Linked data driven EPCIS Event-based Traceability across Supply chain business processes

GS1, 7th November 2014, London
Linked Data Driven, EPCIS Event-Based
Traceability in Supply Chain Business
Processes
Monika Solanki
https://w3id.org/people/msolanki
@nimonika
Aston Business School
Aston University, Birmingham, UK

GS1, 7th November 2014, London
Broad Outline
Motivation
Background
Semantic Web & Linked data
EPC, EPCIS, Pedigrees
Ontologies
Linked Pedigrees
Summary
m.solanki@aston.ac.uk, @nimonika Linked Data Driven, EPCIS Event-Based Traceability in Supply Chains

Motivation GS1, 7th November 2014, London
Part 1
Motivation

The FIspace project: Motivating use cases
Flowers and Plants Supply Chain Monitoring: the
monitoring and communication of transport and logistics
activities focusing on tracking and tracing of shipments,
assets and cargo, including quality conditions and
simulated shelf life.
Meat Information Provenance (GS1 Germany) :
ensuring that consumers, regulators and meat supply
chain participants all have accurate information concerning
where a meat product originated (production farm) and
how it was affected by its distribution (quality assurance).

Observations: Data sharing in supply chains
Existing mechanisms for sharing data and information
along supply chains are highly restricted and extremely
complex.
There is a lack of information models that facilitate the
exchange of end-to-end supply chain product and process
knowledge.
There is a very conservative “need-to-know” attitude such
that essentially information flows only “one-up, one down”.

Observations: Data sharing in supply chains
Traceability datasets curated by partners are inherently
related, however the varied underlying schemas lead to
mapping overheads and interoperability issues.
The semantics of traceability data and data curation
processes are informally defined in
specifications/standards and associated implementations.
Lack of a systematic and standardised way to exchange
traceability information.
Large volumes of traceability data are recorded at each partner’s end.

Requirements
Data Sharing in Supply chains
Information and knowledge need to be interlinked, shared and
made available consistently along the supply chains not least
for regulatory reasons but also due to increasing consumer
demands of being able to track and trace commodities.
Flow of information across stakeholders (Abstraction)

SW/LD for Traceability in Supply chains
Proposed framework
Exploits Semantic Web standards, Linked data principles
and well known ontologies/vocabularies.
Based on GS1’s EPCIS 1.1 and CBV 1.1 standards.
A set of ontologies: EEM, CBVVocab, OntoPedigree.
Event-Based, Provenance-aware traceability artifact:
Linked Pedigrees.
Algorithms for the automated generation of linked
pedigrees from EPCIS events.
ETL processes: EPCIS RDBs R2RML
! Linked data.
Exception detection, constraint validation and inferencing.
...and there is more work-in-progress...
Supply chain domain/sector agnostic, as long as there is conformance to
EPCIS 1.1. CBV 1.1

Semantic Web Linked Data GS1, 7th November 2014, London
Part 2
Semantic Web Linked Data
A minimalistic overview

Semantic Web in 1 slide
Web scale data ! Machines first!
Marks a shift in thinking from publishing data on the Web as
human readable, interlinked HTML documents to publishing
self describing, interlinked data on the Web in
“machine-interpretable” formats.
self describing: associating metadata with data via
vocabularies/ontologies/data dictionaries.
interlinked data: “meaningful” links between “pieces” of self
described data.
machine-interpretable: an underlying model for data that
enables the exploitation of computational power to
automate and improve certain tasks at Web (Big data)
scale e.g. Search, data integration, visualisations and
more...

Linked data in 1/2 slide
Web scale data ! Machines first!
Central idea
Publish data using principles that support Web applications in
discovering and integrating data by complying to a set of best
practices in the areas of linking, vocabulary usage, and
metadata provision.

Principles for 5 star * Linked data
Use (HTTP) URIs as names for things
Provide useful information as structured data
Provide data in non-proprietary formats
Link your data to other datasets using URIs
Linked Open Data *
Publish data under an open license
*http://datahub.io/group/lodcloud
*http://5stardata.info/

The evergrowing LOD cloud*: April 2014
*http://data.dws.informatik.uni-mannheim.de/lodcloud/
m.solanki@aston.ac.uk, @nimonika Linked Data Driven, EPCIS Event-Based2T0ra1ce4a/bilIitySinWSCu-ppRlyDCBha/ins

LOD cloud stats: April 2014
Max Schmachtenberg, Heiko Paulheim and Christian Bizer. Adoption of
Linked Data Best Practices in Different Topical Domains. ISWC 2014

Semantic Web: The W3C Technology stack
HTTP URIs - universal identifiers for resources on the Web.
RDF data model - a “triples” based model.
RDFS and OWL - domain knowledge representation
standards that enable inferencing over asserted facts.
SPARQL - a query language for datasets encoded using
the RDF data model.

