Data Contracts Management
Schema Registry and Beyond
About me
Pietro La Torre
Strategy & Business Development
@ Quantyca SpA
Quantyca is a privately owned technological consulting firm
specialized in data and metadata management based in Italy
quantyca.it
Always a good idea
Data contracts
Data as a Product
Data Centricity Data Fabric
Data Contract
Needed toformalize applications’
accountability on generated data
Foundational to proactively provide
metadatathat enables automation
Essential to clarify meaning, shape, content
and how to access data
images: Flaticon.com
Formal agreement between data producer and consumers
on the structure and semantics of data exchanged
Key characteristics
Good data contracts
Addressable
Expressive
Scoped
Stable
Reliable
Computable
Data Contract
images: Flaticon.com
From an architectural perspective
An interesting paradox
images: Flaticon.com
Slow and ruled context
with vulnerable flows
Fluid landscape resilient
to changes
Schema is not enough
Data Contracts
How do I use this data?
Where can I consume it? How?
How often it is updated?
Who is allowed to use it?
What is its quality?
What does this value mean?
API
How-to
for data interactions
• Service location
• Communication protocols
• Authentication methods
• Endpoints
• Schema
• SLO
• SLA
• Terms and conditions
• Deprecation policy
• Billing policy
• Expected usage patterns
• Schema tagging
• Schema linking
• Semantic linking
Constraints
Rules
to properly use data
Semantic
Context
and meaning
A socio-technical puzzle
Adoption challenges
images: Flaticon.com
Define and enforce
Specifications and tools still immature
Require evolution of operating model,
strong sponsorship and engagement
Technical
perspective
Organizational
perspective
Consumer
Producer
Negotiate and own
Team A Team B
Data Contract
Standardization through a specification
Addressing the technical challenge
dpds.opendatamesh.org
Automation
Interoperability
Scalability
images: Flaticon.com
Validation and Monitoring
Enforcing Data Contracts
CEP
Contract Enforcing Point
Central governance of
data contracts’ lifecycle
CDP
Contract Decision Point
Integration with
underlying tech stack
CIP
Contract Information Point
Census for
archival and retrieval
CEP
Data Contract
CDP
CDP
CDP
1
CIP
4
2
3
5
Validation
CEP CDP
CDP
CDP
CIP
1
3
4
Monitoring
2
A unified approach across service and platforms
Mastering Data Contracts
• Data contracts span across numerous services and platforms
• Schema Registry serves as a cornerstone
• Applying a holistic strategy ensures end-to-end effectiveness
Data
product
Data
contract
CDP
CDP
CIP
Confluent Platform
Schema
Registry
CEP
Data Platform
Data Contracts
Control Center
Thank you. @pietrolatorre
pietro.latorre@quantyca.it
Pietro La Torre

Data Contracts Management: Schema Registry and Beyond

  • 1.
  • 2.
    About me Pietro LaTorre Strategy & Business Development @ Quantyca SpA Quantyca is a privately owned technological consulting firm specialized in data and metadata management based in Italy quantyca.it
  • 3.
    Always a goodidea Data contracts Data as a Product Data Centricity Data Fabric Data Contract Needed toformalize applications’ accountability on generated data Foundational to proactively provide metadatathat enables automation Essential to clarify meaning, shape, content and how to access data images: Flaticon.com Formal agreement between data producer and consumers on the structure and semantics of data exchanged
  • 4.
    Key characteristics Good datacontracts Addressable Expressive Scoped Stable Reliable Computable Data Contract images: Flaticon.com
  • 5.
    From an architecturalperspective An interesting paradox images: Flaticon.com Slow and ruled context with vulnerable flows Fluid landscape resilient to changes
  • 6.
    Schema is notenough Data Contracts How do I use this data? Where can I consume it? How? How often it is updated? Who is allowed to use it? What is its quality? What does this value mean? API How-to for data interactions • Service location • Communication protocols • Authentication methods • Endpoints • Schema • SLO • SLA • Terms and conditions • Deprecation policy • Billing policy • Expected usage patterns • Schema tagging • Schema linking • Semantic linking Constraints Rules to properly use data Semantic Context and meaning
  • 7.
    A socio-technical puzzle Adoptionchallenges images: Flaticon.com Define and enforce Specifications and tools still immature Require evolution of operating model, strong sponsorship and engagement Technical perspective Organizational perspective Consumer Producer Negotiate and own Team A Team B Data Contract
  • 8.
    Standardization through aspecification Addressing the technical challenge dpds.opendatamesh.org Automation Interoperability Scalability images: Flaticon.com
  • 9.
    Validation and Monitoring EnforcingData Contracts CEP Contract Enforcing Point Central governance of data contracts’ lifecycle CDP Contract Decision Point Integration with underlying tech stack CIP Contract Information Point Census for archival and retrieval CEP Data Contract CDP CDP CDP 1 CIP 4 2 3 5 Validation CEP CDP CDP CDP CIP 1 3 4 Monitoring 2
  • 10.
    A unified approachacross service and platforms Mastering Data Contracts • Data contracts span across numerous services and platforms • Schema Registry serves as a cornerstone • Applying a holistic strategy ensures end-to-end effectiveness Data product Data contract CDP CDP CIP Confluent Platform Schema Registry CEP Data Platform Data Contracts Control Center
  • 11.