Know Large Data Volumes (LDV)
Large Data Volumes refer to Salesforce.com Environments that have millions to hundred of millions of records. The underlying multi-tenant architecture of the Force.com platform solves lots of problems and creates some interesting ones. LDV environments pose some interesting technical challenges that other orgs don't often face including but not limited to:
Data Skew
Sharing Calculation Time
Upsert Performance
Report Timeouts
Non-Selective Queries / Indexing Considerations / PK Chunking Mechanisms
There are many questions on the exam that cover LDV scenarios.
How do you build Data Models that support LDV (Skinny Tables, Lookups, ETC)? Do you normalize or denormalize your objects and why?
How do you load millions of records successfully (Use the SOAP Bulk API)?
What settings may you need to turn off during a data load to successfully load data (Workflows, Validation Rules, Defer Sharing Calculations, ETC)?
How will you need to structure the data operations (Perform Seperate Insert & Update rather than the simpler but more time consuming Upsert)?
How do you group the data in batches (Group by Parent Lookup for individual batches to avoid record row lock issues)?
My advice is to read the following material to get familiar with LDV:
LDV Best Practices
Optimizing SOQL, List Views, and Reports
Bulk API Data Loading and Parallel Throughput
https://developer.salesforce.com/docs/atlas.en-us.api_asynch.meta/api_asynch/async_api_headers_enable_pk_chunking.htm
Know Master Data Management
Several of the questions revolve around understanding data management. This includes Data Quality best practices (Using Data Duplicate Matching Rules, Validation Rules, Reports and Dashboards for Data Quality KPI) as well as Data Governance Programs and Systems of Records. There are several questions about maintaining data integrity between SFDC and 3 other systems. How do you maintain a System of Record across the Systems?
My advice is to understand the mechanisms for Data Quality on the Platform including:
Data.com
Duplicate Management Capability of Salesforce.com
Lead Conversion
Merge Accounts
There are also questions about setting up a Data Governance Program. What roles do you need for a Governance Program (Data Stewards, ETC)? This is that architecture considerations and how do you technically run a program as the data architect.
Know Data Loading Options
This is part integration and part data management. You need to know what options are available to integrate data across systems. ETL Tools, Data Loader, SFDC Data Import Wizard, Outbound Messages, SOAP and REST API considerations. For example one question asked about considerations for loading millions of records into the System on a frequent basis (You need to think about API Limits!). Another question asked about using Outbound Messaging to do a outbound read-only synchronization to a ERP system for Opportunities. You should be familiar with all the integration options with respect to data layer integrations.
Conclusion
This exam validates that an architect candidate has performed implementation of, or is familiar with large, complex Salesforce Orgs. I hope the above material will give folks insight into this new exam.
If you want to learn more connect with me on Twitter. I am looking for the next set of great Force.com Engineers to work with. If you are in Chicago especially network with me. I am looking to hire talent that I can mentor and grow our Fusion Engineering Team with!
If you want to learn more connect with me on Twitter. I am looking for the next set of great Force.com Engineers to work with. If you are in Chicago especially network with me. I am looking to hire talent that I can mentor and grow our Fusion Engineering Team with!
