Kinesis를 활용한 실시간 처리

In today's fast-paced world, many applications need to react to data instantly. This is known as real-time data processing.

• Immediate insights: Analyze data as it arrives.
• Quick responses: Trigger actions without delay.
• Enhanced user experience: Personalize content on the fly.

Think of fraud detection, live dashboards, or IoT device monitoring.

Amazon Kinesis is a powerful suite of services designed for collecting, processing, and analyzing streaming data in real-time. It helps you handle massive volumes of data efficiently.

Key Kinesis services include:

• Kinesis Data Streams (KDS): For custom applications needing granular control.
• Kinesis Firehose: For easy delivery to data stores like S3 or Redshift.

Kinesis Data Streams (KDS) is like a continuous pipeline for high-throughput data. It captures data from various sources and makes it available for processing by different applications.

KDS organizes data into shards. A shard is a base unit of throughput. Data producers write to shards, and data consumers read from them.

Applications send data to KDS using operations like PutRecord or PutRecords. Each record includes a partition key, which KDS uses to group data and route it to a specific shard.

A good partition key ensures even distribution of data across shards, preventing hot spots and maximizing throughput.

AWS Lambda functions are excellent consumers for Kinesis Data Streams. You can configure a Lambda event source mapping to automatically invoke your function whenever new data records are available in a KDS stream.

Lambda polls the stream, reads records in batches, and passes them to your function for processing. This makes building real-time processors very efficient.

Here's a simple Python Lambda function that processes records from a Kinesis Data Stream. It iterates through the records in the event and prints their data.

Try running this example:

Kinesis Firehose is a fully managed service for delivering real-time streaming data to destinations like Amazon S3, Amazon Redshift, Amazon OpenSearch Service, or HTTP endpoints.

Unlike KDS, Firehose requires almost no administration. You simply create a delivery stream, specify your source and destination, and Firehose handles all the scaling, buffering, and delivery.

Firehose is designed for simplified data delivery. It automatically batches, compresses, and encrypts data before sending it to your chosen destination.

Common destinations include:

• Amazon S3: For long-term storage and data lakes.
• Amazon Redshift: For data warehousing and analytics.
• Amazon OpenSearch Service: For logging and search.
• HTTP endpoints: For custom integrations.

Kinesis Firehose can integrate with Lambda to transform incoming data before it's delivered to its final destination. This is incredibly useful for cleaning, enriching, or reformatting data on the fly.

You configure a Lambda function within your Firehose delivery stream, and Firehose invokes it for each batch of records, expecting transformed records in return.

Which statements accurately describe the differences or uses of Kinesis Data Streams (KDS) and Kinesis Firehose?

You've learned about Amazon Kinesis, a key service for real-time data processing.

• Kinesis Data Streams (KDS): Offers flexible, shard-based streaming for custom applications, often consumed by Lambda.
• Kinesis Firehose: Provides a managed solution for delivering streaming data to various destinations, with optional Lambda transformation.

These services, combined with Lambda, enable powerful event-driven architectures for handling massive data streams in real-time.