On July 11-12, that is, this Thursday and Friday, a conference will be held . These are two days and two tracks of reports dedicated to distributed computing. The reports are given by the best scientists and engineers who came to St. Petersburg from around the world. The conference is aimed at specialists in the field, no introductory reports!
You can watch a completely free online broadcast. It will only have first day and first room + online interview between reports. What kind of reports - we will discuss below.
It is important that the broadcast will begin at 9:45 am (Moscow time), 15 minutes before the opening, and end closer to 8 pm. All this time you can listen to reports with short breaks. The link will work all day, so you can open it only on the reports that are most important to you.
Link to the site with the video and the program is under the cut. In the same place we will discuss a few things that will not be broadcast, but which are available to participants who came to the conference live.
Where to get broadcast
The broadcast page is waiting for this green link button:
In the presence of a video player and the program of the first hall. The player will come to life only on the morning of July 11, now it shows nothing.
Reports
It all starts with keynote Cliff Click . Cliff is a legend in the Java world, the father of JIT compilation and the magician of low-level performance. We did with him I recommend reading. This is a report about that incredible supercomputer created in the depths of Azul.
The second report is from Ori Lahav from Tel Aviv University. Ori is involved in programming language research, formal verification, and in particular everything related to multithreading. in the report we'll take a look at how C++11's threading model is formally described and how to live with issues like out-of-thin-air.
In the third report, , Heidi Howard of the University of Cambridge will return to the theoretical foundations of Paxos, relaxing the original requirements and generalizing the algorithm. We will see that Paxos is, in fact, only one of the options among a huge range of approaches to consensus, and that other points on the spectrum are also very useful for building good distributed systems. Heidi specializes in consistency, fault tolerance, performance and distributed consensus.
— this is a report by Alex Petrov about how you can reduce the load on the storage if you keep data only on a part of the nodes, and use special nodes (Transient Replica) for failure handling scenarios. In the course of the report, we will look at Witness Replicas, the replication scheme used in Spanner and Megastore, and the implementation of this concept in Apache Cassandra under the names Transient Replication & Cheap Quorums.
Roman Elizarov from JetBrains will talk about . Roman is a team leader in the development of libraries for the Kotlin language and platform, personally involved in the architecture and implementation of coroutines.
And ends the broadcast - Keynote Maurice Herlihy, world-famous scientist and father of transactional memory. We did with Maurice worth reading before attending the report.
Total: six reports, transactional memory, memory models, distributed consensus, structured concurrency and even blockchains. Everything you need to have a great day.
If you want to get access to all reports in general (not only to the first hall) on Thursday and Friday, then you can . In fact, this is the only opportunity for those who have learned about the conference just now and will not have time to get to St. Petersburg. In addition, this way you will have all the videos of what happened. Difficult reports are distinguished by the fact that they are likely to have to be revised.
Not everything is available on stream
If you manage to buy a ticket at the last moment and come to the conference live, then there will be a few more interesting things:
Discussion zones
After each report, the speaker goes to a dedicated discussion area, where you can chat with him and ask your questions. Formally, this can be done during a break between reports. Although the speakers are not required, they usually stay much longer - for example, for the duration of the next report. Sometimes it makes sense to skip a report from the main program (if you bought a ticket, you will still have notes after filling out the feedback) and spend it on focused communication with an important expert.
Two BOF sessions
BOF is now a traditional format at our conferences. Something like a round table or discussion group in which everyone can take part. This format historically goes back to the first informal . There is no division into speaker and participant: everyone participates equally.
Currently scheduled : "Modern CS in the real world" and "Trade-offs in concurrency". Both BOF sessions are held in English only, as are many of the presentations and discussion areas at the conference.
Exhibition area
The exhibition is a zone of stands of the conference partner companies. Here you can learn about interesting projects, technologies and work in a team of IT industry leaders. This is the place where you and the company can find each other. On Hydra with us и .
Party with beer and music
In parallel with the BOFs, a party starts at the end of the first day. Drinks, snacks, music - everything at once. You can chat in an informal setting and discuss everything in the world. You can move from the bof to the party. You can move from a party to a bof.
Next Steps
- If you are watching a free stream: you need to go on Thursday, July 11th. The broadcast will begin around 9:45 am Moscow time.
- If you want to have access to all reports and recordings after the conference: .
- If you change your mind and go live: you have less than a day left to purchase a ticket, all possible options are .
Source: habr.com