---
title: "Data Stream glossary"
slug: "datastream-glossary"
tags: ["Data Stream", "glossary"]
updated: 2026-09-17T09:18:03Z
published: 2026-09-17T09:18:03Z
canonical: "guide.ncloud-docs.com/datastream-glossary"
---

> ## Documentation Index
> Fetch the complete documentation index at: https://guide.ncloud-docs.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Data Stream glossary

Available in VPC

Here are some must-know terms for use of the Data Stream service. The glossary and its description are as follows:

### Value

The actual data of the message sent by the producer, which can be stored in various formats such as JSON, XML, and text.

### Lag

The difference between the message offset sent by the producer and the offset consumed by the consumer.

### Message

The basic unit of data passed between systems. Usually comprised of a key, value, offset, time stamp, header, etc.

### Broker

A server that is part of a Kafka cluster. Stores messages sent by producers and provides messages requested by consumers.

### Offset

A unique identifier indicating where in the partition the message is stored. Messages are stored sequentially within each partition, with offsets starting from 0 and increasing.

### Kafka Connect

A component that connects systems such as databases, key-value stores, and file systems to Kafka.

### Consumer

A client application that reads and processes messages stored in a topic.

### Consumer Group

A unit where multiple consumers are grouped together and process messages in parallel. Each consumer distributes the load by reading and processing messages from different partitions.

### Cluster

A collection of multiple brokers. Used for distributed storage of data and load balancing.

### Key

A unique identifier for messages. Used to distinguish messages. Messages with the same key are stored in the same partition.

### Timestamp

Time information used to record the time a message was created or processed.

### Topic

A unit for storing messages. It represents a category or feed of messages, and is the unit where producers publish messages and consumers subscribe to messages.

### Partition

A unit for dividing topics. Topics divided into multiple partitions have distributed storage and can be processed in parallel.

### Producer

A client application that creates messages in a topic and sends them to a broker.

### Header

A field that contains additional metadata, such as properties or processing information, about a message.

              Note

              

For more information on terms, see [Glossary](https://www.ncloud.com/guideCenter/glossary) on the NAVER Cloud Platform portal.