Stardog 3.0:
A Deep Dive
Michael Grove & Evren Sirin
About Us
- Founded in 2005; offices in Washington DC & Boston
- Customers in US Gov't, banking/financial, energy, health/bio, retail
- Strong academic partnerships in US, UK, Europe, and Mexico
- Expertise in
- Information Integration, Expertise Location, Policy Management, Enterprise Decision Support, Application Development
- OWL/RDF/SPARQL/SWRL
Use Cases
Customer 360
- Unify customer information: integrate all data about a customer as it's discovered
- Past, Present, and Future
- Pull from a variety of sources, including unstructured, many of which are non-relational
- Take advantage of flexible nature of semantic technology
Data Provenance
- Capture the who-what-when-where-why-how of data throughout its lifecycle
-
Utilize this information to enable data governance and regulatory compliance
- Annotate data as it comes in; continuously updated
- W3C spec dedicated to this: PROV
Reference Data
- Create a 'gold standard' for names, labels and identities
- Represent core industry terms and concepts
- Ties into Data Provenance
- Modeling complex relationships between entities can be trivial using semantic technology
- FIBO is a great example
Compliance
- Reduce compliance efforts to query answering and graph analytics
- Legal regulations are complex
- And tracking related policies is a time consuming job
- Cost of implementation high, cost of a failure catastrophic
- Utilize reasoning & rules
- Express regulations and policies as complex relationships
- Workflows and compliance checking can be performed by a reasoner
- Automated compliance analysis with explanations
Analytics & Decision Support
- Empower human decision making with contextualized, relevant information
- There is a lot of value in querying structured information automatically extracted from unstructured data
- As you build up structured corpus from data sources, you create actionable information
- Sift through the data to find the facts so a human can make decisions more quickly and easily
What's the Common Thread?
- All information integration problems
- i.e. not really financial services problems
- So how do you solve them?
- Specifically, what's the best way to perform information integration?
-
Semantic Graphs
Semantic Graphs
- Create graphs with meaning
- Encoded within the graph
- By giving formal, declarative definitions of the nodes and edges
- Using a high-level language
- Specifically, to create computer understandable meaning
- So the computer can help
- This lets us use the appropriate abstractions
- And is the obvious choice for information integration problems
Benefits of Declarative
- Let non-programmers perform complex information processing tasks without writing code
- More directly capture expertise
- By letting the actual experts author the business logic
- Easier and more maintainable for programmers, too
- Using the appropriate abstractions
-
Inference rules & queries
- So the computer can do the work
Stardog
- Leading RDF graph database
- Pure Java
- Enterprise, Developer, Community versions
- Great developer experience
- Rich feature set
- Currently version 2.2.4—3.0 due March 16th
- 59 public releases in 3 years (!)
What's new in 3.0?
- HA Cluster production-ready
- Improved read & write performance...in some cases 10x
- Equality reasoning via hybrid materialization
- only RDF database to support all of OWL 2
- ICV repair plans
- User extensible query aggregates, search analyzers
- LDAP integration
- Full (in-memory) DL reasoning
- Support for SPARQL SERVICE
- First class Maven support
What's coming in 3.x?
- Property graphs & graph traverals via Apache Tinkerpop
- Docker support
- Named graph security
- Geospatial queries
- Cluster 2.0
What about 4.0?
- Backing Stardog with Apache Spark
- Data-center scalability with Yarn, Mesos, etc.
- Building a unified graph processing platform:
- Queries, rule reasoning, OWL
- Graph traversals
- Graph algorithms (shortest path, clustering, etc)
- Graph machine learning
- With a unified SPARQL interface for users
Enterprise Features
- HA Cluster: strong consistency guarantees (2PC)
- JMX server monitoring
-
Hot Backup & Restore
- Access/Audit logging
- Web console built on Stardog Web Framework
- PROV and SKOS support
- ACID Transactions
- Rich Security model
Stardog Cluster
- HA Cluster
- Active Replication
- 2PC-based commit protocol for strong consistency
- Writes processed by coordinator to determine order of operations
- Reads are distributed over all nodes
Performance
- Query
- Query 100M triples, thru-put: 4M+ queries per hour. 1B with ˜650k queries/hour and 10B with ˜50k queries/hour
- BSBM with 64 concurrent clients
- Scale
- ~50B triples/quads on modest hardware
- Load rates up to 500k triples/second
- 100M triples in 3 minutes, 1B in 30, and 20B in 20 hours.
Query
- SPARQL 1.1
- Update, query, graph protocol
- Custom query planner, optimized for complex queries
- Targets BI/analytic queries
- And also reasoning
- But does not sacrifice performance at low scales or with simple queries
- Scalable query answering
- Intermediate results can get big, and fast
- Runtime will automatically flow results off-heap, and then to disk as needed
-
Query management
Developers
- First and foremost, we are developers too
- We intend for the best out of box experience possible
- Excellent documentation
- Easy installation, just unzip
- We <3 the command line
- Modeled after the Git command line, autocomplete support
- Like Java?
- Prefer Javascript, Ruby, Python, .Net, Groovy, Clojure
Full Text Search
- Embeds Lucene
- Automatically managed by database as if another RDF index
-
Enables full-text searches over your RDF
- Literals are indexed by Lucene
- Uses the Lucene query language
- Extensible analyzer framework; customize stopword lists, language support, etc.
