In a primary-replica replication system, what is the most significant challenge when attempting to achieve linearizability across all reads and writes, particularly under high write contention and network partitions?

Maintaining a consistent global order of operations across geographically dispersed replicas.

Ensuring all replicas are always online and synchronized without any lag.

The inherent latency in committing writes to the primary before propagating to replicas.

The complexity of dynamic primary election and failover mechanisms during network splits.

A distributed database uses primary-replica replication. If the primary node experiences a sudden, unrecoverable failure and a replica is promoted, which of the following scenarios is most likely to result in a 'split-brain' situation where data consistency is severely compromised?

The failed primary node recovers and attempts to resume its primary role without coordination.

The promoted replica successfully becomes the new primary and processes new writes.

Other replicas continue to apply old writes from the original primary before receiving notification of the new primary.

A network partition isolates the new primary from some replicas, but the old primary remains offline.

Consider a peer-to-peer replication system designed for eventual consistency with conflict resolution occurring at read time. What is the primary difficulty in guaranteeing 'read-your-writes' consistency for a client that performs rapid successive writes and reads to the same data item?

The possibility of different peer nodes having different versions of the data immediately after a write.

The inherent network latency between the client and different peer nodes.

The overhead of vector clocks or similar mechanisms to track causality.

The difficulty in efficiently propagating conflict resolution logic to all client applications.

Which scenario most strongly favors a primary-replica replication model over a peer-to-peer model when the primary requirement is strong consistency and minimal developer effort for conflict resolution?

An application where data consistency must always reflect the most recent committed write globally, even if it impacts write availability.

A geographically distributed system with high availability requirements and frequent network partitions.

A system requiring high write throughput across multiple active data centers with low read latency.

A collaborative editing application where concurrent modifications are frequent and users expect immediate feedback on their own changes.

When migrating from a single-node database to a primary-replica replicated setup, which architectural change presents the most significant challenge in maintaining application correctness without extensive code refactoring?

Updating the application to direct all write operations exclusively to the primary node.

Configuring the network firewall to allow communication between primary and replica nodes.

Ensuring that read operations are load-balanced across all replicas to offload the primary.

Implementing a robust monitoring system to track replication lag and node health.

In a peer-to-peer replication system using vector clocks for conflict detection, a conflict is detected when two concurrent writes modify the same data item on different nodes. What is the most complex aspect of automatically resolving such a conflict algorithmically without human intervention?

Defining a universally applicable, semantic conflict resolution logic for arbitrary data types.

Determining which write occurred chronologically first based on system timestamps.

Ensuring the chosen resolution strategy is idempotent and always produces the same outcome.

Efficiently propagating the conflict resolution decision to all involved nodes in the cluster.

Why is it generally harder to implement transactional guarantees (like ACID properties) across multiple data centers in a peer-to-peer replication model compared to a primary-replica model?

The absence of a single point of serialization for writes necessitates complex distributed consensus for atomicity.

Peer-to-peer systems inherently prioritize availability over consistency, making strong transaction guarantees difficult.

Network latency between geographically dispersed peers makes two-phase commit protocols prohibitively slow.

Conflict resolution algorithms used in peer-to-peer systems often break ACID properties during merge operations.

A company requires a database system that can withstand the failure of an entire data center without any data loss and with minimal downtime. Which replication strategy, if properly implemented, offers the best chance of meeting the 'no data loss' requirement under this severe failure condition, assuming synchronous replication is employed?

Synchronous primary-replica replication with a quorum of replicas across multiple data centers.

Asynchronous primary-replica replication to a single hot standby in another data center.

Peer-to-peer eventual consistency replication across multiple data centers.

A hybrid model combining asynchronous primary-replica for local replicas and peer-to-peer for disaster recovery.

What is the most challenging aspect of managing schema changes (e.g., adding a column) in a large-scale, globally distributed peer-to-peer replicated database with active writes on all nodes, compared to a primary-replica system?

The potential for schema conflicts if different nodes attempt divergent schema modifications concurrently.

Coordinating the schema change across all nodes simultaneously to prevent downtime.

Ensuring backward and forward compatibility of data during the transition period.

Propagating the schema change definition efficiently to all nodes without interrupting ongoing operations.

Which statement accurately describes a key disadvantage of peer-to-peer replication when strong consistency is a paramount concern?

It generally has higher write latency than primary-replica due to distributed coordination.

It struggles with providing read scalability due to eventual consistency models.

It inherently leads to 'split-brain' scenarios more frequently than primary-replica systems.

It necessitates complex conflict resolution mechanisms that can compromise data integrity.

A primary-replica replication system experiences a network partition where the primary node becomes isolated from all replicas. What is the most appropriate action to maintain data integrity and availability during such a partition, assuming the replicas are in the larger partition?

Attempt to failover to a replica only if a majority of replicas agree the primary is unreachable and the old primary is 'fenced'.

Immediately promote a replica to be the new primary and continue processing writes.

Wait indefinitely for the primary to reconnect before allowing any further writes.

Block all write operations until the network partition is resolved and the primary is reachable by replicas.

In a distributed system using peer-to-peer replication, what is the most significant benefit of implementing a CRDT (Conflict-free Replicated Data Type) for data synchronization?

It allows concurrent updates on different replicas to automatically converge to a consistent state without external resolution logic.

It guarantees strong consistency (e.g., linearizability) without any read latency.

It simplifies read operations by eliminating the need for vector clocks or timestamps.

It ensures that all replicas always have the exact same version of data at any given point in time.

What is the primary reason that a primary-replica replication system, using asynchronous replication, might be preferred over a synchronous peer-to-peer system for an application with extremely high write throughput requirements across multiple geographically dispersed data centers?

Synchronous peer-to-peer requires a distributed commit protocol which adds significant write latency.

Asynchronous primary-replica offers inherently stronger consistency guarantees.

Asynchronous primary-replica avoids network latency for committing writes on the primary.

Peer-to-peer systems are more complex to operate and scale for write-heavy workloads.

Consider a distributed application where each user has a unique ID, and updates to a user's profile are frequent and originate from various geographical locations. If a primary-replica model is used, what is the most significant scaling challenge for write operations as the number of users and update frequency increase globally?

The difficulty in horizontally scaling the primary node itself to handle all write requests.

Managing the replication lag between the primary and numerous globally distributed replicas.

The overhead of electing a new primary node during frequent primary failures.

Ensuring low read latency for users accessing data from distant replicas.

In the context of a globally distributed database, which statement best characterizes the trade-off between strict serializability and availability during a network partition, specific to primary-replica vs. peer-to-peer architectures?

Primary-replica systems must sacrifice write availability in the smaller partition to maintain strict serializability, while peer-to-peer may sacrifice strict serializability in both partitions to maintain write availability.

Primary-replica inherently favors availability by allowing replicas to serve stale reads during partitions, unlike peer-to-peer.

Peer-to-peer with CRDTs maintains higher write availability during partitions but sacrifices strict serializability.

Strict serializability is easily achieved by both models during partitions, but at the cost of significantly increased latency.

primary replication vs peer-to-peer replication

primary replication vs peer-to-peer replication - Test Paper