MuleSoft Certified Integration Architect - Level 1

A global organization operates datacenters in many countries. There are private network links between these datacenters because all business data (but NOT metadata) must be exchanged over these private network connections.
The organization does not currently use AWS in any way.
The strategic decision has just been made to rigorously minimize IT operations effort and investment going forward.
What combination of deployment options of the Anypoint Platform control plane and runtime plane(s) best serves this organization at the start of this strategic journey?

Anypoint Exchange is required to maintain the source code of some of the assets committed to it, such as Connectors, Templates, and API specifications.
What is the best way to use an organization's source-code management (SCM) system in this context?

An organization is designing an integration solution to replicate financial transaction data from a legacy system into a data warehouse (DWH).
The DWH must contain a daily snapshot of financial transactions, to be delivered as a CSV file. Daily transaction volume exceeds tens of millions of records, with significant spikes in volume during popular shopping periods.
What is the most appropriate integration style for an integration solution that meets the organization's current requirements?

A set of integration Mule applications, some of which expose APIs, are being created to enable a new business process. Various stakeholders may be impacted by this. These stakeholders are a combination of semi-technical users (who understand basic integration terminology and concepts such as JSON and XML) and technically skilled potential consumers of the Mule applications and APIs.
What is an effective way for the project team responsible for the Mule applications and APIs being built to communicate with these stakeholders using Anypoint
Platform and its supplied toolset?

A Mule application is being designed to do the following:
Step 1: Read a SalesOrder message from a JMS queue, where each SalesOrder consists of a header and a list of SalesOrderLineltems.
Step 2: Insert the SalesOrder header and each SalesOrderLineItem into different tables in an RDBMS.
Step 3: Insert the SalesOrder header and the sum of the prices of all its SalesOrderLineltems into a table in a different RDBMS.
No SalesOrder message can be lost and the consistency of all SalesOrder-related information in both RDBMSs must be ensured at all times.
What design choice (including choice of transactions) and order of steps addresses these requirements?