{"id":34759,"date":"2019-10-31T22:00:14","date_gmt":"2019-10-31T19:00:14","guid":{"rendered":"https:\/\/prohoster.info\/blog\/osnovopolozhniki-teorii-raspredelennyh-sistem-v-obyatyah-gidry\/"},"modified":"2019-10-31T22:00:14","modified_gmt":"2019-10-31T19:00:14","slug":"osnovopolozhniki-teorii-raspredelennyh-sistem-v-obyatyah-gidry","status":"publish","type":"post","link":"https:\/\/prohoster.info\/en\/blog\/administrirovanie\/osnovopolozhniki-teorii-raspredelennyh-sistem-v-obyatyah-gidry","title":{"rendered":"Founders of Distributed Systems Theory in the Arms of the Hydra","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"

\"FoundersThis Leslie Lamport<\/strong> \u2014 the author of foundational works in distributed computing, and you might also know him from the letters La in the word La<\/b>TeX \u2014 \"Lamport TeX.\" He was the first to introduce the concept back in 1979 of sequential consistency<\/a><\/noindex>, and his article \"How to Make a Multiprocessor Computer That Correctly Executes Multiprocess Programs\"<\/a><\/noindex> received the Dijkstra Prize (specifically, the award was originally known as the \"PODC Influential Paper Award\" in 2000). There is an article about him on Wikipedia<\/a><\/noindex>, where you can find a few more interesting links. If you are excited by solving problems related to happens-before or the Byzantine generals' problem<\/a><\/noindex> (BFT), you should understand that Lamport is behind all of this.
\n<\/p>\n

<\/p>\n

And he's also coming to our new conference on distributed computing \u2014 Hydra, which will take place on July 11-12 in St. Petersburg. Let's see what this creature is.<\/p>\n

<\/a><\/noindex><\/p>\n

Hydra 2019<\/h1>\n

<\/p>\n

Topics like multithreading are among the most popular at our conferences, and have always been. Just a moment ago this room was empty, but now a person appears on stage discussing memory models, happens-before, or multithreaded garbage collection and \u2014 boom! \u2014 nearly a thousand people occupy every available space to sit and listen attentively. What is the essence of this success? Perhaps it's because we all have some hardware capable of organizing distributed computing? Or is it that we subconsciously recognize our inability to fully utilize it? There is a real story of a quantum analyst and developer from St. Petersburg, who has a computing cluster that only he can fully utilize. What would you do if you gained powers many times greater than you currently have for your tasks?<\/p>\n

<\/p>\n

Due to such popularity, the topic of performance and efficient computing tends to spread throughout the conference program. How much of the two days of presentations can be dedicated to performance \u2014 one third, two-thirds? In some places, there are artificial limitations that restrict this expansion: besides performance, there must always be room for new web frameworks, for some devops, or architectural astronautics. No, performance, you won\u2019t consume us all entirely!<\/p>\n

<\/p>\n

We can take the opposite path, give up and honestly hold a conference entirely about distributed computing and only that. And here it is, Hydra.<\/p>\n

<\/p>\n

Let\u2019s honestly admit that today all computing is somehow distributed. Whether it\u2019s a multi-core machine, a computing cluster, or a large-scale distributed service \u2014 there are many processes executing independent calculations in parallel, synchronizing with each other. This is what Hydra will focus on: how it is structured in theory and operates in practice.<\/p>\n

<\/p>\n

Conference Program<\/h1>\n

<\/p>\n

The program is currently being formed. It will include talks from the founders of distributed systems theories and engineers working with them in production.<\/p>\n

<\/p>\n

For example, it is already known that Leslie Lamport from Microsoft Research and Maurice Herlihy from Brown University will participate. <\/p>\n

<\/p>\n

\"Founders Maurice Herlihy<\/strong> is a very well-known and respected professor of Computer Science, and there is also a Wikipedia page<\/a><\/noindex>, where you can browse through links and works. There, you may notice two Dijkstra Awards, the first for his work on \u201cWait-Free Synchronization\u201d<\/a><\/noindex>, and the second, more recent one \u2014 \u201cTransactional Memory: Architectural Support for Lock-Free Data Structures\u201d<\/a><\/noindex>. By the way, the links lead not even to SciHub but to Brown University and Virginia Tech University, so you can open and read them.
\n<\/p>\n

<\/p>\n

Maurice is going to give a keynote titled \u201cBlockchains from a Distributed Computing Perspective.\u201d If you\u2019re interested, you can take a look at a recording of Maurice\u2019s talk from the St. Petersburg JUG. See how clearly and understandably he conveys the topic.<\/p>\n

