- Knowledge of cloud computing and virtualization (computing, memory and network) - Knowledge of the theory of multilayer, recursively factorized Clos topologies for data centers - Knowledge of different virtualization techniques (virtual machines, containers) - Ability to design the topology of the interconnection network of large data centers - Ability to operate Docker containers - Ability to design the IP addressing plan and to implement the routing in a data center - Ability to use network emulators (e.g., kathara.org) for data center networks - Ability to use Linux commands for advanced network configuration - Ability to run an AI application on hardware accelerators

- Main concepts of signal processing and computer networks - IP addressing and IP routing - Linux bash commands (useful)

The class is divided in two parts, the first (Part A) on cloud computing and design of data centers (4CFU) and the second (Part B) on digital architectures and advanced computing architectures (2CFU). Lectures topics and corresponding credits: A.1 Cloud computing (1CFU) - Service models. Web services. - Virtual machines and containers. Docker engine and container networking - Lab on cloud computing and Docker engine. A.2 Design of data centers (3CFU) - Clos-based design of data centers. Recursive construction - Optical interconnections and switching - Lab on data center network design - Software Defined Networking (SDN) and programmable data planes. P4 switches. - Lab on IP routing and on BGP protocol, based on FRRouting daemon. - Lab on BGP-based routing in a data center B.1 Basic digital hardware architectures (1CFU) - Von Neumann architectures and fundamental components of a computing system - Processing systems: pipelining and parallel processing, CPU and GPU architectures - The memory hierarchy and memory types: HDD, SSD, DRAM, SRAM, TCAM - Introduction to the interconnections and interconnections examples: USB, DDR, PCIe - Lab on system benchmarking: computation speed on CPU/GPU, memory transfer speed B.2 Advanced algorithm implementation (1CFU) - Mathematical model of Deep Neural Networks - Encoding of data, implementation of the Multiply-and-Accumulate computation paradigm - Evaluation of the complexity of a Deep Neural Network model - Compression of Deep Neural Network models: quantization, pruning - Lab on compression of Deep Neural Network models

The course provides all the theoretical concepts that are needed for the main laboratory and practical activities. The laboratories are mandatory and will be carried out in the classroom, using the student laptop and/or the PC in the laboratory. The students deliver weekly homeworks, which are mostly evaluated through a peer-grading system available Politecnico Moodle portal. Furthermore, the student documents part of their practical activities on lab reports. Group interaction will be fostered, even if all homeworks and lab reports are individual.

The teaching material (slides, handouts and reference scientific papers) will be made available by the class teachers on the Politecnico courses web portal and it is sufficient to cover all the topics taught in the class.

Lecture slides; Lab exercises; Video lectures (current year); Simulation tools; Self-assessment tools;

Exam: Compulsory oral exam; Individual essay;

The final evaluation is based on the oral discussion on the delivered reports. The report will be evaluated based on correctness, original contribution and quality of presentation. The oral will be aimed at discussing the reports and at assessing the level of understanding of the theoretical topics covered in the class. Let GA be the grade of the oral of Part A and let GB be the grade of the oral of Part B, the final grade G of the exam will be computed based on the credit-weighted average of the two parts as follows: G=nearest-integer(4/6*GA+2/6*GB) Both GA and GB must be >=18 to be considered valid.

In addition to the message sent by the online system, students with disabilities or Specific Learning Disorders (SLD) are invited to directly inform the professor in charge of the course about the special arrangements for the exam that have been agreed with the Special Needs Unit. The professor has to be informed at least one week before the beginning of the examination session in order to provide students with the most suitable arrangements for each specific type of exam.