Automating Cisco Data Center Solutions (DCAUTO): Exam Guide and Study Roadmap
The exam commonly called DCAUTO has moved to Cisco’s current 300-635 DCNAUTO, Automating Cisco Data Center Networking Solutions v2.0. It validates practical knowledge of data-center automation across infrastructure as code, network-element programmability, operations, and AI in automation. This guide is for engineers, automation developers, and certification candidates deciding whether their preparation should target the current blueprint or the legacy DCAUTO material. It also turns the blueprint into a study sequence, so you can choose the right labs, identify weak areas, and schedule the exam only after your skills match the current scope.
Which exam should you prepare for?
Prepare against Cisco’s current 300-635 DCNAUTO blueprint unless your registration specifically identifies a different exam version. Cisco’s current listing calls the exam Automating Cisco Data Center Networking Solutions v2.0, while the older exam-topics document calls 300-635 DCAUTO v1.1 Automating Cisco Data Center Solutions. The name change is not a minor wording difference: the current blueprint adds or emphasizes technologies and workflows that may not appear in legacy study material.
What changed from legacy DCAUTO?
The legacy v1.1 blueprint covered programming concepts, orchestration, and automation tools for implementing automated data-center solutions. Cisco’s current v2.0 scope is organized around Network Automation Foundation, Infrastructure as Code, Network Element Programmability, and Operations. It also identifies Cisco NX-OS, Cisco Nexus Dashboard, Cisco ACI in network-centric mode, and Cisco Nexus Hyperfabric as current technologies. Use the legacy document for historical context only, not as your main checklist.
If a course, book, video, or practice product still uses only the DCAUTO v1.1 title, compare its topic coverage with the current 300-635 DCNAUTO v2.0 document before relying on it. In particular, check for OpenConfig, YANG, NETCONF, gNMI, gRPC, gNOI, YANG Suite, pyang, Jinja2, Ansible, Terraform, GitOps, and troubleshooting workflows. Those are named in the current exam topics.
What does passing the current exam do for your certification plan?
Passing the current exam earns the Cisco Certified Automation Specialist – Data Center Automation and Programmability certification. The exam also satisfies the concentration-exam requirement for the CCNP Data Center and CCNP Automation certifications. Cisco states that the current exam can be used toward recertification requirements. Therefore, your first planning decision is not simply whether you can pass: determine whether the specialist certification, a CCNP concentration, or recertification is the outcome you need.
Who is the exam a sensible fit for?
The exam is most relevant to data-center professionals who automate Cisco infrastructure, developers who build network automation workflows, and engineers responsible for repeatable provisioning and operations. It is also suitable for a candidate combining networking knowledge with Python- or tool-based automation. The blueprint is less suited to someone who wants to memorize product terminology without building or troubleshooting automation.
You do not need to assume that a traditional network engineer and a software-focused engineer should study identically. The network engineer may need extra work on data models, APIs, and code structure. The developer may need extra work on NX-OS, ACI network-centric operation, Nexus Dashboard, and the behavior of the underlying data-center systems. Start with a gap assessment rather than choosing a single generic course.
How is the current exam structured?
The current 300-635 DCNAUTO exam is 90 minutes long. Cisco lists US$300 as the current exam price, with Cisco Learning Credits as an alternative payment method, and lists English and Japanese as available languages. Confirm these details on Cisco’s current exam page when you schedule because registration information can change. The official duration and blueprint should control your pacing and preparation decisions, not a legacy page or third-party summary.
What are the blueprint domains?
The current blueprint weights Network Automation Foundation at 15%, Infrastructure as Code at 25%, Network Element Programmability at 25%, and Operations at 25%. Each percentage belongs to the named domain; do not treat the percentages as interchangeable or compare them without their domain labels. The three 25% domains deserve comparable preparation time, while the 15% foundation domain still supplies concepts used across the rest of the exam.
