Understand RAFT without breaking your brain
Vložit
- čas přidán 6. 05. 2023
- RAFT is a distributed consensus algorithm used by many databases like CockroachDB, Mongo, Yugabyte etc. In this video, understand RAFT simply and get this important piece in your system design arsenal.
#systemdesign #database #computerscience
This is the best, most concise video on RAFT! Thanks!
bro ur video is so awesome like explaining the concept deeply and keeping it short too like that. examples are very much good considering all scenarios
Thank you so much 😀
I see that there are not many views on your channel inspite of the fact that your content is great. I think your videos target the niche tech audience who like to go into nitty gritty of things. Curious people like me, who just want straightforward technical knowhow. I just wanted to let you know that keep at it, don't have a motivation on making it big on CZcams. Just do it for your heart's content and passion like you already do. I am sure you will be rewarded eventually.
I like the new style of editing with after effects, but I would recommend this type of editing to only explain complex topics (Like what you did with RAFT) and stick to your old drawing board + facecam style for your normal videos they are much much better + you don't have to use gifs or stock videos to explain stuff. I like your straightforward style of editing its unique.
I really like your content I would appreciate if you make videos more frequenty like once a week, with minimal compromise on quality. I know its hard with editing, thumbnail, reserach, writing and all. But see to it that you increase upload frequency even if a little bit, because your videos motivates me to learn more about the topics you cover and I learn more in general and I am sure many people do so too.
Thanks for your feedback!! It means a lot! There are no videos now because I have moved houses and am in the process of getting my setup up. There are a lot of videos already in the pipeline... I agree with you on the editing style, I made this video in After Effects as it was too hot to shoot my face, but I will mostly be making videos with facecam
Hope you get more reach, great content
Excellent video with great graphics!
The best explanation I have ever seen of the subject. Keep the good work up 🎉
Thank you! 😃
Great video! Thanks!
Superb explanation!!
Great video! Really helped me out, thanks :)
Glad to be of help! Thanks!
Very nice
Excellent video, made reading the extended raft paper much easier!
Thanks!
Very well explained SIR!
Thanks much :)
Great explanation, thanks!
Glad it was helpful!
awesome explanation, thank you so much!! helped a lot with understanding how to implement my university project :D
Glad to hear! Your university is teaching a lot of good things then. My college was just theoretical.
Thank you 🧡
This was a very nice, to the point explanation of how RAFT works. Thank you!
Glad you enjoyed it!
Thank you 👍❤️ helped me lot for my exam preparation
Nice video brother
Great video!
Thanks!
thanx bro, one of the best explanations on raft
Glad you liked it
It looks similiar to 2 phase commit protocol which is used for distributed transactions
It does not explain leader selection
Nice
wow, my prof talked about 1 hour, i didn't understand it. But you explained it in few minutes. Thank you
Great to know that.. Thanks!!
explained like a cake walk. it really helped. thanks ❤
Glad it helped!
Thank you very much, you helped me with my Distributed Systems exam 😊
Happy to help
Love you man, this is the best explanation on RAFT .. If you have any documented resource on this can you please share that.
Thanks! There was this blog post I wrote many moons ago: ankushc.hashnode.dev/raft-taming-distributed-consensus
I don't even remember what I wrote in that so... best of luck :p
great video!
Glad you enjoyed it
kya mast samjhaya bhai
Thanks bhai!
Beautifully explained. Thanks for sharing bro
Thanks! Do consider sharing the video with you peers!
best video
very high quality presentation
Thanks! Although ... this was probably the only video I made or will make with this much effort on presentation. It takes way to much time with my super noob after effects skills :P
nais video bhai, unique style too
Glad you liked it
Instead of sending prev index, isn't it possible that the concrete follower will check to see if its current index has a diff with its previous index which is > 1?
Great explanation sir
Thank
you@@core_dump
legend
thanks bud 😭
this broke my brain.
Oh No!...sorry for that ... but what doesn't kill you makes you stronger I guess...
what you use to make videos ?
After Effects... This is actually the video I made to learn After Effects from scratch.
Won't it be faster if the 7:10 disconnected node gives its last known log instead of doing this?
I was thinking the same. I took a look on the raft consensus page they have an interactive visual representation of the algo, and I don't see such backwards checking, it was adding it without checking every item.
@@ankush_chatterjee thank you
@@ankush_chatterjee Thank you for the explanation, I checked out the article, and the GeeksForGeeks article on the same topic (the paragraph above the "Safety" section): "The AppendEntries RPC will iteratively send the RPCs with reduced Index Numbers so that a match is found."
Still the question was why is it made like so, because if few nodes missed 1000 transactions, then the load will be dangerous, just like RAID failures putting pressure on the main disc during rebuild and thus can fail.
in raft whitepaper, the author also thought of this, this was written:
"If desired, the protocol can be optimized to reduce the
number of rejected AppendEntries RPCs. For example,
when rejecting an AppendEntries request, the follower
7 can include the term of the conflicting entry and the first
index it stores for that term. With this information, the
leader can decrement nextIndex to bypass all of the con-
flicting entries in that term; one AppendEntries RPC will
be required for each term with conflicting entries, rather
than one RPC per entry. In practice, we doubt this opti-
mization is necessary, since failures happen infrequently
and it is unlikely that there will be many inconsistent en-
tries"
00:04 RAFT is a distributed consensus algorithm for data replication in case of server failures.
01:12 Raft helps achieve consensus by getting all nodes to agree on one state of the data.
02:25 The leader in RAFT is selected through an election process
03:31 Nodes become candidates when the leader dies and vote for themselves
04:33 Log entry is appended but not committed
05:42 Leader cannot append the log entry if there are gaps in the log
06:45 RAFT ensures consistent log across multiple nodes
07:52 Raft can be used on any finite State machine, not just a database.
Crafted by Merlin AI.
Thanks!
Who is Tom Tato?