise exponentially and requiring the network to provide tons of
resources. Currently, the resource management is done by experts which are not very efficient
and is prone to errors requiring a replacement. Considering the complexity of 5G and its
requirements manuscript aims to automate configuration management using IBN (Intent-based
Networking) application and apply ML(Machine Learning) to manage the next generation
platforms for the automatic scaling of resources in accordance with dynamic user demands. The
IBN application will simplify the configuration procedure by enabling what to achieve (highlevel
generic instructions) as input and how to achieve (Policies and Configuration) as its output.
On the other hand, the ML with a monitoring integration proficiently perceive the current state of
the system resources and predicts future utilization to suggest updates in network configuration.
- Author(s)
- Khan Talha Ahmed
- Issued Date
- 2019
- Awarded Date
- 2019. 8
- Type
- Dissertation
- URI
- http://dcoll.jejunu.ac.kr/common/orgView/000000009126
- Affiliation
- 제주대학교 대학원
- Department
- 대학원 컴퓨터공학과
- Advisor
- Ahn, khi Jung
- Table Of Contents
- CONTENTS
ACKNOWLEDGMENTS iv
ABSTRACT vi
CONTENTS viii
I. List of Acronyms xi
II. Chapter 1 1
Introduction 1
III. Chapter 2 6
Related Work 6
IV. Chapter 3 11
The overall system, IBN-Application, and M-CORD platform 11
3.1 Overall system in view point of SDN and NFV platforms: 11
3.2 IBN-Application 13
3. 3 M-CORD with it Components 22
3. 3 M-CORD based Network Slicing Scenario 31
V. Chapter 4 35
Machine Learning Approach and Experimental Results . 35
VI. Chapter 5 59
Conclusions 59
Bibliography . 60
- Degree
- Master
- Publisher
- 제주대학교 대학원
- Citation
- Khan Talha Ahmed. (2019). Dynamic Resource Management On Top of SDN and NFV platforms by applying Machine Learning
-
Appears in Collections:
- General Graduate School > Computer Engineering
- 공개 및 라이선스
-
- 파일 목록
-
Items in Repository are protected by copyright, with all rights reserved, unless otherwise indicated.