What is CAP Theorem?

Continuing our System Design series, in this illustrated video we talk about the CAP theorem. We will learn what the CAP theorem is, what it means for a distributed system to be Consistent, Available, and Partition Tolerant. Also, we will look at some examples to get the proof of the CAP theorem. After watching this video, you will be able to answer the below questions:
- What is CAP Theorem?
- What is a Distributed System?
- What is Consistency, Availability and Partition Tolerance?
- Proof of CAP theorem.
Resources:
github.com/psy...
codahale.com/y...
System Design Series:
• System Design 101
Channel website:
- roadmap.sh
Discord Community:
- / discord
Find us on the internet:
- / roadmapsh
- github.com/kam...
- / roadmapsh

[@omshreepal9174]

the level of example and explanation you gave ,🎩 off , sir....

[@Best-Regions]

Best and most visual explanation I've seen. Thank you

[@cyclingdude7018]

Thanks Kamran! Love how you explain everything building up to the point without leaving us confused in the end.

[@jasper5016]

Always had difficulty understanding this concept. Thanks so much.

[@cartrivia8379]

Most important points at the last, In distributed system we can't avoid partition (network failure) , so system should be always partition tolerance, so now choice is always between either consistency or availability.

[@onlycodingjeet]

great explanation. Thank you

[@mearn_ctaftsman]

Crystal clear sir 🙏

[@RetroFrequency]

Great explanation!

[@sparkestic1238]

Great explanation

[@surajitpradhan446]

Thank you so much

[@JoyWithShorts]

When to choose Consistency and when to choose Availability?

[@anshulnagpal1479]

In case of "Partition Tolerance"
Cannot we achieve Both Consistency and Availability by having redundant connections in between System S1 and System S2. One setting could be having another System S3 that connects both S1 and S2. That way Both Systems are always connected either directly or indirectly.
In fact in real world applications, we have Distributed Systems with more than 2 Nodes.

[@roadmapsh]

It doesn't really matter how you connect the nodes. Partition is not only specific to network issues but the failure of node S2 is also going to be considered as partition - what if S2 itself is dead and can't receive or process any messages; it doesn't really matter how the nodes are connected in that case.
You might find this article interesting:
codahale.com/you-cant-sacrifice-partition-tolerance

[@NguyenTuan-ek1pv]

Good video

[@mherify]

So what is the solution to this problem?

[@nsandeep7574]

smooth!

[@anshulnagpal1479]

BTW, What Software do you use for making dynamic presentations ?

[@roadmapsh]

It's Keynote

[@kamakshijayaraman3747]

thank you

[@sagarmehta3515]

Great explanation