\n

\r\n \"Play\r\n
<\/div>\r\n <\/div><\/center><\/p>\n

\"FoundersThe second keynote titled \u201cDual Data Structures\u201d will be presented by Michael Scott<\/strong> from the University of Rochester. And guess what \u2014 he also has his own Wikipedia page<\/a><\/noindex>. In Wisconsin, he is well-known for his work as a dean at the University of Wisconsin-Madison, and in the wider world, he is the person who, along with Doug Lea, developed those non-blocking algorithms and concurrent queues that Java libraries operate on. He received his Dijkstra Award three years after Herlihy, for the work \u201cAlgorithms for Scalable Synchronization on Shared-Memory Multiprocessors\u201d (as it should be, it is openly available<\/a><\/noindex> in the online library of the University of Rochester).
\n<\/p>\n

<\/p>\n

There is still plenty of time until mid-July. We will keep you updated on the other speakers and their topics as the program is refined and we get closer to July. <\/p>\n

<\/p>\n

In general, the question arises \u2014 why are we doing Hydra in the summer? After all, it's a dead season, vacation time. The problem is that among the speakers are university professors, and at any other time, they're busy. We simply couldn't choose different dates.<\/p>\n

<\/p>\n

Discussion Zones<\/h1>\n

<\/p>\n

At other conferences, it's common for a speaker to deliver their presentation and then leave right away. Participants don't even have time to find them \u2014 the next talk starts almost immediately after. This is very painful, especially if important people like Lamport, Herlihy, and Scott are present, and you're actually attending the conference just to meet them and discuss something. <\/p>\n

<\/p>\n

We have solved this problem. Immediately after their presentation, the speaker goes to a special discussion zone, equipped with at least a whiteboard, and you have quite a bit of time. Formally, the speaker promises to be there at least for the entire break between talks. In reality, these discussion zones use the same name for<\/em> can stretch on for hours (depending on the speaker's willingness and endurance). <\/p>\n

<\/p>\n

As for Lamport, if I understood correctly, he wants to convince as many people as possible that TLA+<\/a><\/noindex> \u2014 this is a good thing. (Article about TLA+ on Wikipedia<\/a><\/noindex>). This might be a great opportunity for engineers to learn something new and useful. Leslie suggests that those interested can watch his past lectures and come with questions. So instead of a keynote, there could be a specialized Q&A session, followed by a discussion zone. I did a little googling and found an excellent course on TLA+<\/a><\/noindex> (officially duplicated in a playlist on YouTube<\/a><\/noindex>) and a one-hour lecture \"Thinking Above the Code\"<\/a><\/noindex> from the Microsoft Faculty Summit.<\/p>\n

<\/p>\n

If you have viewed all these people as immovable names from Wikipedia and book covers \u2014 it's time to meet them in person! Have a conversation and ask questions that academic papers won\u2019t answer, but their authors will gladly engage with.<\/p>\n

<\/p>\n

Call for Papers<\/h1>\n

<\/p>\n

It's no secret that many of those reading this article wouldn't mind sharing something interesting themselves. From an engineering perspective, a scientific one \u2014 in any way. Distributed computing is a very broad and deep topic, where there's room for everyone.<\/p>\n

<\/p>\n

If you want to present alongside Lamport, it's entirely possible. To become a speaker, you need to follow the link<\/a><\/noindex><\/strong>, read everything carefully and act according to the instructions. <\/p>\n

<\/p>\n

Rest assured, as soon as you engage in the process, you will receive assistance. The program committee has sufficient capabilities to help with the presentation, its content, and formatting. The coordinator will help with organizational matters and so on. <\/p>\n

<\/p>\n

Pay special attention to the image with the dates. July is quite far away for a participant, but the speaker needs to start acting now. <\/p>\n

<\/p>\n

\"Founders<\/p>\n

<\/p>\n

SPTDC School<\/h1>\n

<\/p>\n

The conference will take place at the same venue as the SPTDC School, therefore, everyone who purchases a ticket to the school will receive conference tickets \u2014 with a 20% discount<\/strong>.<\/p>\n

<\/p>\n

Summer School on Practice and Theory of Distributed Computing (SPTDC)<\/a><\/noindex> \u2014 a school providing a wide range of courses on practical and theoretical aspects of distributed systems, led by recognized specialists in the corresponding field.<\/p>\n

<\/p>\n

The school will be conducted in English, so here's what the list of topics looks like:<\/p>\n

<\/p>\n