Training course
This is a free, self-paced reading course for DP-900: Microsoft Certified: Azure Data Fundamentals, the entry-level data credential in Microsoft’s 900-series, alongside AZ-900 and AI-900. DP-900 does not ask you to have run a production data platform. It asks whether you can hold the map in your head: structured against semi-structured against unstructured data, relational against non-relational stores, transactional against analytical workloads, and where each Azure data service sits on that map. Reading is how you build the map; practice questions are how you find the holes in it.
The course is organized as one module per official exam domain, in the order Microsoft publishes them, and each module carries the domain’s published weight. Core Data Concepts and Analytics Workload are the two heaviest domains and they get the two longest modules, which matches where candidates actually lose marks, the first because its vocabulary is assumed by every later question, the last because it spans the widest set of services. Every lesson ends with its key terms and suggested further reading, so it works as a glossary on the second pass as well as a syllabus on the first.
What the course covers
Core Data Concepts
Module 1 · 25-30%The foundation terminology the rest of the exam assumes: structured, semi-structured, and unstructured data; the file formats you are expected to place (CSV, JSON, Parquet, and Avro; transactional against analytical workloads; the ACID guarantees; and the canonical data roles) database administrator, data engineer, and data analyst, with a clear line around which part of the stack each one owns. This module is the one most worth reading slowly, because skimming it is where most wrong answers later in the exam originate.
Relational Data on Azure
Module 2 · 20-25%Relational fundamentals (primary keys, foreign keys, normalization, views, stored procedures, and indexes) mapped onto Microsoft’s Azure SQL portfolio. The module works through the choice DP-900 keeps asking about: Azure SQL Database as the most abstracted PaaS option, Azure SQL Managed Instance for near-full SQL Server feature parity as a managed service, and SQL Server on Azure VMs where you need full control of the operating system, plus the managed open-source options Azure Database for PostgreSQL and MySQL. Basic SQL literacy is covered too: SELECT, INSERT, UPDATE, DELETE, and which statement category CREATE TABLE belongs to.
Non-Relational Data on Azure
Module 3 · 15-20%The Azure non-relational lineup: Blob storage with its Hot, Cool, Cold, and Archive tiers; Azure Files for SMB and NFS shares; Azure Table storage as a schemaless key and attribute store; and Azure Cosmos DB across all five of its APIs: NoSQL, MongoDB, Cassandra, Table, and Gremlin. Cosmos DB gets the most room here because it gets the most questions: the multi-model design, the tunable consistency levels, and the global-distribution story come back repeatedly.
Analytics Workload
Module 4 · 25-30%The widest module, and tied for the heaviest. Azure Data Factory and its vocabulary of linked services, triggers, and pipelines; ETL against ELT; Azure Synapse Analytics with dedicated and serverless SQL pools and Spark pools; Azure Data Lake Storage Gen2 and its hierarchical namespace; Azure Databricks and Delta Lake; Microsoft Fabric and OneLake; streaming through Event Hubs, IoT Hub, and Stream Analytics; and the Power BI essentials, dashboards, reports, Power BI Desktop, DAX, and star-schema modeling.
How to use it
Read a module, then take the matching portion of the DP-900 practice exam rather than waiting until you have read all four. DP-900’s difficulty is not conceptual depth, it is discrimination between neighbors, which Cosmos DB API, which SQL deployment tier, which Synapse pool type, dedicated or serverless. That distinction only surfaces when you are choosing between four plausible options, not when you are reading a paragraph that describes each of them in turn. The free sample set is a quick way to calibrate before committing to a full attempt. The course modules themselves open once you are signed in to a free Certifym account.
Aim higher than you think you need to. The pass mark is 700 out of 1000 on Microsoft’s scaled scoring, not a raw 70%, so knowing only the easy items in each domain tends to land in the low-to-mid 600s. Deliberately learn the edge cases, the five Cosmos DB APIs, the five consistency levels, the three-tier Azure SQL spectrum, the Data Factory vocabulary, the Synapse pool types. For exam logistics (question count, timing, the $99 fee, and the free annual renewal) see the DP-900 certification guide.
IaaS and PaaS: choosing an Azure SQL deployment option
This is the decision DP-900 returns to most often in the relational domain. Azure offers several ways to run SQL Server workloads, and they differ in how much of the stack Microsoft operates and how much you do. Understanding the spectrum lets you answer the question from requirements rather than from memory.
The shared responsibility spectrum
Running a database involves layers: physical hardware and networking, the hypervisor, the operating system, the database engine, its patching and backups, high availability, and finally the database and its data. Every deployment option draws a line through that stack. Above the line is your responsibility; below it is Microsoft's.
Moving down the spectrum from infrastructure-as-a-service to platform-as-a-service trades control for reduced operational work. Neither end is better in the abstract. The question is always which requirements you have that would be blocked by giving up control, and how much operational effort you want to keep.
SQL Server on Azure Virtual Machines — IaaS
This is SQL Server installed on a virtual machine you own. Microsoft provides the hardware, hypervisor, and networking; you own everything from the operating system upward — patching Windows or Linux, patching SQL Server, configuring backups, configuring high availability such as Always On availability groups, and sizing the machine.
You get complete control in return. Any SQL Server version, including versions old enough to be out of mainstream support. Any feature, including those that require operating system access. Full administrative rights on the machine. Third-party software installed alongside the engine. Cross-instance features that depend on the server itself.
Choose it when you need an exact version, when you need operating-system-level access or software that must run on the same host, when you are lifting and shifting an application with minimal change, or when a required feature is unavailable in the managed options.
