CAP theorem in DBMS Theory - Step-by-Step Execution

Concept Flow - CAP Theorem

Start: Distributed database system

↓

CAP Theorem: Pick 2 of 3 guarantees

↓

Consistency

↓

Network partition occurs

↓

Choose: Consistency OR Availability?

↓

CP System

↓

End

This flow shows how the CAP theorem forces a choice between consistency and availability when a network partition happens in a distributed database.

Execution Sample

DBMS Theory

Distributed database has 3 nodes: Node_A, Node_B, Node_C
Client writes value X=10 to Node_A
Node_A replicates to Node_B and Node_C
Network partition splits: {Node_A, Node_B} | {Node_C}
Client reads X from Node_C
CP system: Reject read (Node_C may have stale data)
AP system: Return last known X value (may be stale)

This sequence shows how a network partition forces a distributed database to choose between consistency and availability.

Analysis Table

Step	Action	Condition/Check	Result/Decision	Next Step
1	Client writes X=10 to Node_A	N/A	Write accepted at Node_A	Replicate to other nodes
2	Node_A replicates to Node_B, Node_C	Network healthy?	Yes, all nodes updated	System fully consistent
3	Network partition occurs	Node_C cannot reach Node_A or Node_B	Partition detected	Choose CP or AP
4	Client sends read request to Node_C	Can Node_C confirm latest value?	No (partitioned from Node_A)	Apply CAP trade-off
5	CP system response	Consistency prioritized	Reject read with error	Wait for partition to heal
6	AP system response	Availability prioritized	Return last known value of X	Accept possible staleness
7	Partition heals	Nodes reconnect	Sync data across all nodes	End

💡 After partition heals, all nodes synchronize to reach eventual consistency. The CAP trade-off only applies during the partition.

State Tracker

Variable	Start	After Step 2	After Step 3	After Step 5 (CP)	After Step 6 (AP)
X at Node_A	undefined	10	10	10	10
X at Node_B	undefined	10	10	10	10
X at Node_C	undefined	10	10 (stale?)	Read rejected	Returns 10 (may be stale)
system_status	Healthy	Consistent	Partitioned	Unavailable to Node_C	Available but inconsistent

Key Insights - 3 Insights

Why can't a distributed system guarantee all three of C, A, and P simultaneously?

When a network partition (P) occurs, the system must choose: either reject requests to stay consistent (CP) or serve possibly stale data to stay available (AP). You cannot have both C and A during a partition, as shown in steps 5 and 6.

What does Partition Tolerance mean and why is it non-negotiable?

Partition tolerance means the system continues operating despite network splits between nodes. In real distributed systems, network failures are inevitable, so P is always required — making the real choice between C and A.

How do real databases handle the CAP trade-off?

CP systems like HBase and MongoDB (in strict mode) reject operations during partitions to maintain consistency. AP systems like Cassandra and DynamoDB continue serving requests and reconcile data after the partition heals.

Visual Quiz - 3 Questions

Look at the execution_table at step 5. What does a CP system do when a client reads from a partitioned node?

A Returns the latest value from the primary node

B Rejects the read request with an error

C Returns a cached value from the partitioned node

D Redirects the client to another node

💡 Hint

Check the 'Result/Decision' column in step 5 where consistency is prioritized.

According to the variable_tracker, what happens to X at Node_C after an AP system responds during a partition?

A It is deleted

B It returns the confirmed latest value

C It returns a value that may be stale

D It blocks until the partition heals

💡 Hint

Look at the 'After Step 6 (AP)' column for Node_C.

Why is partition tolerance considered non-negotiable in distributed systems?

A It is the easiest guarantee to implement

B Network failures are inevitable in distributed systems

C Clients always prefer partition tolerance over consistency

D Databases cannot function without partitions

💡 Hint

Think about what happens in real networks spanning multiple data centers.

Concept Snapshot

CAP theorem: distributed systems can guarantee only 2 of 3 properties.
Consistency: all nodes see the same data at the same time.
Availability: every request gets a non-error response.
Partition Tolerance: system works despite network splits.
Since partitions are inevitable, the real choice is C vs A.

Full Transcript

This visual execution shows the CAP theorem in action. A distributed database with three nodes accepts a write and replicates it across all nodes. When a network partition splits Node_C from the others, the system faces a choice. A CP system rejects reads from Node_C to maintain consistency, while an AP system returns the last known value to stay available, accepting possible staleness. Once the partition heals, nodes synchronize to restore consistency. The key insight is that since network partitions are inevitable in distributed systems, the real trade-off is always between consistency and availability.