Distributed Systems Seminars (DSS)
Corso di Laurea Specialistica in Ingegneria Informatica

AA 2010/2011

Principal Instructor Roberto Baldoni


News:

Next Exam of seminars of Distributed Systems will be Oct 4th at 12.00 pm room B214 (Via Ariosto 25) )

List of projects AA 2010/2011and assignments.

If you are intersted to a specific project, contact Roberto Baldoni sending the number of the project you wish to be assigned to (select a project that has not been assigned to any other student). Assignments will be done on a FCFS basis. If you wish to propose a your own project, please contact Roberto Baldoni with your idea.


DSS is a course in computing systems, with a strong focus on operating systems and distributed computing. The course is aimed at students seeking broad background in the areas of importance to modern systems researchers, or having an interest in possible research topics. The course involves a certain amount of reading, and all students who take the class are expected to read the papers for the experience to be satisfactory.  Most papers are chosen from the best conferences and journals in systems, and most are available on-line.The topics we'll span a range of hot areas within modern systems. 

The course is organized into three series of lectures called seminars:

 

Seminar I. Dr. Giorgia Lodi:

Monitoring large scale complex systems through Complex Event Processing Technologies

The course will introduce students to Complex Event Processing (CEP) Technologies and will provide practical examples of applicability of those technologies in two real case studies of monitoring large scale complex systems. In particular, the course consists of 3 lectures:

Lectures: March 25th, April 1st, April 8th, April 15th

Instructors: Luca Montanari, Leonardo Aniello

Seminar II. Prof. Roberto Baldoni:

Building reliable distributed systems

This series of lectures will introduce students to methodologies, techniques and tools for building reliable distributed computing systems and applications. In particular, it will be shown how the replication paradigms can be used to implement reliable systems, as well as middleware supporting developers during the implementation of highly available applications. This will be achieved also by introducing the most relevant technologies suitable for these ends in specific application contexts, such as interceptors and group communication toolkits. Practical works on one of such technologies will be part of the exam.

Instructors: Marco Platania, Giorgia Lodi

Lectures: April 29th, May 6th, May 13th, May 20th

 

Seminar III. Dr. Vivien Quema

Byzantine Fault Tolerance

web page of the seminar

Description : This series of lectures will introduce students to state-machine replication techniques tolerating Byzantine (i.e. malicious) faults. These protocols are commonly called BFT protocols. We will first explain why it is crucial to devise BFT protocols. We'll then present different state-of-the-art BFT protocols. The series of lectures is divided into 3 lectures:

1st lecture: 27/5/2011, 14:00-17:15, room A6. This lecture starts with an introduction about BFT protocols. We then present the seminal work made by Castro and Liskov on PBFT, the first practical Byzantine Fault Tolerant protocol. In particular, we explain how symmetric cryptography helps drastically decreasing the cost of the BFT protocol. We conclude this course by a description of several works aiming at decreasing the cost (in terms of hardware) of BFT protocols.

2nd lecture: 30/5/2011, 9:00-13:00, "Aula Magna". This lecture presents various protocols aiming at improving the performance of BFT protocols. In particular, we'll study in details the Q/U, HQ and Zyzzyva protocols. We will conclude this lecture by a presentation of Abstract, a framework easing the construction of efficient BFT protocols.

3rd lecture: 31/5/2011, 9:00-13:00, "Aula Magna". This lecture presents various protocols aiming at improving the robustness of BFT protocols. A BFT protocols is robust when it is able to sustain a good throughput in periods during which (clients or replica) faults occur. During the lecture, we will study in details the Prime, Aardvark and Spinning protocols.


Exam Rules

After each seminar, students students have to either write a (max 6 pages) report or do a small project. The report has to focus on a specific aspect of the seminar, for example by investigating a paper or a topic discussed during the seminar. For example considering the Seminar I, a report could focus on: failure prediction, port scanning, event processing systems. Alternatively the student could do a project on Esper to be decides with the Principal Instructor. Then students have to email to me (baldoni@dis.uniroma1.it) and to the instructor the report. In case of assignment of a small project, students have to contact the instructor for project evaluation.

To pass the exam:

1. Each student will be then assigned to a final project on a theme related to one of the three seminars and he has to submit to me and to the intructor a high-quality document about the project (max 20 pages).  The final project could be also the continuation of one of the seminar specific project/report. Assignments to final projects will be done on June 3rd.

2. The exam will end with a 20 minutes presentation of the student on the final project with the help of slides.

3. Each student has to attend 75% of the classes. In case this constraint is not satisfied, the student has to undergo a written exam on the content of the course to be defined with the Intructor.

4. The day of the exam, the student has to bring: the report/project done for each seminar, the final project report and the power point presentation.


List of Projects(2010 to be updated for 2011)

select the project you are interested in and contact the person assigned to the project. Assignments will be done on a FCFS basis. If you wish to propose a specific project, please contact Roberto Baldoni with your idea.