HTTP URIs
In order to publish data on the Web, resources, i.e., items
in the dataset and their relationships must be uniquely
identified.
HTTP URIs provide a simple way to create globally unique
names in a decentralised manner.
Besides identifying resources uniquely, they also serve as
a means to access further information about the resources.
Identifying Cologne:
http://live.dbpedia.org/resource/Köln
Identifying Germany:
http://live.dbpedia.org/resource/Germany

RDF: Resource Description Framework
RDF is a data model.
Basic building block: a triple, a statement
A triple is composed of:
subject predicate (property) object
Each RDF Triple is a complete and unique fact.
Abstract data model with several concrete syntaxes.
Most common informal syntax: Directed Graph
Most common formal syntaxes: Turtle, RDF/XML

RDF: Examples
Informal statement (Implicit semantics):
Cologne is in Germany
Informal statement (Explicit semantics):
Cologne has country Germany
RDF “triple” statement:
Cologne hasCountry Germany
Cologne country Germany

RDF: Examples
RDF triple in Turtle:
@prefix: http://fispace.aston.ac.uk/cities#.
:Cologne :country :Germany.
Adding more statements
:Cologne :country :Deutschland,
:Germany;
:leaderName :Jürgen_Roters;
:leaderTitle ‘‘Lord Mayor’’@en;
:areaTotal 405150000.0000^^xsd:double.

RDF: Directed Graph representation
Graphical representation of the RDF data model

HTTP URIs

RDF: HTTP URIs Examples in Turtle
@prefix dbpedia: http://live.dbpedia.org/resource/.
@prefix dbprop: http://live.dbpedia.org/ontology/.
dbpedia:Köln dbprop:country dbpedia:Deutschland,
dbpedia:Germany;
dbprop:leaderName dbpedia:Jürgen_Roters;
dbprop:leaderTitle ‘‘Lord Mayor’’@en;
dbprop:areaTotal 405150000.0000^^xsd:double.

Schemas for RDF triples: RDFS OWL 2
Resource Framework Description Language (RDFS)
Web Ontology Language (OWL 2)
Ontologies: Specification of domain knowledge
Definition of standardised vocabularies used in RDF
triples, e.g, country in
dbpedia:Cologne dbprop:country dbpedia:Germany
RDFS: Class hierarchies, property hierarchies, basic
property restrictions, Individuals(real world entities).
OWL 2: RDFS + (very) expressive constraints + rules + ...
RDFS syntax: RDF/XML, Turtle
OWL 2 syntax: RDF/XML, Turtle, Manchester syntax

Class hierarchy
City and Country are Geographical entities.
City is related to Country through the property country

Defining City: Manchester syntax
Prefix: wiki: http://en.wikipedia.org/wiki/
Class: http://purl.org/ontology/places#City
SubClassOf:
http://purl.org/ontology/places#GeographicalEntity
Annotations:
rdfs:comment A large settlement;
rdfs:label City ,
rdfs:label City@de ,
rdfs:label City@en ,
rdfs:label City@fr-fr ,
rdfs:label Ciudad@es,
rdfs:seeAlso wiki:City,
EquivalentTo:
http://dbpedia.org/ontology/City

Properties

RDFS Property Restrictions
rdfs:domain and rdfs:range specify permitted subjects
and objects for a property respectively.
dbprop:country rdf:type owl:ObjectProperty ;
rdfs:comment identifies the country for a city;
rdfs:domain ns2:City;
rdfs:range ns2:Country.
dbprop:leaderName rdf:type owl:DatatypeProperty ;
rdfs:comment identifies the mayor for a city;
rdfs:domain ns2:City;
rdfs:range xsd:String.
Several other restrictions on properties can be specified in
OWL.