Use the domain weights to allocate review effort, not to predict the number of questions. Cisco’s supplied facts identify the percentages but do not establish a question count, passing score, item format, or delivery method here. Avoid study plans that claim to reproduce the live exam or that assign a guaranteed number of questions to a domain.
What should you do with the blueprint?
Convert every topic in the official exam-topics document into an observable task. For example, “NETCONF” should become an ability to explain its role and work through a relevant automation flow; “Jinja2 templates” should become an ability to reason about variable substitution and rendered configuration; “troubleshooting Ansible- and Terraform-based automation” should become a structured diagnosis exercise. A topic is not ready merely because you can define it.
Mark each task as explain, configure, inspect, compare, or troubleshoot. That classification exposes shallow preparation. A candidate who can explain gRPC but cannot trace a failed workflow needs lab time. A candidate who can write a template but cannot identify an incorrect model path needs protocol and data-model review.
Which foundation topics deserve early attention?
Begin with the concepts that connect tools to network behavior: automation intent, structured data, APIs, models, authentication, state, idempotence, and error handling. The current blueprint names Network Automation Foundation as a domain, and the remaining domains depend on these ideas. Early foundation study prevents the common mistake of treating Ansible, Terraform, or a programmatic API as isolated products rather than parts of an automation system.
Build a vocabulary that supports decisions
Your notes should distinguish a desired state from an imperative sequence, a configuration model from a rendered payload, and a transport from an API operation. Record what each mechanism is used for, what representation it exchanges, and where failure can occur. For every term, add one operational question: What does the automation read? What does it change? How can you validate the result? What should happen if the target is already in the desired state?
Include the current technologies in the same mental map. Cisco NX-OS, Cisco Nexus Dashboard, Cisco ACI in network-centric mode, and Cisco Nexus Hyperfabric are not interchangeable labels. Study the automation context in which each appears in the current blueprint, then connect the platform to the model, interface, tool, or operational task being tested.
Do not overinvest in syntax memorization
Syntax matters when you must recognize a valid structure or troubleshoot a failure, but syntax without an execution model is fragile. Spend less time copying isolated snippets and more time tracing inputs, rendered data, transport, device response, validation, and rollback or remediation. This approach also helps when an item presents unfamiliar but conceptually related code.
How should you study infrastructure as code?
Treat infrastructure as code as a lifecycle rather than a collection of commands: define intent, represent it in code, review the change, plan or render it, apply it, inspect the result, and reconcile drift. The current blueprint explicitly includes Jinja2 templates, Ansible playbooks, Terraform plans, and GitOps. Your preparation should therefore cover both the artifacts and the decisions made around them.
Jinja2 and Ansible study tasks
Create a small data structure representing devices, interfaces, or policy values, then use a Jinja2 template to render the intended configuration. Change one variable and predict the resulting output before running it. Check how loops, conditionals, variable scope, and missing data affect the rendered result. The objective is not to memorize a favorite template; it is to understand how input data becomes a device-facing artifact.
For Ansible, trace inventory, variables, task order, modules or connection methods, result handling, and idempotent behavior. Practice reading a playbook and identifying whether it changes configuration, gathers facts, validates state, or delegates work. Then introduce a controlled error and determine whether the fault is in inventory, credentials, variables, the rendered content, transport, or the target response.
Terraform and GitOps study tasks
Use Terraform study to separate configuration, plan, apply, state, and drift. Be able to explain what a plan is intended to show and why a planned change may differ from the live environment. Review how a resource definition expresses desired state and how an operational team should inspect an unexpected change rather than applying it blindly.
For GitOps, map the path from a version-controlled change through review and automation to the target environment. Identify where validation, approval, secrets handling, and reconciliation belong in that path. The exam topics include GitOps and troubleshooting Ansible- and Terraform-based automation, so practice diagnosing the entire workflow instead of studying Git commands as an unrelated subject.
What must you know about network-element programmability?
