2. What is a Graph Database System
Unlike RDBMS, it uses graphs to represent data.
Data is stored using a collection of nodes and edges.
Graph Database use graph theory to solve graph problems
Examples :Horton, Neo4j,Graphbase,Titan, OrientDB , VertexDB, etc .
Graph databases are very fast in execution of complex pattern matching
queries.
3. World's first Graph Database System .
Uses cypher ,a declarative graph querying language .
It’s highly scalable ( can store 32 billion edges,nodes and relationships)
Runs on Java Virtual Machine
Architecture :
4.
5. Multi-Mode Open Source NoSQL DBMS
It uses SQL with some extensions as a query language
Document databases store information in documents like JSON or XML
Uses an HTTP REST API to access/ edit the database
Runs on Java Virtual Machine
6. BENCHMARKING
Metrics
Clustering Workload
Convergence time for modularity optimization using Louvain Method
Massive Insertion
Time for the creation of the whole graph when populated with massive data
Single Insertion Workload
Time taken for to upload a block, which consist of one thousand edges and
vertices
Query Workload
Time taken to find neighbours of all the nodes
Dataset
Movies dataset
Amazon Dataset
Youtube Dataset
LiveJournal dataset
7. Modularity in a graph Using Louvain Method
● Modularity
○ Measure of a structure of graph
○ Fraction of the #edges inside of a cluster to the #edges outside
○ Ranges between [-½,1]
● Louvain Method
○ Greedy optimization method
○ Performed in two steps
9. Conclusion
Neo4j does not support multi master
replication and Native HTTP REST/JSON
Both support ACID Transaction
OrientDB supports server-side functions
OrientDB provides a better graph editing tool
Notas do Editor
Combines the power of graph and document into one scalable high performance Database