💻 IT & Coding concepts,
clear in 60 seconds.

Elasticsearch uses master election and quorum to prevent split brain scenarios.

gRPC allows efficient communication between services using a structured request-response model.

RPC allows one service to call a function in another service over the network.

Causality in distributed systems helps understand how events influence each other, beyond just timestamps.

Physical time tells when events happen, while logical time tells their order.

Network partitions occur when servers are operational but can't communicate, affecting system performance.

ZooKeeper watches notify clients of changes, avoiding constant polling and reducing server load.

Crash failures stop a service, while network failures disrupt communication.

Fail-stop means a system stops and stays down, while fail-recover means it can come back but needs to be ready.

ZooKeeper uses ephemeral and sequential nodes for managing distributed services effectively.

Raft and Paxos are protocols that help computers agree on values in distributed systems.

Paxos is a protocol that helps distributed systems agree on one value despite failures.

Quorum and majority help distributed systems maintain data consistency and availability.

Raft's safety rules guarantee that committed entries are preserved even after leader failures.

Raft handles leader failure by allowing followers to elect a new leader through voting.

Raft consensus helps distributed systems elect leaders and replicate logs reliably.

matchIndex tracks what each follower has, while commitIndex shows what is committed.

Leader election helps distributed systems select one coordinator from many nodes.

Bully, Raft, and ZooKeeper are distinct algorithms for leader election in distributed systems.

Kafka Streams uses local state for quick data access and changelogs for recovery.

Kafka uses state stores to remember previous events in stateful stream processing.

Kafka consumers resume processing from the last committed offset after a crash.

Kafka rebalance can lead to lost or duplicated work if not handled carefully.

Kafka's assign() method allows manual control of partition assignments for consumers.