Networking for Big Data and Laboratory
Professors: Paolo Di Lorenzo, Andrea Baiocchi
Seminars: Antonio Cianfrani
Objectives: The first aim of the Networking for Big Data and Laboratory course is to provide students the principles of Internet networking, illustrating also the basics of digital communications and signal processing. The course focuses on recent advances in networking protocols to efficiently support Distributed Data Centers infrastructures. Finally, Data Center architecture is presented and two major issues are investiaged with some detail, namely scheduling and congestion control.
Prerequisite: Understanding of programming logic. Basics of probability and statistics, matrix algebra, and calculus.
Final Exam: Written exam + Oral exam
Written exam on the first two parts of the course. Multiple choice and open-ended questions. 0 : 22 points
Oral exam on the third part of the course. Typically, two-three questions. 0 : 11 points
Final grade is the sum of the two grades. 30 e lode requires achieving at least 32 points.
Classroom code (2023-2024): 2ywbqbz
Lectures Time (a.y. 2023-2024)
Monday 15:00 - 17:00 (classroom B2 - building RM102 - via Ariosto 25)
Tuesday 15:00 - 19:00 (classroom 7 - building CU033 - piazzale Aldo Moro 5)
Friday 13:00 - 15:00 (classroom B2 - building RM102 - via Ariosto 25)
Office Hours
Prof. Paolo Di Lorenzo - Office hours will take place on Thursday from 3 pm to 4 pm. It is possible to fix videocalls in different days sending an email to paolo.dilorenzo@uniroma1.it
Prof. Andrea Baiocchi - Office hours will take place on Monday morning (from 11:00 to 12:00). It is possible to arrange dedicated Skype calls sending an email to andrea.baiocchi@uniroma1.it
Contents
Part 1 - Networking Fundamentals (Prof. Paolo Di Lorenzo, Prof. Antonio Cianfrani)
TCP/IP protocol stack
Transport layer: TCP and UDP
IP layer: addressing and routing
Link layer: forwarding
References: [1], [2]
Part 2 - Signal Processing and Digital Communications (Prof. Paolo Di Lorenzo)
Fundamentals of signal processing
Spectral analysis of signals and images [1], [3 - Chapters 2, 3, 4], [4, Chapter 7.3]
Linear Processing, filters, sampling and interpolation [1], [3 - Chapters 5, 9]
Fundamentals of digital communications:
Source coding: Signal and image compression, transform coding, JPEG [1], [4, Chapter 8.2]
Digital modulations: PAM, PSK, QAM. AWGN channel and MAP receiver [1], [3 - Chapter 12]
References: [1], [3], [4]
Part 3 - Data Centers (Prof. Andrea Baiocchi)
Outline of cloud computing.
Data centers architectures, topologies, addressing, routing.
Job scheduling and load balancing.
Congestion control (QCN, DCTCP).
References: [1], [5]-[9]
Textbooks and resources:
[2] James F. Kurose and Keith Ross, “Computer Networking: A Top-Down Approach Featuring the Internet”, Pearson Ed.
[3] Prandoni, P., & Vetterli, M. (2008). Signal processing for communications. EPFL press.
[4] Gonzales, Rafael C., and Paul Wintz. Digital image processing. Addison-Wesley Longman Publishing Co., Inc., 1987.
[5] Liu, Y., Muppala, J. K., Veeraraghavan, M., Lin, D., & Hamdi, M. (2013). Data center networks: Topologies, architectures and fault-tolerance characteristics. Springer Science & Business Media.
[6] Da Fonseca, Nelson LS, and Raouf Boutaba, eds. Cloud services, networking, and management. John Wiley & Sons, 2015.
[7] Zhou, X., Wu, F., Tan, J., Sun, Y., & Shroff, N. (2017). Designing low-complexity heavy-traffic delay-optimal load balancing schemes: Theory to algorithms. Proceedings of the ACM on Measurement and Analysis of Computing Systems, 1(2), 1-30.
[8] Alizadeh, M., Atikoglu, B., Kabbani, A., Lakshmikantha, A., Pan, R., Prabhakar, B., & Seaman, M. (2008, September). Data center transport mechanisms: Congestion control theory and IEEE standardization. In 2008 46th Annual Allerton Conference on Communication, Control, and Computing (pp. 1270-1277). IEEE.
[9] Alizadeh, M., Greenberg, A., Maltz, D. A., Padhye, J., Patel, P., Prabhakar, B., ... & Sridharan, M. (2010, August). Data center tcp (dctcp). In Proceedings of the ACM SIGCOMM 2010 Conference (pp. 63-74).