Data Consistency Levels

2 minute read

Measure of how *up-to-date a piece of data is in a distributed systems with geographically distributed nodes.

Consistency Levels

EveryRead operation results in the MOST up-to-date data
To achieve this a single-threaded single server may be used.So every read and write request will always be ordered
Every request is processed linearly one by one
- Others wait while one request gets processed
- If the processing request fails or takes a long time or sleeps, every other request has to wait (Head-of-line Blocking)
- Single point of failure
- High Latency - Time taken to process the request successfully.
- Low Availability - Chance of a system to be able to respond to a request
- High Consistency -

we can send stale data for a read request,but eventually,return the latest data after updates are done
Until we are returning the stale data our system is not consideredconsistent but we can guarantee that after some time it will be consistent
To achieve this we can process read and write requests parallelly (using multiple servers) or concurrently (using multiple threads)
Multiple servers can send requests to a DB at the same time
- If WRITE (then)-> READ is expected and both requests went from 2 different instances of a service,
  - may be possible that Read goes in first and reads the data before writing it
Same with email - if the mail is in the outbox, eventually it will be sent

if a previous operation is related to the current operation then the previous operation has to be executed before the current operation
Extracts the requests based on the keys and order sequentially and perform operation
Causal consistency is stronger and slower than eventual consistency because operations for the same key are processed sequentially.
Fails on Aggregation
- causal consistency orders query based on IDs but aggregation operations use all IDs

At this consistency level, we have multiple replicas of the database and these replicas may or may not be in sync.
For a read query, we get the data from all the replicas and return the mostappropriate values (Majority value, Latest updated value etc).
It works on some kind of consensus in the distributed system.
Eventually consistent
Distributed Consensus
It provides fault tolerance and depending upon the values R, W and N we can either have an
- eventually consistent system (R + W <= N) or
- a strongly consistent system (R + W > N)

A system can be strongly consistent by specifying the minimum number of replicas it needs to read the data from.

R + W > N where,

Consistency Level	Consistency	Efficiency
Linearizable/Serializable	Highest	Lowest
Causal Consistency	Linealizable > Causal > Eventual	Linealizable < Causal < Eventual
Quorun	Configurable	Configurable
Eventual Consistency	Lowest	Highest

Diodes