- Seamless integration via SPARQL
- Join results of full-text searches with regular SPARQL query
- Also available via SNARL Java API
Graph Versioning
- Version control is insanely useful
- Sometimes I wonder how people live without it
- So why not for an RDF database?
- Stardog adds commit management features similar to many popular VCS systems
- Add metadata, like comments, to commits
- Create tags
- Revert to a previous version
- Get diffs between versions
- Oh, all of this is stored as RDF
- So you can query your version history
What is reasoning?
-
Make implicit information explicit
- Implicit in the schema, or data, or both
- Represent domain knowledge in a formal declarative model
- Called an ontology
- Like UML, but with formal semantics
- W3C specification called OWL, Web Ontology Language
- Reasoners consume ontologies to derive new information
- Answer queries, find inconsistencies
- Complex, but manageable
- OWL divided into profiles with less expressivity, but better computational properties
Reasoning
- Unmatched OWL support
- All OWL2 profiles (RL, EL, QL, DL) and Stardog profile (SL)
- Caveats, no equality reasoning, no datatype reasoning, no DL reasoning over your ABox
- Query time reasoning
- No write performance penalty
- Pay for what you use
- Explanations
- Inference you don't understand?
- Reasoner will give you the proof used to derive it!
- Reasoning Services
- Consistency checking, satisfiability
Stardog Rules
- Stardog supports SWRL
-
Part of the SL profile
- You cannot write it by hand, SWRL/RDF is unusable
- Much easier use Stardog Rules
- If-Then style rules based on SPARQL syntax:
PREFIX :
PREFIX math:
IF {
?c a :Circle ;
:radius ?r
BIND (math:pi() * math:pow(?r, 2) AS ?area)
}
THEN {
?c :area ?area
}
ICV
- Integrity Constraint Validation keeps data safe and consistent
- Prevent modifications that violate your integrity constrains
- 'Guard mode'
- Constraint violations abort transactions
- Also support 'oracle' mode, aka 'middleware' mode
- Outside of a transaction
- Check if data valid w.r.t some constraints
- Violations can be explained
- Inferences can satisfy or violate a constraint
- Constraints expressed in SPARQL, OWL, SWRL, or Stardog Rules
- High-level declarative languages make it easy to write simple constraints, possible to write complex ones
ICV Example
Every supervisor should supervise at least one employee
Supervisor subClassOf supervises some Employee
IF {
?x a Supervisor
}
THEN {
?x supervises ?y .
?y a Employee
}
select * {
?x a Supervisor.
FILTER NOT EXISTS {
?x supervises ?y .
?y a Employee
}
}
ICV Explanations
- If you are using ICV
- You may not understand why a violation occurred
- Or want to communicate it to the user
- Explanations
- Tells you why the violation occurred
- Shows exactly the data that caused the violation
- Gives you the proof used to derive the violation
Another ICV Example
If a project is funded by only internal funding sources, then it should be approved by the internal budget office
Project and (fundedBy only InternalFundingSource) subClassOf approvedBy value InternalBudgetOffice
select * where {
?x a Project .
FILTER NOT EXISTS {
?x fundedBy ?y .
FILTER NOT EXISTS {
?y a InternalFundingSource
}
} .
FILTER NOT EXISTS {
?x approvedBy InternalBudgetOffice
}
}
ICV Explanation Example
Every Supervisor should supervise at least one Employee
Supervisor subClassOf supervises some Employee
Alice a Supervisor
VIOLATED Supervisor subClassOf (supervises some Employee)
ASSERTED Alice a Supervisor
NOT_INFERRED x a Employee
Alice supervises x
Graph Analytics
- Coming in Stardog 3.1
- RDF graphs are still just graphs
- Graph measures: in-degree, out-degree, PageRank, betweenness centrality
- Clustering: weak/strongly connected components, clique finding
- Path finding: BFS and shortest path
- Seamless SPARQL integration
- Adding support for (de facto) graph standard: TinkerPop 3
- Native implementation for Gremlin, TinkerPop 3
Admin Console
- In Stardog 2.0 we added the Web Console
- Expose the features of the stardog
CLI in an easy to use web interface
- Add/Remove data, execute queries, etc.
- Or simply browse your data
- In 2.2, we added an administrative web console
- Create and drop database, manage security, etc.
- Everything you can do via the stardog-admin CLI
Questions?
Thanks!
Transactions & Security
- Transactions
- ACID
- Guarded (optionally) by ICV
- 2 Phase Commit over all database components
- RDF Index, Lucene, KB, etc.
- Automatically managed by the database
- Security
- RBAC model
- Based on Apache Shiro
- R/W ACLs for access to individual databases
- Administrative controls for actions against DBMS
- Online/offline a database, modify security settings, etc.
Reasoning Example
- For example, enforcing security (ACLs)
- Can Bob access Resource1?
Bob is-a Admin OR Bob created Resource1 OR (Bob hasRole ?r AND ?r canAccess Resource1) OR ...
- Hard to maintain, encoded domain knowledge into the query
- Can leverage reasoning to simplify
Bob canAccess Resource1
- More concise and maintainable
- Reasoner handles the implementing logic transparently
Stardog 3: A Deep Dive
By Michael Grove
Stardog 3: A Deep Dive
An overview of the features and performance of Stardog
- 1,614