ENEE 769N: CONTROL PROBLEMS
IN MODERN TELECOMMUNICATION NETWORKS
Modern telecommunication networks are infiltrating every aspect of life. Modern telecommunication networks are implemented using different physical layers: wireless, wireline-copper, wireline-optical fiber, wireline-cable, satellite and various combinations of these media. Modern telecommunication networks carry a mixture of information strings like voice, music, video, images, data etc. Modern telecommunication networks serve a great diversity of customers with great variability in their needs and habits. The Internet and the Web have created a need for re-examining the fundamental models, laws and designs that have been used for telecommunication networks for many years based principally on the needs for telephony.
As current and future networks increase in their size and diversity, a key technical area that emerges at the heart of many problems, and is perhaps the most challenging and promising, is that of control and management. In this course we treat management technically as part of control, since management is the static or dynamic allocation and control of network resources. We address several advanced topics in network management and control tuned to the needs of current and future Internet-like networks. The technical term we use for the latter is heterogeneous internetworks. We cover topics from the following areas: traffic modeling, traffic measurements, flow control, congestion control, queuing control, resource allocation, dynamic bandwidth management, dynamic caching, probing and channel estimation, interference mitigation in wireless, capacity increase using advanced modulations, adaptive dynamic routing, management information bases, knowledge acquisition and learning in networks, automated fault and network management. The unifying techniques come from control systems and optimization. This is indeed the central point of the course: that most of the current and future problems in networks can be best formulated and solved within the framework of controlled dynamical systems. We close the course with topics from complex and dynamical systems, which leads to the analysis of self-configuration and automated self-control of communication networks.
We will assume knowledge of elements of system theory (from ENEE 663 or the equivalent), optimization and control (from ENEE 664 or the equivalent), probability and estimation (from ENEE 620 or the equivalent), elements of digital communication systems (from ENEE 623 or the equivalent), knowledge of telecommunication networks at the level of a senior undergraduate or first year course in networks (from ENEE 426, or elements from ENEE 625, or the equivalent). The courses in parentheses are given so as to clearly define the material.
Student background is assumed and expected to be very diverse. Background will be built as appropriate from reading assignments.
This is an advanced topics course suitable for second (and higher) year graduate students from ECE, CS, BMGT, MATH. It is targeted for students searching for MS or PhD theses topics or already working on their MS and PhD theses in the area of telecommunication networks. It should also be of interest to students wishing to learn in a compact form and in a unified way some of the most exciting and challenging current research problems and techniques motivated by broadband internetworks, the Internet and the Web.
Students are assumed to have reached a mature self-learning style based on reading assignments, so that the lectures can concentrate on the most advanced and sophisticated topics.
Textbook and Class Notes:
There is no required textbook for the course.
We will use a variety of sources, principally from technical research articles, and chapters from recent research monographs. Additional notes and slide copies will be provided by the instructor. Copies of all materials will be distributed electronically or via hardcopy.
A class library will be set up electronically as well as a class web page. Extensive links to other related URLs will be provided.
All network models, programming software and testbeds available either in the Systems Engineering and Integration Laboratory or the Hybrid Networks Laboratory will be made available to class students for project work
- Traffic models and measurements, conventional, self-similar, fractal, multi-fractal
- Flow control algorithms, versions of TCP and beyond, analysis and performance
- Congestion control schemes, control formulations, comparisons
- Wireless channels, models, estimation, probing, interference mitigation, control perspectives
- Capacities and advanced modulations, unified schemes for compression/coding/modulation, the future
- Queuing control, conventional models, non-conventional models, large deviations
- Dynamic resource allocation problems, bandwidth allocation, effective bandwidth, large deviations
- Dynamic adaptive routing, hierarchical routing, selected topics
- Media access protocols, selected topics
- Network management, a control perspective, learning and information storage, database techniques
- Topics in fault management, pattern learning and storage, automation
- Network security, intrusions, learning and classifying intrusion patterns, detection and reaction: a control systems perspective, dynamic games
- Information distribution over the Internet, Web servers characteristics, Web user characteristics, dynamic caching, the Web as a dynamical system, global problems
- Networks as autonomous complex dynamical systems, patterns, new network laws, self-organization, self-control
- Network economics, costs, service pricing, a dynamics and control perspective
Homework and Exams:
- There will be about 8 homework assignments. I hope to make them challenging. The homework will constitute an essential part of the course. Most homework problems will be mini-research projects (partly solved already).
- There will be several tutorial reviews of the material.
- There will be a take home mid-term exam (at a time about 2/3 of the duration of the class).
- A class term paper; by each student individually.
Final grade in the course will be arrived approximately with the following scheme:
25% for the take home mid-term exam,
45% for the class term paper.
The final grade will be a representation of total performance with appropriate adjustments made due to a mishap or exceptional performance in one of the graded efforts