Distributed Systems in One Lesson by Tim Berglund
Vložit
- čas přidán 31. 05. 2024
- Normally simple tasks like running a program or storing and retrieving data become much more complicated when we start to do them on collections of computers, rather than single machines. Distributed systems has become a key architectural concern, and affects everything a program would normally do-giving us enormous power, but at the cost of increased complexity as well.
Using a series of examples all set in a coffee shop, we’ll explore topics like distributed storage, computation, timing, messaging, and consensus. You'll leave with a good grasp of each of these problems, and a solid understanding of the ecosystem of open-source tools in the space. - Věda a technologie
Tim's lectures are so funny and captivating that I can binge on them instead of Netflix.
Tim has a knack for explaining things in a clear and intuitive way. A great teacher, I hope he does more of this. Regards from Chennai.
Many thanks!
I don't know if it is a natural skill from this speaker, but he speaks in a very clear way that i can watch the talk in 1.5 speed. Thank you.
Amazing single place to know about Distributed system, tools and techniques.
Thanks for sharing...
"I am the very definition of mutable state" Awesome. I think I may need to steal this. Thank you Tim!
A great speaker , at the same time an expert and pro educator :) I love your speech, wish I could have your presentation skills..
Love this video ❤️ didn't expect such an amazing content to be available for free. Internet is really a bliss most of the times
Glad you enjoyed it!
Well the idea is that you later go buy Confluent’s services 😉 Kafka is a monster to maintain yourself
wonderful presenter, great delivery of crisp information.
He had worked in cassandra in earlier and now for kafka. Dude, you are the evolution of the distributed systems.
Thanks for the lessons m Berglund. This was by far one of the educating 40 minutes I had in years. I'd like to thank you for sharing your knowledge.
Glad it was helpful!
I love this guy. Cristal clear and definitely he loves distributed systems.
This talk brings some light into all these technologies and helps decide what to learn next.
the original lesson is on Oreilly and its amazing
made it pretty clear from start to end. thank you for sharing.
Gavin Belson? :O
Lol
:xD Haha...
I got confused after seeing this comment :p
Dude same here ! I also thought on what earth Gavin Belson is here...
Yeah. You kind of nailed it. ;-)
WTF lol hooli
Simply love the way Tim Bergland covers the topics. Excellent skills displayed.
Many thanks!
What a great speaker
very concise and to the point !
This is Jeff Winger fron Community tv show LOL
Same thoughts.
@@hellelo.5840 d d. Nnfnd. F d. De. D. N n. D nd. B ndd nn. D y b. D dd d. De. D. Fb. D. D. D. D. L D. D. De. D. Ddn.. D. F d f. DIY d. Xl dd. D d. D d d. D d. Dad d d d d d. D. D. D d. D d
@@hellelo.5840 d d. Nnfnd. F d. De. D. N n. D nd. B ndd nn. D y b. D dd d. De. D. Fb. D. D. D. D. L D. D. De. D. Ddn.. D. F d f. DIY d. Xl dd. D d. D d d. D d. Dad d d d d d. D. D. D d. D d. Y d. Fu d
Great introduction to Distributed Systems. So often the industry gets bogged down in buzzwords and cliche terms that newcomers find it difficult to know where to begin. This is a great starting point.
Marketing loves to obfuscate.
"I like americano today, tomorrow maybe [...] mocca with extra foam - I'm the definition of mutable state" 😂
Mllk
Can you share the link to the 4-hour lecture about distributed systems you mentioned in the start, please?
A great piece of presentation from a great speaker.
I was inspired by this and another talk about system design (parking lot with premier parking spaces). I n your honor I am adding an external drive, for the hidden- read/write copy of a groupware system with an asynchronous don't ask don't tell, storage management unit. The best i ever had... The best i ever had ......
Great speaker n teacher! Thanks for sharing!
Great introduction to Distributed Storage, Computation, & Messaging.
Glad you liked it
Great content on introduction to Distributed Computing. I enjoyed the session. Thank you Tim.
Glad it was helpful!
one hell of insight full talk on distributed system
Can we have the link to the 3-4 hr long video Tim talked about?
Awesome talk, great learning on distributed systems
Amazingly explained. Interesting speaker :)
Really enjoyed listening to this lecture, thanks :)
It was a very good lecture. Thanks for the talk
Great job Tim, nice stuff.!!
Excellent speaker, that was pure joy
Presentation was great and explanation was clear. Thank you!
Glad you enjoyed it!
Thanx sir very helpful..
Nice talk. It feels like watching a good movie :)
Great presentation
Thats a great insight on distributed systems. One thing at 42:00 Tim mentioned about consistent hashing, where as in example of topic partitioning he used modules operation, which doesn't derive consistent hashing.
