Got a question? Feel free to drop them in the comments below! We will make sure all your queries are answered. For more insightful videos, don't forget to subscribe to our channel. Please don't turn forget to turn on the bell 🛎 notification as well! You don't want to miss out on exclusive content. ⚡Take the Hunger test today: bit.ly/41fhgDW
Mr Anshuman, your experience is very much reflecting in such high quality content you have shared. I shared your video to my folks and they were so happy to find this good content. So much thanks for sharing this with everyone .. 🙏
I didn't understand why a system that is partitioned has to choose between consistency and availability but your example at @8:53 made it clear. Thank you. 😊
Super simplified with real world example, teaching / trainings with these kind of explanations will stay in peoples mind forever. Thank you so much for the video.
Hey Anshuman, the example and analogy You chose to explain CAP theorem, made it too easy to understand it and remember it. Thank You for sharing Your knowledge.
00:04 Understanding CAP theorem and its significance 01:39 A premium number providing memory services for users. 03:19 Increased capacity to handle 2000 requests 04:57 Ensuring data consistency through shared diary entries 06:30 Availability problem: Unable to take new requests 07:57 Choosing between consistency and availability in network partition 09:42 In a network partition, you have to choose between consistency and availability. 11:25 Choosing between latency and consistency
wait! can't we solve this problem using even driven arch? like if explain in a layman lang, then why can't we have intermediates like one or more person who can tell you and your wife to note down data? like using a pub/sub or an api gateway and balance the load?
The reason I involved my wife was to increase the number of requests I can process. If you pass it down to another person, then that person becomes the bottleneck. If there are more than one person noting the data, then the same cases that I had with my wife should be into consideration between them.
Hey. I got an assignment as part of the interview process. Any guidance around this Make a end to end movie recommendation system where we need to provide an HIGH LEVEL DEGIGN of which components can be used and how they would interact with each other to provide movie recommendation for OTT client application. The requirement is 1. Capture all the data from user analytics. 2. Store it internally in the right system that could be used for such a large volume. 3. Stream data-set 4. Please note that the recommendation should be in real time to the OTT application. I am not sure how can I proceed with HLD and LLD
We recently did a masterclass on Personalisation @ Hotstar. And we keep doing masterclasses on System Design where we discuss an ideal way to design systems like Facebook News Feed, Hotstar, TH-cam ingestion, Instagram, etc. To get better in this component, please register for our free masterclass here: www.scaler.com/events
There might be a solution , if you take notes and replicate to your wife's diary and let a demon process took it in async mode , then you will have a higher chance of meeting consistency , but you have to make sure your deamon mode process will guarantee a job done , if the job is done you will send a notification to user , else design a failed job que and perform reprocess till you hit a counter .
I have a question wouldn't it be easy to use a central system where everything is stored. instead of storing it on their on personal diary what if they stored on a space which is shared between them in that case bot consistency and availability is possible.
What about CA (Consistent & Available but not network partition tolerant) & CP (Consistent & Partition Tolerant but not Available) systems, do they exist, if so how do they work?
Quick question : when u tell ur wife to note it down or vice versa , Doesn't it increases the load , which hampers the sole purpose of bringing ur wife in first place i.e to handle load ( more request to process) Plz help me out to understand
There are 2 kinds of requests. Read (where some calls and asks for relevant entries in my diary) and Write (Noting down entries in my diary). You are right. Load of writes does not reduce. However, read traffic is distributed between me and my wife. Most systems in the world are read heavy and hence this is okay. However, if you want to reduce write traffic, you will have to shard / split (I can only process users with names starting with a-m, while my wife handles everyone with their names starting from n-z).
Got a question? Feel free to drop them in the comments below! We will make sure all your queries are answered. For more insightful videos, don't forget to subscribe to our channel. Please don't turn forget to turn on the bell 🛎 notification as well! You don't want to miss out on exclusive content. ⚡Take the Hunger test today: bit.ly/41fhgDW
Mr Anshuman, your experience is very much reflecting in such high quality content you have shared. I shared your video to my folks and they were so happy to find this good content. So much thanks for sharing this with everyone .. 🙏
Thank you very much for your kind words
I didn't understand why a system that is partitioned has to choose between consistency and availability but your example at @8:53 made it clear. Thank you. 😊
Simple and clear! This would not need a revision from my side... Cheers!
I knew CAP theorem but the way you taught is awesome..You are a great teacher..Keep up the good work to help everyone.
Super simplified with real world example, teaching / trainings with these kind of explanations will stay in peoples mind forever. Thank you so much for the video.
This could not be any clearer, best explanation i've seen with perfect example. Thank you for sharing.
Hey Anshuman, the example and analogy You chose to explain CAP theorem, made it too easy to understand it and remember it. Thank You for sharing Your knowledge.
