Відео

Ничего не понятно, можно говорить по-русски?

shipilev is my number one engineer!

Трипонятна

Great talk! At the end, due to the question, I got the feeling that most people didn't understand the talk. That is very normal because Pedro Ramalhete tried to insert a huge amount of very nuanced material into 1 hour. This is not to his fault, it is necessary for people outside of the field to understand the talk, but it probably isn't feasible in 1 hour.

Reading his "The paradigm of Multiprocessor Programming " , these guys are out of this world 🌎 so intelligent

Excellent presentation. Interesting new approach to storage fault recovery.

11:45 I don't get the reference to golang. if async/await come as a pair, then kotlin only uses async, whereas golang only uses await ('go func()' turns a sync function into an async function).

Thank you!👋

Amazing talk

Mind blown! 1:08:42

36:27 that's why kotlin Smart Cast work.

Unfortunately i only see PPT talks about TLA+ but never live coding in TLA+ from scratch. But that it an actual barrier. How to get your first model running. The "hello world" of TLA. Never. No one. And from this i come to questioning its practical applicability. Yes, everyone has heard anecdotes about Amazon. But what use are the talks about someone driving a car, where they show you all the possibilities you gain if you start using a car, but no one tells you how to start using it.

Thank you, I got a ton of insights!

Hello, very good presentation. I have not found a detailed analysis of how to deal with low-level failures (in the RPC layer). Leader and followers failures always occur, but there may be processor or network or time failures, as well as network partitions. The state in which the system should remain (in each of its members) are not detailed. Neither does it refer to membership or fault detection. These are the questions developers ask themselves when attempting an implementation

😝 *Promo SM*

The Chinese literature referred to the systems as "Distributed NewSQL".

Are you son of Steve Jobs?

Спасибо большое, но ребят, было бы неплохо если бы преза была на русском, если вы конечно хотите до русскоязычной аудитории достучаться. Когда Никита рассказывает еще норм, но когда вы начинаете что-то обсуждать то вся мысль уходит) и боюсь для нейтив инглиш слушателей тоже)

Slides: assets.ctfassets.net/oxjq45e8ilak/KU9K9OecXlAig5n5cVnYQ/e05f158e22e2a34722918b01f8204472/Jack_Vanlightly_Distributed_systems_showdown__TLA_vs_real_code_2021_06_15_14_08_33.pdf

I love how he took time to explain the concepts in depth. Great talk btw.

Great content, but video and sound quality need improvement. Please re-record this video =)

Great talk, thank you.

доклад -- моё почтение

Thank you for posting! I've been looking for this info, it was difficult to find.

дегротский юморок

гидру преступно мало людей смотрит(

Nice talk!

today I think for HFT rust would be a superior choice to java. because it can reach C++ performance and is memory safe without the big bloat & overhead of garbage collector.

Привет, Вадим!

I am a bit confused on how does the higher order function run is a call to a suspending function ? Should it be launch instead of run ?

Incredible presentation ! Rare gem! And I am really blown away by the fact that AtomicInteger constructor doesn't publish safely !

Brilliant illustration. I am sure that after watching this, many Java developers will be shocked that how incorrectly they were writing concurrency codes :)

Great talk! Andy missed an important player in the distributed SQL landscape: MariaDB Xpand 🙂 (previosly Clustrix). Cheers!

starts at 5:48

by mixing the syntax of PlusCal and Python, this looks more confusing than it needs to be.

Good Effort.Waiting For more

Where do I find an explanation of the diagram shown at 7:00 minutes? I see the x= global state. How do you interpret the "II", "ID", and "DD"? My best attempt at modeling this starts with global state "x = 1" and then does a D state transaction on the left and an I state transition on the right followed by an I and a D state transition to reach the final two states.

Im so frustrated that this video has only 150 views , this man is showing a way to to create AGI in using minisicule power like an iPhone in our hand

Your son got paid: loved that! 😁 not to mention the chest of knowledge that I am taking from this talk! Thanks, all three of you.

7:20 S3 судя по документации byte ranges всё-таки поддерживает.

it is a brilliant conceptual explanation for JMM

Almost 2 years later, still a fantastic talk 👍

Invaluable insight well put together into a brilliant talk 🙏

Rob pike wasn't kidding when he said "Simplicity Is Hard". On the outside golang seems to borrow alot of ideas from Erlang.

Great content, could definitely be 1/3 as long and just as effective, and could have better folks mc'ing.

The start is on 5:30

Starts from 5:40

This is pure gold. Thank you so much!

Great !

It is a really good explanation of Distributed condenses. I have the below queries It will be really helpful to understand. 1 . Is there any documentation on how Paxos or Raft implemented using Zookeeper or Distributed locking? 2. What will happen if primary goes down before distributing term to followers. Is there any data loss 3. If Initial primary comes back it will act like follower. Will there any chance to become primary after term sync ?