Theory is when you know everything but nothing works.
Practice is when everything works but no one knows why.
distributed systems, theory and practice are combined:
nothing works and no one knows why.
To prove that the joke in the epigraph is absolute nonsense, we are holding SPTDC (school on practice and theory of distributed computing) for the third time. About the history of the school, its co-founders Petr Kuznetsov and Vitaly Aksyonov, as well as the participation of JUG Ru Group in the SPTDC organization, we have already told on Habr. Therefore, today is about the school in 2020, about lectures and lecturers, as well as about the differences between the school and the conference.
The SPTDC school will be held from 6 to 9 July 2020 in Moscow.
All lectures will be in English. Lecture topics: persistent concurrent computing, cryptographic tools for distributed systems, formal methods for verifying consensus protocols, consistency in large-scale systems, distributed machine learning.
Did you immediately guess what military rank the characters in the picture are? I adore you.
Lecturers and lectures
Nir Shavit (Nir Shavit) is a professor at MIT and Tel Aviv University, co-author of a great book The Art of Multiprocessor Programming, owner Dijkstra Prizes for the development and implementation software transactional memory (STM) and Gödel Prize for his work on the application of algebraic topology to the simulation of shared memory computing, co-founder of the company Neural Magic, which creates fast machine learning algorithms for conventional CPUs, and, of course, has its own Wikipedia pages with dashing and sultry photography. Nir already participated in our school in 2017, where he gave an exhaustive review of blocking techniques (Part 1, Part 2). What Nir will talk about this year, we do not yet know, but we hope for news from the cutting edge of science.
Michael Scott (Michael Scott) is a researcher in University of Rochester, known to all Java developers as the creator of non-blocking algorithms and synchronous queues from the Java standard library. Of course, with Dijkstra's Design Award synchronization algorithms for shared memory computing and own Wikipedia page. Last year, Michael gave a lecture at our school on non-blocking data structures (Part 1, Part 2). This year he will tell about programming using non-volatile memory (NVM), which reduces program complexity and memory overhead compared to "regular" random-access memory (DRAM).
Idit Keidar (Idit Keidar) is a professor at the Technion and the owner of Hirsch index about 40 (which is very, very much) for two hundred scientific articles in the field of distributed computing, multithreading and fault tolerance. Eidit participates in our school for the first time, where she give a lecture about the basic aspects of the work of distributed data warehouses: distributed memory emulation, consensus development and configuration changes.
Rodrigo Rodriguez (Rodrigo Rodrigues) - professor at Técnico, member of the laboratory INESC ID and author research work in the field of distributed systems. This year at our school Rodrigo will tell about consistency and isolation in distributed data warehouses, and will also analyze using CAP theorems feasibility in practice of several models of consistency and isolation.
Chen Ching (Jing Chen) is a professor at the State University of New York at Stony Brook, author research work in the field of blockchain and a leading scientist in Algorand — a company and a blockchain platform using a consensus algorithm entirely based on Proof of Stake. This year at our school, Chen will talk about the Algorand blockchain and ways to achieve its interesting properties: undemanding to network computing resources, the impossibility of splitting the transaction history, and guaranteeing the end of transaction processing after it is added to the blockchain.
Christian Kashin (Christian Cachin) is a professor at the University of Bern, head of a research group in the field of data protection, co-author of the book "Introduction to Reliable and Secure Distributed Programming”, blockchain platform developer Hyperledger Fabric (about her even was post on Habré) and author research work in the field of cryptography and security in distributed systems. This year in our school Christian give a lecture in four parts about cryptographic tools for distributed computing: symmetric and asymmetric cryptography, and also about shared key cryptography, pseudo-random numbers and verifiable random number generation.
Marko Vukolich (Marko Vukolic) is a researcher at IBM Research, author works in blockchain and developer of Hyperledger Fabric. We don’t yet know what Marco will talk about at our school this year, but we hope to learn about his latest developments in the field of blockchain: research performance degradation distributed consensus protocols on clusters of up to 100 machines, broadcast Mir protocol with global order and Byzantine fault tolerance or blockless blockchain StreamChainminimizing transaction processing time.
Prasad Jayanti (Prasad Jayanti) is a professor at Dartmouth College, part of the elite ivy league, and the author research work in the field of multithreaded algorithms. This year in our school Prasad give a lecture about thread synchronization and algorithms for implementing various options mutex: with interrupt or restore functions in non-volatile memory models, and with separate read and write operations.
Alexey Gotsman (Alexey Gotsman) is a professor at IMDEA and an author research work in the field of program verification of algorithms. We do not yet know what Alexey will lecture at our school this year, but we are looking forward to a topic at the intersection of software verification and distributed systems.
Why is this a school and not a conference?
Firstly, the lecturers speak in an academic format and read two pairs of each large lecture: "an hour and a half - a break - another hour and a half." Many years out of college, with a habit of hour-long conference talks and 10-minute YouTube videos, this can be tricky. A good lecturer will make all three hours interesting, but everyone is responsible for the plasticity of their own brain.
Helpful Hint: Practice on video recordings of school lectures in 2017 year and 2019 year. Goodbye, work - hello, Byzantine generals.
Secondly, the lecturers focus on scientific research and talk about the fundamentals distributed systems and parallel computing, as well as news from the cutting edge of science. If your goal is to quickly code something and deploy it to production the next day after school in hot pursuit, this can also be difficult.
Helpful Hint: Look for the research papers of the school's lecturers at Google Scholar и arXiv.org. If you enjoy reading scientific papers, you will enjoy the school too.
Thirdly, the SPTDC 2020 school is not a conference, because the conference on distributed systems and parallel computing is Hydra 2020. Recently on Habré there was a post with review of its program. Last year, SPTDC and Hydra took place simultaneously and on the same site. This year they do not overlap in dates, so they do not compete with each other for your time and attention.
Helpful Tip: Check out the Hydra conference program and consider attending the conference after school as well. This will be a good week.
How to get to school?
- Write down the dates from July 6 to July 9, 2020 in the calendar (or better, by July 11 to go to the Hydra conference after school).
- Take heart, get ready.
- Choose tickets and go to school.
Source: habr.com