Network-element programmability is a major current domain, not an optional protocol appendix. Study the relationship among models, encoding, transport, RPCs, and device behavior. The current topics include OpenConfig, IETF and native YANG models, NETCONF, gNMI, gRPC, gNOI, YANG Suite, and pyang. Organize these into workflows so you can select or evaluate an approach rather than merely expand acronyms.
Use one comparison table of your own
Build a table with columns for model source, transport or protocol, operation type, data representation, typical purpose, validation method, and likely failure points. Populate it from the official topic list and your study material. The act of filling the table reveals confusing overlaps, especially when a tool, model, transport, and RPC are discussed together.
Compare OpenConfig, IETF, and native YANG models by asking what each model represents and how you would inspect or validate it. Then compare NETCONF and gNMI as mechanisms used to interact with modeled data. Keep the distinctions precise: do not collapse a model into a transport or describe a tool as though it were the device protocol.
Practice model inspection
Use pyang or YANG Suite in a lab or learning environment to inspect a model, identify hierarchy and types, and reason about whether a requested path or value is valid. Trace how a model becomes an operation against a network element. When an operation fails, check the model path, namespace or encoding assumptions, capability support, authentication, transport, and device-side response in that order.
For gRPC and gNOI, focus on their role in an automation workflow and on the evidence you would use to diagnose an unsuccessful call. Do not rely on a memorized command sequence if you cannot explain what the request is asking the target system to do and how you would confirm the outcome.
How should you prepare for operations and troubleshooting?
Operations preparation should make you comfortable diagnosing a failed automation run from evidence. The current blueprint gives Operations a 25% weighting and includes troubleshooting Ansible- and Terraform-based automation. Practice separating an authoring defect from a data defect, an authentication problem, a transport failure, a target capability issue, and a post-change validation failure.
Follow a fixed troubleshooting order
Start with the intended outcome and the exact change. Inspect the input variables, inventory or resource definitions, rendered payload, selected target, credentials, and connection details. Next inspect the tool output, device response, resulting state, and any plan or state information. This sequence reduces random edits and helps you identify the first point at which actual behavior diverged from intent.
Record the symptom, evidence, hypothesis, test, and result in a short troubleshooting log. For example, if a playbook reports success but the target state is wrong, do not immediately rewrite the task. Check whether the task changed the intended object, whether the response was interpreted correctly, and whether validation examined the same state that the task modified.
Practice safe operational judgment
A good automation engineer knows when not to apply a change. Review changes before execution, validate assumptions, isolate test targets, and make the resulting state observable. When a plan or rendered configuration is unexpected, stop and inspect the source rather than trying repeated executions. These are practical recommendations, not additional Cisco exam requirements, but they create the reasoning habits the operations domain tests.
Where does AI in automation fit?
Cisco’s current exam listing says the exam covers AI in automation. Treat this as a current-scope signal and verify the detailed treatment in the current exam-topics document and Cisco learning materials. Prepare to reason about how AI-assisted automation fits into an engineering workflow, including validation, human review, data quality, and operational accountability, without assuming that generated output is correct.
Use AI as an object of evaluation, not a shortcut
When studying an AI-assisted workflow, ask what input it used, what output it produced, how the output is checked, and what safeguards prevent an unsafe change. Compare generated code or configuration with the intended model and target state. This is a practical recommendation; it is not a claim about undisclosed exam item formats or a promise that a particular AI tool appears in the exam.
What lab environment is enough?
Build the smallest environment that lets you observe the complete automation loop. You need a way to work with structured data and templates, execute or inspect Ansible and Terraform artifacts, examine YANG models with YANG Suite or pyang, and reason about Cisco data-center platforms named in the blueprint. Use official training or lab resources where available, and do not imply that a specific commercial or hosted lab is required by Cisco.