SPARQL: Querying RDF datasets
SPARQL is a syntactically-SQL-like language for
querying RDF datasets via pattern matching.
SPARQL queries contain a set of triple patterns called a
basic graph pattern (BGP).
Triple patterns are like RDF triples except that each of the
subject, predicate and object may be a variable.
A BGP matches a subgraph of the RDF data when RDF
terms from that subgraph may be substituted for the
variables.

SPARQL example
PREFIX dbprop: http://live.dbpedia.org/ontology/
SELECT ?city ?country ?leader
WHERE
{
?city rdf:type ns2:City;
dbprop:country ?country;
dbprop:leaderName ?leader.
?country rdf:type ns2:Country.
}

Vocabularies in the LOD cloud
Well-Known Vocabularies used by more than 5% of all datasets
in the LOD cloud.

Proprietary Vocabularies in the LOD cloud

....and then there is schema.org....
From Guha’s SemtechBiz 2014 Keynote
Since 2010: Google, Yahoo!, Microsoft then Yandex.
One vocabulary understood by all the search engines.
Make it very easy for the (5 million) webmasters.
Syntax: Microdata, RDFa, JSON-LD
*http://www.slideshare.net/rvguha/sem-tech2014c

Linked data principles?
5 star linked data?
Authoritative URIs for entities?
Dereferenceable URIs for entities with content negotiation?

Only a few of the classes and properties are actually used
*http://www.slideshare.net/bizer/
schmachtenberg-bizerpaulheim-lodbestpracticesiswc2014

EPC, EPCIS, CBV Pedigrees GS1, 7th November 2014, London
Part 3
EPC, EPCIS, CBV Pedigrees
A minimalistic overview

EPC, EPCIS, CBV
EPC: provides products with unique, serialised identities.
EPCIS 1.1: provides a set of specifications for the syntactic
capture and informal semantic interpretation of EPC based
product information.
CBV 1.1 supplements EPCIS by defining the structure of
vocabularies and specific values for the vocabulary
elements.
Events as abstractions for traceability: One generic (EPCIS
Event) and four speciliased (Object, Aggregation,
Transaction, Transformation) physical event types.

Data model components
What(product(s)), Where(location), When(time), and
Why(business step and status) of events (product movement)
occurring in any supply chain.
EPCs (SGTINs)
Time
Read Points
Business Location
Business steps
Disposition
Transaction types
Action
Quantities and measurements
Sources and Destinations
ILMD (Instance Lot Master Data)

Pedigrees
Most widely prevalent in the pharmaceutical industry.
Pedigree (e-pedigree) is an audit trail that records the path
and ownership of a drug as it moves through the supply
chain.
Each stakeholder involved in the manufacture or
distribution of the drug adds information to the pedigree.

SW LD for Visibility in Supply chains
Problem statement
* Can we formalise EPCIS using the underlying standards
for Semantic Web and principles of linked data to
represent traceability-specific domain knowledge in
supply chains?
* Can we exploit EPCIS events for the automated
generation of provenance-based traceability/visibility
artifacts that can be shared across supply chain partners?

Ontologies GS1, 7th November 2014, London
Part 4
Ontologies

EEM*: The EPCIS Event Model
A domain specific, ontological information model.
Focuses on a tight conformance with the EPCIS 1.1
standard and Simplicity.
Explicitly defines relationships with CBV 1.1 entities
through CBVVocab*.
EEM has been mapped* to PROV-O*.
*http://purl.org/eem#
*www.w3.org/ns/prov-o
*http://purl.org/cbv#
*http://fispace.aston.ac.uk/ontologies/eem_prov.html

EEM Modules

Modelling the generic EPCISEvent
An EPCIS event has three temporal properties associated
with it.
An EPCIS event occurs at a unique location and is part of
a singular business process.

Modelling the generic EPCISEvent
Class: EPCISEvent
SubClassOf:
eventTimeZoneOffset exactly 1 xsd:dateTime,
eventRecordedAt exactly 1 xsd:dateTime,
eventOccurredAt exactly 1 xsd:dateTime
ObjectProperty: hasReadPointLocation
Characteristics:
Functional
Domain:
EPCISEvent
Range:
ReadPointLocation

Modelling ObjectEvent
An ObjectEvent is an EPCISEvent.
An ObjectEvent is required to have associated EPCs,
and an action.

