Thema der Dissertation:
Performance and Reliability Evaluation of Apache Kafka Messaging System Thema der Disputation:
The ACID guarantees in streaming data processing
Performance and Reliability Evaluation of Apache Kafka Messaging System Thema der Disputation:
The ACID guarantees in streaming data processing
Abstract: Streaming data can be sensor data transmitted across IoT devices, user activity records
from websites or financial transactions in e-commerce analysis. This type of data is generated
continuously in large volumes and needs to be processed by data analysis applications in real-time.
The traditional batch-oriented big data processing systems are insufficient for the real-time
requirements because they store data in distributed databases before processing the data, thus the
earlier arrived data has to wait for accumulating batches. Consequently, streaming data processing
systems (SDPSs) are developed to continuously process each piece of data immediately as it arrives.
This talk will introduce the streaming data processing workflow of the SDPS in comparison with
traditional batch-oriented systems. While most batch-oriented systems provide strong ACID (atomicity,
consistency, isolation, durability) using transactional databases, guaranteeing the ACID in SDPS
remains challenging: traditional transactional processing technology affects the speed of streaming
data processing. This talk will explain the differences between batch-oriented and stream-oriented
processing, and what this implies for maintaining the ACID properties.
from websites or financial transactions in e-commerce analysis. This type of data is generated
continuously in large volumes and needs to be processed by data analysis applications in real-time.
The traditional batch-oriented big data processing systems are insufficient for the real-time
requirements because they store data in distributed databases before processing the data, thus the
earlier arrived data has to wait for accumulating batches. Consequently, streaming data processing
systems (SDPSs) are developed to continuously process each piece of data immediately as it arrives.
This talk will introduce the streaming data processing workflow of the SDPS in comparison with
traditional batch-oriented systems. While most batch-oriented systems provide strong ACID (atomicity,
consistency, isolation, durability) using transactional databases, guaranteeing the ACID in SDPS
remains challenging: traditional transactional processing technology affects the speed of streaming
data processing. This talk will explain the differences between batch-oriented and stream-oriented
processing, and what this implies for maintaining the ACID properties.
Zeit & Ort
14.01.2021 | 15:00