Great presentation👏
great video, awesome explanation
Thanks , its a great session !!!
Best explanation of Cassandra and consistent Hashing in all of CZcams
Thank you
very well explained core concepts and problems about distributed systems, thanks Tim
Glad you enjoyed it!
Very good presentation
Excellent speaker holy moly.
Awesome speaker. Wish I could deliver talks in this manner.
Love him as well
he has good explanations
great presentation, thank you!
Glad it was helpful!
As a full stack developer who does stand up on the side, gotta say that was one tough crowd :D
Sad
Great lecture!
Thank you!
Such a powerful video
very inspiring!
Tim is amazing
great content!
Super cool presentation. I am also not a fan of football, but I like Cervantes, so my favorite football club is Real Madrid.
awesome talk.
Question on Topic Partitioning (at 41:00): together with each message, can we not include the timestamp when it was produced? Wouldn't it provide the global ordering?
Really appreciated very nice
Is "read replication" synonymous with multiversion concurrency control? Meaning, you have different versions of data items for each transaction that are distinguished by time stamp, and therefore avoiding conflicts?
No, it means a given version has X identical redundant copies.
and they're usually kept on separate servers, racks, or data centers.
nice talk thank you!
4:29 I thought my laptop was being possessed by Satan for a second.
Its really a good talk...
i didn't know about all these pain involved with distributed system.
45:37 "They say the best code is the code you never write and the worst code would be the code you write two or more times" 👏
Akshay AS except good code is usually rewritten until it’s great, so this isn’t quite true.
Anyone else noticed: "Kakfa" in the heading of the slide :O..
ya thats awsome
Amazing
Can someone point me to the link for the 4 hr video, he referenced to?
Also interested.
He is also Hooli's CEO.
This is correct
Very nice explanation
Thanks for liking
6:41 "In a relational database, reads are usually than writes" I don't understand this. I thought reads are generally more expensive since you might have joins? While when writing you typically add some rows to a single table and that's it.
Can somebody help explain it to me?
Or I guess it's because in a distributed system you have to synchronize the writes to other machines?
I think he meant more reads than writes in terms of volume.
Because OLTP databases do a great job of caching and using indexes to optimise read. Even the storage characteristics and disk layout is usually optimised for read traffic.
He meant there are usually more reads than writes, this is why the first step is to have some replicas to use for the reads
Very clear
Instant like for over simplified CAP theorem at 20:00
Nice talk.
"zed's dead" - the audience didn't get this. Hope so I did :)
Misspelled Kafka ("kakfa") on the slide at 34:10.
10:03 picks up bottle. 10:08 opens it to drink water (presumably) 10:24 shuts and keeps it back without drinking
Haha that's swag
23:28
great talk
Tim's a beauty.
Thanks
Love this video as I prepare for sys design interview
Glad it helped you.
You can skip to 2:47 if you want.
Dude... F*** you...
Nice talk..
One of few people where you can run the video at 2x speed and still understand what he is saying.
That's not a good thing
All this Hadoop/Spark thingy are so abstract that I have no clue what is what anymore. 🤣😂
23:00 about cap theorem.. did he confuse himself?
He almost made it seem like by not having availability, you would also lose consistency, but what he means is, if the node just doesn't respond, it's still consistent because it's not giving out bad/inconsistent data. In a big distributed system, this data could be retrieved from elsewhere while tolerating the consistency.
The focus of this video is Distributed Systems when writes and reads getting slower on Databases. What about application server getting too many requests? Why is that not covered as part of a problem that DS solves?
Because you scale the application server horizontally and load balance across the servers/processes
17:00 that order though
Continue watching: 12:42
Why still water over sparkling water?
apache sparkling water?
33:05 kakfa
What if I tell you that you read 'kakfa' as 'kafka' 33:07
33:03 kafka is not spelled correctly at the top there
This is good, but of course, very briefly.
"I draw examples from a coffee shop just to be cute". Ha ha.. Of course there is a dire need to be cute in this otherwise one hell of a hard core tech talk.
I am here for the second time!
Awesome
This is Jeff Winger fron Community tv show LOL
Good entry-level talk, but also would be great to give credits to Leslie Lamport, touch upon CRDTs (Conflict-Free Replication Data Types), mention consensus solutions like RAFT and Paxos, explain SQL vs NoSQL vs NewSQL, say PACELC (extended CAP), add an overview of consistency models (what is strong serializability?), and talk about leader election.
I know, too much, but that's the essence of distributed computing!
shameless plug at 5:46
Gavin Belson ! :p
that z thing is funny, especially in Poland :)
:)
If you are in front of the mic, always go with the still water :P
It is one of the Top10 Speakers' pro tips 😂
@@DevoxxPoland It should be the most important tip :P
Fyi, Kafka is not a message queue...