Modelling ObjectEvent
Class: ObjectEvent
SubClassOf:
(action some Action)
and (associatedWithEPCList some SetofEPCs),
EPCISEvent

EEM Entities: Axiomatisation

EEM Entities: Mapping to PROV-O

Implementing EEM: LinkedEPCIS library
EEM is a complex data model.
Non trivial to generate class assertions and complex
queries without knowing the structure of the model and
nomenclature of the entities.
LinkedEPCIS* - an open source Java API to,
Capture EPCIS events as linked data.
Encourage the uptake of EEM among EPCIS conforming
organisations and industries
Ease the creation of EEM instances
Provides classes, interfaces and RESTful Web services for
capturing, storing and querying EPCIS events.
* https://github.com/nimonika/LinkedEPCIS

Interlinking EPCIS Event data

Applying EEM to the Agri-food domain
The tomato supply chain involves thousands of farmers,
hundreds of traders and few retail groups.

Agri-food scenario: Subset of EPCIS events
Supply chain operation EPCIS event type Business Step Disposition Action type
1. Commissioning crates for tomatoes Object event commissioning active ADD
2. Storing crates Quantity event storing in_progress -
3. Aggregating crates in pallets Aggregation event packing in_progress ADD
4. Loading and shipping pallets Transaction event shipping in_transit ADD

Linked Pedigrees GS1, 7th November 2014, London
Part 5
Linked Pedigrees

Event-based Linked Pedigrees
Encapsulate EPCIS event-based knowledge required to
trace and track products in supply chains.
Facilitate the interlinking of a variety of related and relevant
data, i.e., product master data with event data and other
pedigrees.
Enable sharing of knowledge among partners - pedigrees
are exchanged as products physically flow downstream or
upstream in the supply chain.

OntoPedigree: A CO design pattern
Competency questions:
Who is the creator of the pedigree?
What is the supply chain creation status of a given
pedigree?
Which are the business transactions recorded against a
particular consignment?
What are the events associated with pedigrees created
between dates X and Y?
Which products have been shipped together?
Which other pedigrees are included in the received
pedigree?

OntoPedigree: A CO design pattern

Pedigree: Axiomatisation
Class: ped:Pedigree
SubClassOf:
(hasPedigreeStatus exactly 1 ped:PedigreeStatus)
and (hasSerialNumber exactly 1 rdfs:Literal)
and (pedigreeCreationTime exactly 1 xsd:DateTime)
and (prov:wasAttributedTo exactly 1 ped:PedigreeCreator)
and (ped:hasConsignmentInfo some eem:SetOfEPCISEvents)
and (ped:hasTransactionInfo exactly 1 eem:SetOfEPCISEvents)
and (ped:hasProductInfo min 1),
(prov:wasGeneratedBy only ped:PedigreeCreationService),
(ped:hasReceivedPedigree only eem:Pedigree),
prov:Entity
*possible integration with GTIN+ on the Web
http://www.gs1.org/digital

Generating Linked Pedigrees event URIs
Events incorporated in pedigree creation
commissioning: uniquely identifying products
aggregation: uniquely identifying aggregations
shipping: associating products with orders
receiving: associating received products with orders
Pedigree Component Linking relationship Resource identifier
Product information hasProductInfo Product data URIs
Serialised product data URIs
Consignment information hasConsignmentInfo Commissioning events -
Object event/Aggregation event URIs
Transaction information hasTransactionInfo Shipping events -
Transaction event URIs
Direct linkages in the linked pedigree generated by each supply
chain trading partner

Linked Pedigree: An example
### http://fispace.aston.ac.uk/joetrader/
pedigrees/JoeTomatoTraderPedigree456
jsc:JoeTomatoTraderPedigree456 rdf:type ped:Pedigree
ped:hasSerialNumber joeTradePed456^^xsd:String;
ped:hasStatus ped:Intermediate;
ped:hasConsignmentInfo jci:JoeTraderObjectEvent20,
jci:JoeTraderObjectEvent30;
ped:hasTransactionInfo jti:JoeTraderTransactionEvent40;
ped:hasProductInfo jpi:JoeTradesMay2013Info.
ped:hasReceivedPedigree fsc:FranzTomatoFarmerPedigree123,
bsc:BobTomatoFarmerPedigree123.

Linked Pedigrees: Agri-food supply chains

Linked Pedigrees: Healthcare supply chains
Flow of linked pedigrees (Abstraction)
M. Solanki and C. Brewster. EPCIS event-based traceability in
pharmaceutical supply chains via automated generation of linked pedigrees.
ISWC 2014. Springer-Verlag.