Absolutely loved the illustration you used. It sure drives home the point of CAP Theorem. Thank you.
Very easy explanation. Thanks!
Wow..... This so simpl but elegant example ...... Also great explanation.... Bro. Thank you 👍
Hey Man! Good example which you have taken, very impressive.I'll never forget to explain in the interview. Thank you so much!
Glad you liked it!
Excellent explanation!!! Subscribed
well explained. kept me stay on video for 12 min though i read cap theorem before, now i don't forget it
Thanks Prudhvi. Will do more videos in coming weeks. Will look forward to feedback :)
thanks man, great video
Simplest and best explaination ever.. cheers!
The best explanation for CAP Theorem in youtube
Super easy explanation 👍
Very well explained!! Thanks a lot !
Awesome explanation.
Perfectly explained
Wish we could have such amazing individuals teaching in our education system !
Thanks so much
plz make few videos for low level system design
Great Explaination!! 💯
Thank you very much for make it simple & easy to grasp
Simply awesome explanation Anshuman !!
Nicely explained. Thank you anshuman sir🙏
Enjoyed this, thanks!
Savior of my day
Very well explained sir! Thank you.
Great explanation!! Could you please also make a video on how to design a recommendation system?
Thank you so much!
brilliant explanation !
nice explanation
Thats the best explanation ever.
Love your teaching style. Looking forward to more videos on system design
00:04 Understanding CAP theorem and its significance
01:39 A premium number providing memory services for users.
03:19 Increased capacity to handle 2000 requests
04:57 Ensuring data consistency through shared diary entries
06:30 Availability problem: Unable to take new requests
07:57 Choosing between consistency and availability in network partition
09:42 In a network partition, you have to choose between consistency and availability.
11:25 Choosing between latency and consistency
superb.
great explaination sir it was helpful.
Excellent explanation. Pretty detailed & the use case has been fantastic.
Thanks😊
Simply superb.. Subscribed!
very nicely explained. Thank you Scaler Academy :)
wait! can't we solve this problem using even driven arch? like if explain in a layman lang, then why can't we have intermediates like one or more person who can tell you and your wife to note down data? like using a pub/sub or an api gateway and balance the load?
The reason I involved my wife was to increase the number of requests I can process. If you pass it down to another person, then that person becomes the bottleneck. If there are more than one person noting the data, then the same cases that I had with my wife should be into consideration between them.
Good explained 👍
Glad you liked it
🔥 🔥 🔥
amazing explanantion
I find it so good way of learning although I knew cap theorem
Thanks. Glad you liked it.
Hey. I got an assignment as part of the interview process. Any guidance around this
Make a end to end movie recommendation system where we need to provide an HIGH LEVEL DEGIGN of which components can be used and how they would interact with each other to provide movie recommendation for OTT client application.
The requirement is
1. Capture all the data from user analytics.
2. Store it internally in the right system that could be used for such a large volume.
3. Stream data-set
4. Please note that the recommendation should be in real time to the OTT application.
I am not sure how can I proceed with HLD and LLD
We recently did a masterclass on Personalisation @ Hotstar. And we keep doing masterclasses on System Design where we discuss an ideal way to design systems like Facebook News Feed, Hotstar, TH-cam ingestion, Instagram, etc. To get better in this component, please register for our free masterclass here: www.scaler.com/events
How to achieve this theorem using coding??
Take the Hunger test today: bit.ly/41fhgDW
Amazing
There might be a solution , if you take notes and replicate to your wife's diary and let a demon process took it in async mode , then you will have a higher chance of meeting consistency , but you have to make sure your deamon mode process will guarantee a job done , if the job is done you will send a notification to user , else design a failed job que and perform reprocess till you hit a counter .
I have a question wouldn't it be easy to use a central system where everything is stored. instead of storing it on their on personal diary what if they stored on a space which is shared between them in that case bot consistency and availability is possible.
What is partition tolerance?
What about CA (Consistent & Available but not network partition tolerant) & CP (Consistent & Partition Tolerant but not Available) systems, do they exist, if so how do they work?
Banking systems should be based on the CP principle whereas customer-centric companies like google & FB should be AP?
Quick question : when u tell ur wife to note it down or vice versa , Doesn't it increases the load , which hampers the sole purpose of bringing ur wife in first place i.e to handle load ( more request to process)
Plz help me out to understand
There are 2 kinds of requests. Read (where some calls and asks for relevant entries in my diary) and Write (Noting down entries in my diary). You are right. Load of writes does not reduce. However, read traffic is distributed between me and my wife.
Most systems in the world are read heavy and hence this is okay. However, if you want to reduce write traffic, you will have to shard / split (I can only process users with names starting with a-m, while my wife handles everyone with their names starting from n-z).
You guide is for college kids
worst example you took and didn't make clearance with the concepts
THIS WAS BRILLIANT
Nice explanation