Potpourri

In this blog, I'll try to put together the Neo4j topics to read and the resources for it. I'll hope that it can act as a trigger to learn Neo4j and to do a quick recap as and when desired.

1. Graph and Graph Theory
• History
• 7 Bridges of Konigsberg by Euler
• Eulerian path
• Field of Study
• Social Science: Interaction, influence and idea sharing between people.
• Biological: graphs describe metabolic pathways. Link
• Computer Science: Path finding algorithms to analyze effect of change in design of artifacts.
• Flow: Flow network, Maximum flow problem
• Route Problems: Hamiltonian path problem, Route inspection problem, Shortest-path problem, Travelling salesman problem, dijkstra, A*
2. Graph Database
• online DBMS with CRUD operations working on graph data model.
• Generally built for OLTP systems.
• Engineered with transactional integrity and operational availability in mind.
• Properties
• Graph storage: 
• native storage: defined to store and manage graph.
• relational or OO storage. It is obviously slower.
• Graph Processing Engine: Native graph processing a.k.a. index free adjacency is the most efficient way to process graphs and nodes physically point to each other.
• Advantages
• Minutes-to-millisecond performance.
• Accelerated development cycles.
• Extreme business responsiveness.
• Enterprise ready (ACID, availability, horizontal read scalability, Storage of billion entities)
• Common Use Cases
• Fraud Detection, Real-time recommendation engines, Master Data Management, Identity and Access Management, Graph based search
• Where not to use?
• Large set-oriented queries - RDBMS is better.
• Simple aggregate-oriented queries - Document database is better.
3. Neo Databases
• Network-oriented (ordered in complex n/w and deep trees) and semi-structured data.
• Neo is an embedded persistence engine.
4. Installation and Getting Started
• http://neo4j.com/docs/stable/server-installation.html
5. Data Model http://neo4j.com/developer/guide-data-modeling/
• Best Practices
• Design for query ability.
• as an employee, I want to know who in the company I work for has similar skills to me so that we can exchange knowledge
• Align relationship with use cases.
• Look for n-ary relationship.
• Granulate nodes.
• Use in-graph indexes when appropriate
• Pitfalls
• Rich properties
• Node representing multiple concepts e.g. country, language and currency.
• Unconnected graph.
• Dense node pattern. - Madonna and her fans problem.
6. Cypher Query language
• This is a vast topic in itself.
• Tried to cover more of it at http://www.i-satyam.blogspot.in/2016/03/neo4j-cypher-query-language.html
7. References
• Capabilities
• Data Security: Neo4j does not deal with data encryption explicitly, but supports all means built into the Java programming language and the JVM to protect data by encrypting it before storing.
• Data Integrity: transactional architecture ensures that data is protected and provides for fast recovery from an unexpected failure.
• Data Integration:Event based synchronization, Periodic synchronization, Periodic full export/import data.
• Availability and Reliability: Cold Spare, Hot Spare, High Availability Cluster
• Capacity: File Size, Read Speed, Write Speed, Data Size
• Transaction Management
• read-committed isolation level
• Neo4j Java API enables explicit locking of nodes and relationships which gives the opportunity to simulate the effects of higher levels of isolation by obtaining and releasing locks explicitly.
• Default Locking Behavior:
• When adding, changing or removing a property on a node or relationship a write lock will be taken on the specific node or relationship.
• When creating or deleting a node a write lock will be taken for the specific node.
• When creating or deleting a relationship a write lock will be taken on the specific relationship and both its nodes.
• Handling Deadlock
• TransactionTemplate class

• TransactionTemplate template = new TransactionTemplate(  ).retries( 5 ).backoff( 3, TimeUnit.SECONDS );

• We can also use our own retry-loop code.
• Creating unique nodes
• Single Threaded Environment ensures it.
• Unique constraints and cypher can also help with this.
• Uniqueness is guaranteed by using a legacy index in case of putIfAbsent.