メインコンテンツまでスキップ
バージョン: 3.13

Glossary

This glossary includes database and distributed-system terms that are often used when using ScalarDB.

ACID

Atomicity, consistency, isolation, and durability (ACID) is a set of properties that ensure database transactions are processed reliably, maintaining integrity even in cases of errors or system failures.

concurrency control

Concurrency control in databases ensures that multiple transactions can occur simultaneously without causing data inconsistency, usually through mechanisms like locking or timestamp ordering.

consensus

Consensus in distributed systems refers to the process of achieving agreement among multiple computers or nodes on a single data value or system state.

data federation

Data federation is the process of integrating data from different sources without moving the data, creating a unified view for querying and analysis.

data mesh

A data mesh is a decentralized data architecture that enables domain teams to independently perform cross-domain analysis, addressing scalability and ownership challenges.

data virtualization

Data virtualization is similar to data federation in many aspects, meaning that it virtualizes multiple data sources into a unified view, simplifying queries without moving the data.

database anomalies

Database anomalies are inconsistencies or errors in data that can occur when operations such as insertions, updates, or deletions are performed without proper transaction management.

federation engine

A federation engine facilitates data integration and querying across multiple disparate data sources, often as part of a data federation architecture.

global transaction

A global transaction spans multiple databases or distributed systems and ensures that all involved systems commit or roll back changes as a single unit.

heterogeneous databases

Heterogeneous databases refer to systems composed of different database technologies that may have distinct data models, query languages, and transaction mechanisms.

HTAP

Hybrid transactional/analytical processing (HTAP) refers to a system that can handle both transactional and analytical workloads concurrently on the same data set, removing the need for separate databases.

JDBC

Java Database Connectivity (JDBC) is an API that allows Java applications to interact with databases, providing methods for querying and updating data in relational databases.

linearizability

Linearizability is a strong consistency model in distributed systems where operations appear to occur atomically in some order, and each operation takes effect between its start and end.

NoSQL databases

NoSQL databases are non-relational databases designed for specific data models, such as document, key-value, wide-column, or graph stores, often used for handling large-scale, distributed data.

Paxos

Paxos is a family of protocols used in distributed systems to achieve consensus, even in the presence of node failures.

PITR

Point-in-time recovery (PITR) allows a database to be restored to a previous state at any specific time, usually after an unintended event like data corruption.

polystores

Polystores are database architectures that allow users to interact with multiple, heterogeneous data stores, each optimized for a specific workload or data type, as if they were a single system.

read-committed isolation

Read-committed isolation is an isolation level where each transaction sees only committed data, preventing dirty reads but allowing non-repeatable reads.

relational databases

Relational databases store data in tables with rows and columns, using a structured query language (SQL) to define, query, and manipulate the data.

replication

Replication in databases involves copying and distributing data across multiple machines or locations to ensure reliability, availability, and fault tolerance.

Saga

The Saga pattern is a method for managing long-running transactions in a distributed system, where each operation in the transaction is followed by a compensating action in case of failure.

serializable isolation

Serializable isolation (serializability) is the highest isolation level in transactional systems, ensuring that the outcome of concurrently executed transactions is the same as if they were executed sequentially.

snapshot isolation

Snapshot isolation is an isolation level that allows transactions to read a consistent snapshot of the database, protecting them from seeing changes made by other transactions until they complete.

TCC

Try-Confirm/Cancel (TCC) is a pattern for distributed transactions that splits an operation into three steps, allowing for coordination and recovery across multiple systems.

transaction

A transaction in databases is a sequence of operations treated as a single logical unit of work, ensuring consistency and integrity, typically conforming to ACID properties.

transaction manager

A transaction manager coordinates the execution of transactions across multiple systems or databases, ensuring that all steps of the transaction succeed or fail as a unit.

two-phase commit

Two-phase commit is a protocol for ensuring all participants in a distributed transaction either commit or roll back the transaction, ensuring consistency across systems.