Azure SQL Managed Instance — PaaS
Managed Instance is a fully managed instance of the SQL Server engine with near-complete feature compatibility. Microsoft operates the operating system, patching, backups, and high availability. You manage databases, security, and performance.
What distinguishes it from Azure SQL Database is instance-scoped capability. It supports cross-database queries within the instance, SQL Server Agent for scheduled jobs, Service Broker, the Common Language Runtime, linked servers, and native database restore from backups taken on-premises. It is deployed into your virtual network, which suits organizations requiring private connectivity by policy.
Choose it when migrating an existing SQL Server estate that uses instance-level features, when you want to stop managing operating systems without rewriting the application, or when the migration path must accept an existing backup file.
Azure SQL Database — PaaS
Azure SQL Database is the most abstracted option: a managed database rather than a managed instance. Microsoft handles everything below the database itself, including patching, backups, high availability, and much performance tuning through automatic tuning. There is no operating system to see and no instance to configure.
It offers capabilities the others do not. It scales without downtime, including a serverless tier that scales compute automatically and can pause during inactivity. Elastic pools let several databases share a pool of resources, which suits software-as-a-service applications running one database per tenant with unpredictable individual demand. Hyperscale supports very large databases with fast backup and restore. Built-in high availability comes as standard rather than as something you configure.
The constraints follow from the abstraction. Instance-scoped features are unavailable or limited: cross-database queries need elastic query, SQL Server Agent is replaced by other scheduling mechanisms, and there is no native restore of an on-premises backup file. Deployment models are the single database, where each database has its own resources, and the elastic pool.
Purchasing models and service tiers
Azure SQL Database offers two ways to buy capacity, and the exam expects recognition rather than detail. The vCore model lets you choose the number of virtual cores and the amount of storage independently, which makes the sizing legible and allows the Azure Hybrid Benefit for organizations with existing SQL Server licenses. The older DTU model bundles compute, memory, and I/O into a single blended unit called a database transaction unit, which is simpler but harder to reason about when only one dimension is the bottleneck.
Within the vCore model, service tiers describe the storage and availability architecture: General Purpose for balanced workloads, Business Critical for low latency and high availability with local SSD storage and readable replicas, and Hyperscale for very large databases with rapid scaling and near-instant backups. Confirm current tier capabilities and limits in Microsoft's documentation before designing against them, as the specifics change.
Reading a requirement onto the spectrum
Work from the constraints in the question.
- Operating system access, a specific SQL Server version, or third-party software on the host — SQL Server on Azure Virtual Machines.
- SQL Server Agent, cross-database queries, Service Broker, linked servers, or native restore of an existing backup — Azure SQL Managed Instance.
- A new cloud application, minimum administration, automatic scaling, per-tenant databases, or a serverless model for intermittent use — Azure SQL Database.
Also note the two managed open-source engines, Azure Database for PostgreSQL and Azure Database for MySQL. If a question names PostgreSQL or MySQL rather than SQL Server, the answer is that service and not the SQL portfolio at all — a substitution the exam makes regularly to check you are reading the engine name rather than pattern-matching on the word "managed."
Frequently asked questions about the DP-900 training course
Is the DP-900 training course free?
Yes. The course costs nothing to read and opens once you are signed in to a free Certifym account, no payment and no card.
How is the course structured?
One module per official DP-900 exam domain, in Microsoft’s published order, with each module weighted to the domain’s published percentage range. Within each module the material is broken into short lessons, followed by key terms and further reading, the key terms matter here, because much of DP-900 is knowing which service name goes with which description.
Does this replace Microsoft’s official training?
No. Microsoft Learn publishes free official learning paths for DP-900, and the exam guide on Microsoft’s credentials site is the authoritative statement of what is examinable. This course is an independent study companion, written to be read quickly and to slot alongside practice questions.
Do I need database experience before starting?
No. DP-900 has no prerequisites and no requirement for hands-on production experience. If you have written a SQL query before you will move faster through module 2, but the course introduces relational concepts from the beginning rather than assuming them.
What should I do after finishing the course?
Work the DP-900 practice exam until you are scoring consistently across all four domains rather than carrying a weak Analytics Workload result on a strong Core Data Concepts one. Because scoring is scaled, leave yourself margin above the nominal cut before you book.
Is the course current with the latest DP-900 outline?
The course is built against the exam skills outline as currently published by Microsoft, with the four domains and their weightings as Microsoft lists them. Microsoft can revise the outline at any time; confirm the current exam guide, weightings, and logistics on Microsoft’s credentials site before you schedule.
Trademark notice & independence. Certifym.net is operated by Certifym Exam Services, LLC and is not affiliated with, endorsed by, or sponsored by Microsoft Corporation. Microsoft®, Azure®, DP-900, and the Microsoft Certified logos are trademarks or registered trademarks of Microsoft Corporation, referenced here solely to identify the certification these study materials target. The DP-900 exam objectives are the property of Microsoft Corporation; candidates should always confirm the current exam guide, weightings, and logistics directly on Microsoft’s credentials site.
All course content, questions, answers, and explanations on Certifym are original content created for study purposes. No Microsoft exam content, question text, or objective language is reproduced, and none of it is represented as official Microsoft training material. Studying with these materials does not guarantee a passing result on DP-900 itself. Exam requirements, format, domain weights, pricing, and renewal policies are set by Microsoft and may change; always verify current details on Microsoft’s credentials site before scheduling your exam.