Graph Applications: Big Data and Scientific Applications
3
Definitions
4
GraphOne: Background and Architecture
5
Hybrid Store Details
6
Hybrid Store: Optimizing the Archiving Phase
7
Hybrid Store: Optimizing Memory Usage
8
Results: Hybrid Store Composition and Ingestion Rate
9
Results: Dynamic Graph Systems
10
Results: Cost of Data Management Over Static Graph Engine
11
GraphOne: Conclusion
Description:
Explore a 27-minute conference talk from FAST '19 on GraphOne, a data store designed for real-time analytics on evolving graphs. Learn about the challenges of supporting diverse data access while ingesting high-velocity updates, and discover how GraphOne addresses these issues through a combination of complementary graph storage formats and dual versioning. Understand the innovative GraphView data abstraction that enables efficient data access at different granularities. Examine experimental results demonstrating GraphOne's superior ingestion rate and analytics performance compared to existing graph databases and dynamic graph systems. Delve into topics such as hybrid store composition, optimization techniques for archiving and memory usage, and the overall architecture of GraphOne. Gain insights into the growing importance of real-time graph analytics in big data and scientific applications.
GraphOne - A Data Store for Real-time Analytics on Evolving Graphs