Orchestration · Data store
Shared Blackboard Store
Data storeOrchestrationOrchestration & Toolsarc:SharedBlackboardStore
A common memory space to which agents publish hypotheses, evidence and results and from which they consume others' contributions, enabling implicit, opportunistic coordination.
Responsibility. Holds shared problem state through which agents coordinate implicitly.
Also known as: Blackboard, Shared memory, Shared context, Shared memory layer, Shared context pool, Central context buffer, Shared memory space, Shared workflow state
Relationships
is read by dependency
- Worker Agent abstract Ch1.3 Ch2.6 +3
is written by dependency
- Worker Agent abstract Ch1.3 Ch2.6 +3
sends data to dynamic
- Worker Agent abstract Ch1.3
has access controlled by control
is constrained by control
- State Concurrency Controller abstract Ch5.9 Ch8.2A
Design guidance
- SHOULD be chosen when workflows are unknown or dynamic, many agents contribute to overlapping problems, and agent populations change frequently.
- MUST apply concurrency control and versioning; without conflict resolution, concurrent writes create race conditions.
- SHOULD be distributed with replication and automatic failover; a centralised blackboard is a bottleneck and single point of failure.
- SHOULD log every read and write with timestamp and agent identifier to allow reconstruction of problem-solving paths.
- SHOULD NOT be used where strong consistency or low-latency coordination is required.
- SHOULD scope agent access so no agent can undo another agent's completed work.
- MUST make updates atomic and immediately visible so agents never reason over stale or partially modified context.
- SHOULD apply negotiation or priority mechanisms when agents compete for the pool's collective token budget.
Quantitative guidance
As stated by the sources; verify before use.
- Suitable when scalability to hundreds of agents is required (Ch1.3 selection criteria).
Classification
- Patterns
- Blackboard architectureOpportunistic problem-solvingShared working memory across collaborating agentsCollective token budgetCoordination via shared memory without direct communication (Ref5.05)
- Quality attributes
- Maintainability (ISO/IEC 25010)Performance efficiency (ISO/IEC 25010)
- Risks mitigated
- Rigid predefined workflows
Sources
- Ch1.3: T. Nguyen, "Multi-Agent Systems," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 1.3. ISBN: 9798244538229.
- Ch2.6: T. Nguyen, "Tool Integration and Function Calling," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 2.6. ISBN: 9798244538229.
- Ch5.9: T. Nguyen, "Working Memory," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 5.9. ISBN: 9798244538229.
- Ch8.2A: T. Nguyen, "Error Rates and Reliability," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 8.2A. ISBN: 9798244538229.
- Ref5.04: C. Stryker, "What is AI agent memory?," IBM Think. Accessed: Sep. 27, 2026. [Online]. Available: https://www.ibm.com/think/topics/ai-agent-memory
- Ref5.05: Model Context Protocol, "Model Context Protocol servers," GitHub repository. Accessed: Sep. 27, 2026. [Online]. Available: https://github.com/modelcontextprotocol/servers