SPARK-CYPHER: HOW WE WILL BE DOING GRAPH ANALYTICS FROM 2020 ONWARDS
summary
Spark 3 promises up to 17× faster execution (Alibaba’s GTPc benchmarks) and built‑in features like adaptive execution and GPU support, but the planned unified Spark‑Cypher API is feature‑frozen and absent from 3.0. Current options are: low‑level RDD GraphX (official but dated), DataFrame‑based GraphFrames (external, Python support), and Morpheus implements openCypher on top of dataframes to run Cypher queries in Spark. The result: developers must still stitch GraphX/GraphFrames/Morpheus together for scalable graph analytics.
AI-written summary · about
chapters (20)
- 0:00Intro
- 0:35Brief History of Spark
- 1:39Spark is fast in multiple ways
- 3:36Spark 3 is coming
- 5:18Graphs: Common usages
- 6:16Common Graph Algorithms
- 8:52What is a graph?
- 15:14Graphs vs Tables: Graphs
- 16:31Graph Databases
- 18:40Cypher Query Language
- 23:16Spark Graph X & the other side of graphs
- 29:38GraphX: The Good
- 31:09GraphX: The bad & the ugly
- 32:34GraphFrames
- 38:08Morpheus: Reading inputs
- 38:22Morpheus: Building a Graph
- 38:38Morpheus: Cypher Queries
- 39:36Spark - Cypher
- 40:42Spark Graph-Based Trolling
- 41:46Summary: How will we be doing graph analytics?
related talks
19mWhat's Next for Apache Spark™ Including the Upcoming Release of Apache Spark 4.0
38mA Tale of Two Graph Engines on Spark
47mThe future of data: Apache Spark 3.1, Delta Lake and Koalas
55mUse Spark from Anywhere
30mIntroducing Spark Cyclone Accelerating Spark with...
11mThe Evolution of Apache Spark™
17mBringing Cypher to Apache Flink
52mSolving Analytical Problems using Apache Spark