Exam AZ-204 topic 3 question 44 discussion

Strong consistency offers a linearizability guarantee. Linearizability refers to serving requests concurrently. The reads are guaranteed to return the most recent committed version of an item. A client never sees an uncommitted or partial write. Users are always guaranteed to read the latest committed write. Source: https://learn.microsoft.com/en-us/azure/cosmos-db/consistency-levels

I believe its C 'bounded staleness' Because strong consistency is not supported for multiple regions

This page explains it pretty well: https://learn.microsoft.com/en-us/azure/cosmos-db/consistency-levels#strong-consistency It shall be strong, as the write is only commited once it's replicated to all regions, and therefore all reading parties (from all regions) read the last inserted value (a tradeoff between availability + throughput and consistency)

D. strong: Strong consistency guarantees that reads always return the most recent committed version of an item. It also provides linearizability, which ensures that all operations are seen in a strict order. This meets both requirements of returning the most recent committed version and providing ordering guarantees.

There is a trick here - the SOLUTION is deployed to multiple regions. That only contains the function app. Nowhere does it say that the COSMOS db is in multiple regions. Therefore, Strong is perfectly acceptable.

D is correct .According to the requirements that item results from Azure Cosmos DB must return the most recent committed version of an item and items written to Azure Cosmos DB must provide ordering guarantees, the Strong consistency level should be used. Strong consistency offers linearizability, which is the highest consistency level available. It guarantees that once an operation completes, the result will be immediately visible to all subsequent operations. This ensures that the most recent committed version of an item is always returned and also provides ordering guarantees, but this comes with a trade-off as strong consistency may introduce higher latencies and reduced availability compared to other consistency levels.

"Most recent"

The problem doesn't specify any condition (versions or time lag) which may allow bounded staleness. Only Strong consistency guarantees the reading the more recent commit.

Item results from Azure Cosmos DB must return the most recent committed version of an item: Strong consistency ensures that once a write operation is acknowledged, all subsequent read operations will return the most recent committed version of the data. This means that any read operation will see the effects of the latest write operation. Items written to Azure Cosmos DB must provide ordering guarantees: Strong consistency also ensures ordering guarantees. It means that if one write operation is acknowledged before another, then the order of these operations is preserved for all subsequent read operations.

Strong - doesnt mention multi-region writes - the multi-region functions are triggered hence reads

The key is the multiple regions

Azure Cosmos DB accounts configured with multiple write regions can't be configured for strong consistency as it isn't possible for a distributed system to provide an RPO of zero and an RTO of zero. Additionally, there are no write latency benefits on using strong consistency with multiple write regions because a write to any region must be replicated and committed to all configured regions within the account. This scenario results in the same write latency as a single write region account. https://learn.microsoft.com/en-us/azure/cosmos-db/consistency-levels#strong-consistency-and-multiple-write-regions

strong is correct ans.: Strong consistency: Strong consistency offers a linearizability guarantee. Linearizability refers to serving requests concurrently. The reads are guaranteed to return the most recent committed version of an item. A client never sees an uncommitted or partial write. Users are always guaranteed to read the latest committed write. Bounded staleness consistency: With Bounded Staleness consistency, reads issued against a non-primary region may not necessarily return the most recent version of the data globally, but are guaranteed to return the most recent version of the data in that region, which will be within the maximum staleness boundary globally. https://learn.microsoft.com/en-us/azure/cosmos-db/consistency-levels

I think it's D. Strong consistency is not supported for multiple WRITE regions, and I don't see any such requirements in the question. https://learn.microsoft.com/en-us/azure/cosmos-db/high-availability#multiple-write-regions

I think it's Bounded Staleness because it's not mentioned that read is executed right after write operation so I think it's acceptable to have small latency between write and read operations. I would go with this!

But for Strong consistency, it also says Bounded Staleness in a multi-write account is an anti-pattern. So I really don't understand what the correct answer is. Bounded Staleness or Strong?

C. bounded staleness Item results return the most recent committed version of an item: Bounded staleness ensures that reads are served from a version of the data that is within a specified lag from the current write operation. This means that you can read the most recent committed version of an item while still providing a level of consistency. Items written to Azure Cosmos DB must provide ordering guarantees: Bounded staleness guarantees that writes are ordered and committed before being read. While other consistency levels like "Strong" and "Consistent prefix" also provide strong consistency, "Bounded staleness" offers a balance between consistency and availability, which is suitable for scenarios where the most recent data is needed with some level of delay.