Runtimes for concurrency and distribution
Master Degree in Computer Science
University of Padova, a.y. 2020/2021
Instructor: Tullio Vardanega

(Latest update: 18 December 2020, 16:20)
[published specification of exam requirements
published material for lecture 22]

Outline of course

Learning outcomes

The higher the level of abstraction (and thus the expressive power) of the programming language -- which actually means the more distant from the actual operation of the underlying processor hardware --, the larger the role played by the software layer that implements the fraction of semantics of it that cannot be fully resolved at compile time. That software layer is usually termed language runtime.

Most evidently, such language features as concurrency and distribution involve semantics that is not primitive to the computer hardware, and therefore require the compilation system of the corresponding programming language to be augmented with software libraries dedicated to make the relevant semantics happen at run time.

This class aims at exploring the requirements placed on such runtimes, and their typical operation, considering exemplary paradigms of language-level concurrency and distribution, progressively ranging from single-CPU computers to the Cloud.

The instructional method used in this class will combine frontal lectures with multimedia resources, homework assignments, and flipped-class activities.

Useful reading

Concurrent and Real-Time Programming in Ada
Alan Burns and Andy Wellings
Cambridge University Press ISBN 978-0-521-86697-2
Distributed Systems - Principles and paradigms (2nd edition)
Andrew S Tanenbaum, Maarten van Steen
Pearson Prentice Hall ISBN 0-13-239227-5

Technology resources

Lecture schedule

Week	Date	Lecture #	Lecture Topic
1	28 September	1	On the notion of run-time support
1	30 September	2	On multi-programming
2	05 October	3	On virtualization Homework: Investigate the inner workings (under the hood) of Virtualbox
2	07 October	4	On concurrent programming Homework: Scrutinize the "driving motivation" of modern languages that have chosen to support co-routines or other forms of concurrency Resources: C#'s concurrency model Resources: Per Brinch Hansen on Java concurrency Resources: The inheritance anomaly problem
3	12 October	5	Continuation of lecture 4
3	14 October	6	On communication among threads
4	19 October	7	Continuation of lecture 6 Resources: Ada's base concurrency model Resources: About Ada tasking Resources: Ada for the C++ or Java developer Example: The notion of master & dependent
4	21 October	8	On synchronous communication Example: Programmatic realization of the Dining Philosophers case
5	26 October	9	Continuation of lecture 8 In-class exercise: analysis of Eratosthenes' sieve: synchronous version Homework: try and reproduce the above program in your language of election
5	28 October	10	On asynchronous communication Example: Application-level control of call queuing Example: A producer-consumer system with bounded buffer Homework: Study Eratosthenes' sieve: asynchronous version and reproduce in your language of election Resources: Per Brinch Hansen: "Structured Multiprogramming" (CACM, 1972)
6	02 November	11	Continuation of lecture 10
6	04 November	FC1	(*Flipped class*) You report on other models of language-level concurrency and we all discuss your findings
7	09 November	12	On the multiple facets of synchronization Example: Application-level allocation of resources Resources: Toby Bloom: "Evaluating Synchronization Mechanisms" (CACM, 1979)
7	11 November	FC2	(*Flipped class*) You constructively solve a design-and-programming problem posed by the instructor
8	16 November	FC3	Continuation of FC2 Example: An elegant solution in Ada Homework: Try and reproduce (or improve) the example solution in your language of election and single out the comparative pros and cons
8	18 November	13	An introduction to distributed systems Resources: On system scalability
9	24 November (08:30-10:15)	14	Continuation of lecture 13
9	25 November	15	Distributed inter-process communication Resources: Naming & DNS
10	30 November	16	Continuation of lecture 15
10	02 December	17	Distributed concurrency Resources: Statelessness
11	09 December	18	Distributed synchronization Homework: Study the inner workings of this Ada demonstrative concurrent implementation of Chandy-Lamport's distributed snapshot Resources: Chandy and Lamport's "distributed snapshot" Resources: Lecture material @ Cornell University (Ken Birman)
11	10 December (14:30-16:15)	19	The native distribution model in (early) Java
12	14 December	20	The native distribution model in Ada Example: Distributed version of Eratosthenes' sieve in Ada - source code Resources: Another use-case example Resources (implementazione middleware): GLADE, RT-GLADE Resources: YAMI4: un middleware per scambio messaggi in distribuito
12	16 December	21	Continuation of lecture 20
			Exam requirements
13	21 December	22	Cloud computing - general introduction - Ken Birman @ Cornell view of the Cloud Homework (self-study, material in Italian): - on Cloud-enabled applications - on scalability and multitenancy
15	11 January	23	Cloud computing: orchestration as the new middleware [Closing seminar]

Hall assignment and Zoom links

The hall assignment for the lectures in the schedule are reported in the relevant UNIPD's app.

The corresponding Zoom links will be communicated via the Moodle page of this class.

Appointments with students

Students are welcome to request personal or group appointments with the instructor, who will offer either a face-to-face or a Zoom-based meeting as the circumstances will allow.