Explanation:

NTO needs sales reps to view and update real-time order information in Salesforce, while ensuring the order data only persists in the external system (not copied into Salesforce).

This requirement is best met by the Data Virtualization pattern, typically implemented with Salesforce Connect and External Objects:

Data remains in the external (remote) system.
Salesforce accesses the data on demand via real-time callouts (e.g., OData).
Users can view and, if configured, update the external data directly from Salesforce pages, related lists, and search.
No data replication or synchronization into Salesforce is required.

Why the other options are incorrect

A. Data Synchronization
Involves copying/replicating data between systems (batch or near-real-time). This violates the requirement that data should only persist in the external system.

C. Process Orchestration
Focuses on coordinating multi-step business processes across systems. It is not the primary pattern for providing real-time access to external data without persistence in Salesforce.

Key takeaway
When the business requires real-time access to external data without storing a copy in Salesforce, recommend the Data Virtualization pattern (Salesforce Connect / External Objects).

Explanation:

The key to this question is recognizing that the Case Management System is the legacy system being replaced by Salesforce. When implementing Salesforce for Case Management, the architect's primary job is to identify which systems should be retired versus which must be integrated.

Why the Case Management System is excluded
Salesforce Service Cloud provides native, best-in-class case management capabilities. Since the entire purpose of the implementation is to use Salesforce for Case Management, the existing standalone Case Management System becomes redundant and should be retired. Integrating with a system you are actively replacing would create data fragmentation and defeat the purpose of the implementation.

Why the remaining three systems require integration
Email Management System: This system handles inbound customer communications. An integration (such as Email-to-Case or a connector) is needed so that incoming emails automatically generate or update cases in Salesforce.

Order Management System (OMS): Support agents frequently need to view order history or status to resolve customer inquiries. Integrating Salesforce with the OMS provides a 360-degree view, enabling agents to see relevant order data directly within the case console.

Data Warehouse: For long-term reporting, trend analysis, and a unified customer profile, case data from Salesforce needs to be pushed to the Data Warehouse. This ensures downstream analytics tools can report on support metrics alongside other enterprise data.

Why the Other Options Are Incorrect

Option A: Includes the Case Management System, which is the system being replaced by Salesforce's native capabilities.

Option B: Also includes the Case Management System, repeating the same error.

Exam Focus Point
This question tests the ability to evaluate a system landscape and identify systems of record versus systems to be retired. When Salesforce is being implemented for a specific function (like Case Management), the architect should retire any legacy system that duplicates that function, while integrating with surrounding systems that provide complementary data or capabilities.

Reference
Salesforce Integration Architecture Designer Resource Guide
ExamTopics: Salesforce Certified Integration Architect discussion confirming C as the consensus answer

Explanation:

Why B is correct

This is a close sibling of the earlier question about timing. Here the option that gives the architect the concrete integration inputs is B. To move from an abstract "data updates between systems" requirement to an architecture, the architect needs to know:

Source and target systems: which systems participate, what APIs or interfaces they expose, and which system is the master for each data domain.

Directionality: inbound to Salesforce, outbound from Salesforce, or bidirectional. This decides between patterns such as Remote Call-In, Remote Process Invocation, and Batch Data Synchronization, and whether conflict resolution is needed.

Data volume: this drives the choice between REST, Bulk API, Pub/Sub API, and Data Virtualization, and how close the design comes to governor and API limits.

Transformation complexity: heavy mapping, orchestration, or protocol conversion points to middleware rather than point-to-point logic in Salesforce.

Available middleware: an existing ESB or iPaaS (such as MuleSoft) can be leveraged for routing, transformation, retry, and monitoring.

Why A is wrong

User experience design, community choices, and licensing are general program concerns. Encryption may matter for security, but the option as a whole does not address the data flow characteristics that shape integration design.

Why C is wrong

Skills, SME availability, and governance affect delivery and staffing, not which patterns and technologies the integration should use.

Reference

Salesforce Integration Patterns and Practices: Pattern Selection Guide and the criteria for choosing a pattern (systems, direction, volume, transformation)
Trailhead: Integration Architecture concepts in the Architect learning paths

Explanation:

1. Why B is Correct (Synchronous Dependency & Immediate Feedback)

Real-Time Synchronous Requirement:
Agents are opening accounts on the spot and must inform the customer immediately of the newly created bank account ID generated by the core banking system before the interaction concludes.

Request and Reply Pattern:
This business scenario demands a synchronous Request and Reply pattern. Salesforce must make a direct, blocking API call (such as a REST callout) to the core banking system, wait for the transaction to complete, receive the newly generated account ID in the immediate response payload, and present it back to the agent in the UI.

2. Why the Other Options are Incorrect

A. Salesforce platform event:
Platform events are asynchronous and event-driven (Fire and Forget). While they handle data synchronization reliably, they do not inherently provide the immediate, synchronous response payload required for an agent to wait on the screen and read back an ID in real time.

C. Streaming API to generate PushTopic:
PushTopics are a legacy streaming mechanism used to stream outbound notifications regarding record changes within Salesforce to external subscribers. They do not facilitate a real-time, two-way request-reply flow for generating records and returning IDs from an external system of record.

Explanation:

The requirement is very high-level:

Business processes on Salesforce require data updates between internal systems and Salesforce.

To design the integration architecture, the architect needs to determine when and how data must move between Salesforce and the internal systems.

Important details include:

Real-time vs. near-real-time
Synchronous vs. asynchronous
Batch processing
Update frequency
Potentially the required response time and processing window

These factors directly influence the choice of integration pattern and mechanism, such as synchronous request/reply, fire-and-forget, event-driven integration, or batch data synchronization.

Therefore, B is the most directly relevant detail.

Why A is incorrect
A. Core functional/non-functional requirements for UX, encryption, Community, and licenses ❌
These can certainly be important to the overall Salesforce program, but they do not specifically provide the key information needed to determine the integration pattern and data-exchange mechanism.

Why C is incorrect
C. Integration skills, SME availability, and Program Governance details ❌
These are important implementation and governance considerations, but they do not define how the systems need to exchange data.

Exam Tip
When a question asks what detail an integration architect should seek to determine the integration architecture, focus on:
When does data need to move? → How fast? → Synchronous or asynchronous? → Batch or real-time? → How frequently?

Final Answer: B