Architecture and Implementation

Transformation Events: Wine production
EPCIS events generated during the wine processing stages
M. Solanki and C. Brewster. Modelling and Linking transformations in EPCIS
governing supply chain business processes. EC-Web 2014. Springer-LNBIP.

Typical queries
1 Tracking ingredients: What were the inputs consumed
during processing in the batch of wine bottles shipped on
date X?
2 Tracking provenance: Which winery staff were present at
the winery when the wine bottles were aggregated in
cases with identifiers X and Y?
3 Tracking external data: Retrieve the average values for
the growth temperature for grapes used in the production of
a batch of wine to be shipped to Destination D on date X.

Transformation Events: ETL Framework

EPCIS Exceptions
Typical examples
(e1) Pedigree serial number discrepancy
(e2) product inference problem - the inability to infer about
products contained in an outer container without
disaggregation using pedigree information
(e3) quantity inference problem - the inability to derive the
total quantity of items packed in an outer container without
disaggregation using pedigree information
(e4) missing or incorrect containment hierarchy between
items and their containers - source of counterfeits.
(e5) incomplete pedigree data
(e6) pedigree data with broken chains, i.e., missing
intermediate stakeholder pedigree information.

Hierarchy of EPCIS Exceptions

EPCISExceptionEvent: Axiomatisation
M. Solanki and C. Brewster. Detecting EPCIS Exceptions in linked
traceability streams across supply chain business processes. SEMANTiCS
2014. ACM-ICPS.

Summary GS1, 7th November 2014, London
Part 6
Summary

EEM: EPCIS Event Model
Data visibility (tracking and tracing) in supply chains has
received considerable attention in recent years.
EEM based linked datasets can be exploited in order to
improve visibility, accuracy and automation along the
supply chain.
EEM along with CBVVocab can be used to derive implicit
knowledge that can expose inefficiencies such as shipment
delay, inventory shrinkage and out-of-stock situation.

Linked Pedigrees
Semantic Web standards, ontologies and linked data can
be utilised to record and represent real time supply chain
knowledge via “linked pedigrees”.
EEM forms the basis for traceability in supply chains -
Event-based Linked Pedigrees.
Complex Event Processing over continuous streams of
semantically interlinked EPCIS event datasets enable
automated generation of linked pedigrees, detection of
exceptions and validation of integrity constraints.
The proposed approach is domain independent and can
be widely applied to most scenarios of traceability as long
as there is conformance to EPCIS 1.1 in the supply chain.

Further information
M. Solanki and C. Brewster. A Knowledge Driven Approach towards the
Validation of Externally Acquired Traceability Datasets in Supply Chain
Business Processes. EKAW 2014. Springer-Verlag.
M. Solanki and C. Brewster. EPCIS event-based traceability in
pharmaceutical supply chains via automated generation of linked
pedigrees. ISWC 2014. Springer-Verlag.
M. Solanki and C. Brewster. Modelling and Linking transformations in
EPCIS governing supply chain business processes. EC-Web 2014.
Springer-LNBIP.
M. Solanki and C. Brewster. Detecting EPCIS Exceptions in linked
traceability streams across supply chain business processes.
SEMANTiCS 2014. ACM-ICPS.
M. Solanki and C. Brewster. Consuming Linked data in Supply Chains:
Enabling data visibility via Linked Pedigrees. COLD2013 at ISWC,
volume Vol-1034. CEUR-WS.org proceedings, 2013.
M. Solanki and C. Brewster. Representing Supply Chain Events on the
Web of Data. DeRiVE at ISWC. CEUR-WS.org proceedings, 2013.
http://windermere.aston.ac.uk/~monika/ontologies.html
http://windermere.aston.ac.uk/~monika/publication.html

Linked data driven EPCIS Event-based Traceability across Supply chain business processes

Recommended

Recommended

More Related Content

What's hot

What's hot (7)

Viewers also liked

Viewers also liked (11)

Similar to Linked data driven EPCIS Event-based Traceability across Supply chain business processes

Similar to Linked data driven EPCIS Event-based Traceability across Supply chain business processes (20)

More from Monika Solanki

More from Monika Solanki (17)

Recently uploaded

Recently uploaded (20)

Linked data driven EPCIS Event-based Traceability across Supply chain business processes