Data Across Machines
Understand copies, partitions, coordination and the limits of promises made over a network.
Replication: More Than One Copy
A second copy of a database can help answer questions, survive some failures or move data closer to readers. It also creates a new question: which copy…
3,831 words/17 min read34Leaders, Followers and Failover
A leader is not simply the machine with the newest-looking screen. It is a participant currently authorized by a defined protocol to coordinate…
3,693 words/16 min read35Partitioning: Dividing the Work
Partitioning divides a collection into pieces. It can make some operations smaller, distribute capacity or align storage with the way records are…
3,701 words/16 min read36Consistency, Availability and Network Partitions
“The system is consistent” is incomplete unless we state which observations are allowed. A database can obey its local constraints while different…
3,753 words/16 min read37Consensus: Agreeing Despite Failure
Consensus helps participants agree on a decision despite specified failures. In a replicated-log system, the decisions determine which commands occupy…
3,678 words/16 min read38Distributed Transactions and Sagas
A local transaction can protect changes inside one database boundary. A workflow touching several independent resources needs an additional agreement or…
3,642 words/16 min read39Caches, Queues and What They Do Not Guarantee
Caches reduce repeated work by keeping derived copies. Queues separate the arrival of work from its processing. Both can improve a system while also…
3,717 words/16 min read