AWS prescriptive-guidance documentation change
Summary
Updated section headings, simplified terminology (removed 'with AWS Step Functions'), corrected image paths, fixed typos and formatting in tables/links, and removed redundant sentences.
Security assessment
Changes are editorial and structural (image path updates, terminology simplification, formatting fixes). No security vulnerabilities, configurations, or features are mentioned or modified. The content remains focused on orchestration models without security implications.
Diff
diff --git a/prescriptive-guidance/latest/agentic-ai-serverless/orchestration-models.md b/prescriptive-guidance/latest/agentic-ai-serverless/orchestration-models.md index 7feebff8e..ad8a4fde5 100644 --- a//prescriptive-guidance/latest/agentic-ai-serverless/orchestration-models.md +++ b//prescriptive-guidance/latest/agentic-ai-serverless/orchestration-models.md @@ -7 +7 @@ -Rule-based orchestration with AWS Step FunctionsAI-native orchestration with Amazon Bedrock AgentsRule-based or AI-native: When to use which?Event-driven orchestrationStrategic perspective +Rule-based orchestrationAI-native orchestration with Amazon Bedrock AgentsRule-based or AI-native: When to use which?Event-driven orchestrationStrategic perspective @@ -22 +22 @@ Each model plays a distinct role in building flexible, reactive, and intelligent -## Rule-based orchestration with AWS Step Functions +## Rule-based orchestration @@ -39 +39 @@ The following diagram shows the workflow of an example use case of document inge - + @@ -68 +68 @@ Each task is a defined state with built-in error handling. No decisions are made -Where Step Functions manages how things happen, agents for Amazon Bedrock decide what should happen based on user goals. An [Amazon Bedrock agent](https://docs.aws.amazon.com/bedrock/latest/userguide/agents.html) or agents built on Amazon Bedrock AgentCore combine the following: +Where Step Functions manages how things happen, agents for Amazon Bedrock decide what should happen based on user goals. An [Amazon Bedrock agent](https://docs.aws.amazon.com/bedrock/latest/userguide/agents.html) or agents built on Amazon Bedrock AgentCore, combine the following: @@ -72 +72 @@ Where Step Functions manages how things happen, agents for Amazon Bedrock decide - * A set of tool integrations such as Lambda functions (or Model Context Protocol (MCP) client to execute MCP integrations) + * A set of tool integrations such as Lambda functions (or MCP client to execute MCP integrations) @@ -98 +98 @@ The following diagram shows the workflow of an example use case of customer supp - + @@ -125 +125 @@ In this example, a user on a retail website types a message in the support chatb -The entire workflow demonstrates how Amazon Bedrock Agents orchestrates complex business logic through defined action groups. By connecting customer intent with backend systems and processes, it delivers an automated yet contextually appropriate customer service experience. +The entire workflow demonstrates how Amazon Bedrock Agents orchestrates complex business logic through defined action groups. @@ -133 +133 @@ The following diagram shows the workflow of an example use case of customer supp - + @@ -135 +135 @@ The following diagram shows the workflow of an example use case of customer supp -This example follows the same actions as the earlier Amazon Bedrock Agents example: A user on a retail website types a message in the support chatbot. The following workflow occurs: +This example follows the same actions as the earlier Amazon Bedrock Agents example: a user on a retail website types a message in the support chatbot. The following workflow occurs: @@ -150 +150 @@ AgentCore introduces three key capabilities that complement existing orchestrati - * **AgentCore Memory** – Provides persistent, structured storage for context, state, and task history. This enables agents to maintain continuity across invocations and workflows, supporting both ephemeral and long-term memory modes. Memory data can be synchronized with DynamoDB or Amazon Simple Storage Service (Amazon S3) for observability and compliance. + * **AgentCore Memory** – Provides persistent, structured storage for context, state, and task history. This enables agents to maintain continuity across invocations and workflows, supporting both ephemeral and long-term memory modes. Memory data can be synchronized with DynamoDBor Amazon S3 for observability and compliance. @@ -167 +167 @@ AWS Step Functions and Amazon Bedrock Agents each excel in different orchestrati -**Use case type** | **Step Functions** **(Rule-based)** | **Amazon Bedrock Agents** **(AI-native)** +Use case type| Step Functions (Rule-based)| Amazon Bedrock Agents (AI-native) @@ -190 +190 @@ Whether using rule-based or AI-native orchestration, events are the mechanism th - 3. AgentCore agents can emit and subscribe to EventBridge events natively by using the [AgentCore SDK](https://aws.github.io/bedrock-agentcore-starter-toolkit/). With this approach, agents participate directly in serverless workflows while maintaining long-term context through AgentCore Memory. This integration forms a _dual communication layer_ : + 3. AgentCore agents can emit and subscribe to EventBridge events natively by using the[AgentCore SDK](https://docs.aws.amazon.com/bedrock-agentcore/latest/devguide/agentcore-sdk-memory.html). With this approach, agents can participate directly in serverless workflows while maintaining long-term context through AgentCore Memory. This integration forms a _dual communication layer_ : @@ -192 +192 @@ Whether using rule-based or AI-native orchestration, events are the mechanism th - * EventBridge provides deterministic, auditable event routing. + * EventBridge provides**** deterministic, auditable event routing. @@ -194 +194 @@ Whether using rule-based or AI-native orchestration, events are the mechanism th - * AgentCore Memory plus the Agent2Agent Protocol (A2A) provides semantic state sharing and capability discovery. + * AgentCore Memory + A2A**** provides semantic state sharing and capability discovery.