Redis Deep Dive w/ a Ex-Meta Senior Manager
Вставка
- Опубліковано 22 лип 2024
- Redis (Valkey, MemoryDB, Garnet, Dragonfly, etc.) is one of the most versatile and useful technologies to know for your system design interview with companies like Meta, Google, Amazon, and others. Walk through essential elements of Redis in system design interviews: how it's used, how it works under the covers, and common patterns you're likely to see in practice.
00:50 Why Redis
01:51 Redis as a User
05:36 Redis Cluster
09:55 As Cache
12:41 As Rate Limiter
15:26 Streams and Async Job Queues
19:44 Sorted Sets and Leaderboards
23:25 Geospatial Indexes
26:40 Pub/Sub
Full writeup here:
www.hellointerview.com/learn/...
Redis commands docs are a great reference for the capabilities of Redis:
redis.io/docs/latest/commands/
Connect with me on LinkedIn:
/ stefanmai
Preparing for your upcoming interviews and want to practice with top FAANG interviewers like Evan? Book a mock interview at:
www.hellointerview.com
Good luck with your upcoming interviews!
I've been waiting for videos on deep dives. More deep dives please!! These rock
🫡
Can't thank you enough. You made system design so much easier to understand.
Just love the way you add practical aspects! Thankyou for the thoughtful, concise and articulate deep dive! Was definitely a high ROI on time :)
So glad it was useful!
Loving the consistency of uploads! :)
would love to see a deep dive on Kafka as well
Kafka is in the works! Written resource will be posted to the website this week or early next!
Wait .... I thought I clicked on a Dua Lipa music video. 😀. Just kidding. Stefan I love this - clear, concise, articulate and I learned something. Keep 'em coming.
Lol. The intro is a banger, no :)?
The problem of hot key is clear but solution of hashtag is very difficult to understand. do we have to handle that all in the client? like figure out what are the hot keys?? and what will adding hashtag do really?
Great watch! A similar deep dive video for Kafka would be amazing 🙏
Kafka is in the works! Written resource will be posted to the website this week or early next!
@@hello_interview amazing!
Excellent video. Please also make video on mock interviews
Mock interviews are what we do! What is it you want to see?
Absolutely loved it !!!
Can we have one deep dive on SQL vs NoSQL, a topic always asked in system design interview.
We actually have an opinion on that very topic, check this out: www.hellointerview.com/learn/system-design/in-a-hurry/key-technologies#core-database
Can we have a deep dive on Apache Flink please ?
We'll put it on the list!
Very informative and helpful video as always. I would appreciate if the visualization of the internals could be a bit more detailed and in sync with what you describe, rather than just a couple of boxes. Regardless, this is a refreshing take to step away from the usual system design videos to dive into technologies that are useful in both interviews and real life. Hope to see more of this in the future.
Good feedback, thank you! Will try to make the visuals more engaging in the future.
Great content. Really liking it. Can you please do a deep dive on TiDB, Rocks DB and Cassandra too
Cassandra is soon!
It's a great deep dive. I'd like to give a suggestion, Can you guys add chapters to the video, so that, if I want to rewatch again about certain portion of the video then that would be easy thing to do.
You should have chapters on this one! If you look at the description "Key Moments" will take you to the relevant sections. LMK if we're not doing something else we should be doing!
At 22:55 how does hashing tweet IDs help if we're looking for the most liked tweets overall? Or were you referring to getting top liked tweets for specific keywords when splitting that key across multiple nodes?
The latter!
Amazing Content.. Can't thank you enough!! Any Plans for DSA Mock Interviews 👀
What are you looking to see?
@@hello_interview DSA Mock Interviews with FAANG Candidates..
Can i book sys design coaching sessions on hello interview or is it only for mock interviews?I'm a senior engineer and I've just started preparing do you recommend to jump straight into mocks?
Send us an email at support@hellointerview.com and we might be able to make something work. Have some products planned here for people in your shoes.
@@hello_interview done
10:00 didn't quite catch how adding random suffix would solve hot node problem
Splits the hot shard into multiple, so you can distribute that load across nodes.
@@hello_interview sorry, what's the approach? I think somewhere else in the video talks about it, but I also somehow did not understand that.
Is it roughly the following?
Generate X random suffixes. Maintain a map on the (Redis') client side of .
Whenever we want to do some sort of "write", we get a random suffix from our list and append it to our key in the Redis command.
Whenever we want to do some sort of "read", we get _all_ random suffixes (X of them) from our set, make X reads from Redis, and aggregrate on the client side.
----
Or maybe it's sometimes the opposite way. For example, in a cache, I'm guessing we'd want to write our value to each and every one of the key+randomsuffix.
Then for a read, we can (randomly) pick one of the suffixes.
----
Maybe the missing link (for me and others similarly confused) is how/where the random suffixes are managed.
There's also the question of how we (automatically?) identify a hotkey, or develop an algorithm that generates (or increases) the number of suffixes depending on some sort of access rate.
Can you please do deep dives on Kafka and Spark too ? I love your videos
Written deep dive on Kafka coming in a couple days!!
Why would I want to use Redis stream over something like Kinesis or Kafka streams
In practice? Substantially faster and cheaper if you can accept some of the (many) compromises. In an interview setting it can be helpful to have a few multifaceted tools vs having to know all about Kafka (video on that soon).
If your business was entirely built around append-only logs I probably wouldn't take Redis as an acceptable substitute, though it works in a pinch.
But how will you handle the at most one delivery of redis in the case messaging?
Check out our Whatsapp guide for one example of this: www.hellointerview.com/learn/system-design/answer-keys/whatsapp
Re-upload, sorry! Stefan is a UA-cam noob.
Facts.
🤣
Couldn't focus, too handsome
The GOAT has arrived and has something important to say!
@@hello_interview Gigachad onlyfans allocator (of) toes, that's very kind of you
I have never clicked so fast!
Nice! Can you also do Deep Dive on Graph Databases(eg, neo4j) as well. please!
Will add it to the list! Typically, graph dbs are not as commonly used at scale.
do a deepdive on ethereum latest version, l2 rollups
It would be nice to see a Deep Dive video about location database. Pretty hard to understand for me
Which part is most confusing? Geohashing? Quad trees? Where to use them?
@@hello_interview Geohashing algorithm is too complex. I just cant take in all this diagrams. I also dont understand how proximity systems can search something inside the circle using square blocks of geohashes
@@maxvettel7337 Gotcha! I'll earmark this for follow-up. The nice thing is you don't necessarily need to know about the geohashing internals to use something like Redis' proximity search in practice!
Kafka next, please!
This week!
I see , I like.