Option B is correct because putting the business purpose in the Description of each field is a method that an architect can use to trace back configuration changes to the detailed requirements1. The Description of each field provides a brief explanation of what the field is used for and why it is needed2. Option D is correct because using the Force.com IDE to save the metadata files in source control is another method that an architect can use to trace back configuration changes to the detailed requirements1. The Force.com IDE is an integrated development environment that allows developers to work with Salesforce metadata files and Apex code3. Source control is a system that tracks and manages changes to code and configuration files4. Option A is not correct because reviewing the setup audit trail for configuration changes is not a method to trace back configuration changes to the detailed requirements, but a way to monitor and audit the changes made in the setup area. Option C is not correct because maintaining a data dictionary with the justification for each field is not a method to trace back configuration changes to the detailed requirements, but a document that provides information about the data entities and attributes in a system.

Creating a report in an external system combining Salesforce Account data and Invoice data from the ERP, and creating a Visualforce page combining Salesforce Account data and Invoice external object data are two approaches that an architect can suggest for achieving this solution. Both of these approaches can display data from both Accounts and Invoices in one place, using either an external reporting tool or a custom web page.
Creating a report combining data from the Account standard object and the Invoices external object is not possible because Salesforce does not support reporting on external objects. Creating a separate Salesforce report for Accounts and Invoices and combining them in a dashboard is not ideal because it will not show the data in one report, but rather in two separate components.

According to the Salesforce documentation2, extracting large amounts of data from Salesforce can be challenging and time-consuming, as it can encounter performance issues, API limits, timeouts, etc. To extract
100 million records from Salesforce for migration to an enterprise data warehouse (EDW), a data extraction strategy that can provide maximum performance is:
Utilize PK Chunking with the Bulk API (option C). This means using a feature that allows splitting a large query into smaller batches based on the record IDs (primary keys) of the queried object. This can improve performance and avoid timeouts by processing each batch asynchronously and in parallel using the Bulk API3.
Installing a third-party AppExchange tool (option A) is not a good solution, as it can incur additional costs and dependencies. It may also not be able to handle such a large volume of data efficiently. Calling the REST API in successive queries (option B) is also not a good solution, as it can encounter API limits and performance issues when querying such a large volume of data. Using the Bulk API in parallel mode (option D) is also not a good solution, as it can still cause timeouts and errors when querying such a large volume of data without chunking.

Improving reporting performance by replacing the existing reports in Sales Cloud with new reports based on Analytics Cloud can avoid the time-out issues by leveraging the power and scalability of Analytics Cloud. Analytics Cloud can handle large volumes of data and provide faster and more interactive reports than Sales Cloud .