Design labs around failure and recovery
A successful run teaches execution; a failed run teaches diagnosis. For each lab, deliberately alter one input, model path, variable, target value, or access assumption. Capture the error, locate the first failed layer, correct the cause, rerun safely, and verify the resulting state. Keep the exercise bounded so that you can repeat it and explain why the correction worked.
If you cannot access every named platform, separate platform-specific facts from transferable automation skills. Study Cisco’s documented behavior and terminology for the platforms you lack, while using an available environment to practice models, templating, plans, validation, and troubleshooting. Do not claim hands-on proficiency with a platform you have only read about.
Keep a lab notebook that mirrors the blueprint
Create four headings in the notebook: Network Automation Foundation, Infrastructure as Code, Network Element Programmability, and Operations. Under each, record the task, evidence, error encountered, correction, and remaining question. Add a separate technology index for NX-OS, Nexus Dashboard, ACI in network-centric mode, and Nexus Hyperfabric. This makes revision targeted and prevents familiar tools from crowding out weaker domains.
What should a six-stage study roadmap look like?
A staged plan is more reliable than switching tools every day. First establish the current version and baseline your skills. Next learn the foundation, then infrastructure as code, then network-element programmability, and finally operations and integrated troubleshooting. End with blueprint review and scheduling checks. Adjust the pace to your availability; the sequence matters more than an invented calendar.
Stage 1: confirm scope and measure gaps
Download and read the current 300-635 DCNAUTO v2.0 exam-topics document. Write each domain and topic in a checklist, then mark each item green, amber, or red. Green means you can explain and demonstrate it; amber means you recognize it but need practice; red means you cannot yet connect it to an automation workflow. Remove legacy-only notes from your primary checklist unless you are studying historical material.
Stage 2: establish foundation concepts
Review structured data, models, intent, state, APIs, authentication, idempotence, and validation. For each concept, write a short explanation and one failure example. Link the concept to a platform or workflow from the current scope. Finish this stage only when you can explain why a particular tool or protocol is appropriate, not merely what its name expands to.
Stage 3: build infrastructure-as-code fluency
Work through Jinja2 rendering, Ansible playbook execution, Terraform planning and state reasoning, and GitOps flow. Alternate reading with execution: predict the result, run or inspect the artifact, compare the result, and document the discrepancy. Include at least one troubleshooting exercise for both Ansible- and Terraform-based automation because the current topics name those troubleshooting skills explicitly.
Stage 4: work through programmability methods
Study OpenConfig, IETF, and native YANG models, then connect them to NETCONF, gNMI, gRPC, and gNOI. Use YANG Suite and pyang for model inspection where possible. Your checkpoint is a diagram or written explanation showing how a model, request, transport, target, response, and validation step relate to one another.
Stage 5: integrate operations
Take a complete desired-state change from source data through execution and validation. Break it, collect evidence, and repair it. Repeat with a different layer as the fault. Include the Cisco technologies named in the current blueprint in your review, but keep the exercise focused on a specific operational decision rather than trying to build an oversized demonstration environment.
Stage 6: perform a readiness review
Revisit every amber and red checklist item. Explain each domain without notes, inspect unfamiliar examples, and troubleshoot representative failures under a time constraint that is shorter than the official 90-minute exam duration. This is a practice recommendation, not a prediction of the live exam. Schedule only after you can justify readiness with demonstrated tasks rather than confidence from repeated memorization.
Which preparation mistakes create false confidence?
The most damaging mistakes are version confusion, passive reading, tool worship, and ignoring troubleshooting. A candidate can recognize many product names yet fail to trace a change from intent to device state. Correct these problems by tying every study item to an observable task and by using the current blueprint as the controlling scope.
Mistake: studying the legacy title without checking the version
Legacy DCAUTO v1.1 material is not automatically a guide to current DCNAUTO v2.0. Cisco’s current listing uses the DCNAUTO title, and the current blueprint names technologies and topics that must be checked explicitly. Keep the v1.1 document only as background or as a way to understand older references in training material.
Mistake: memorizing dumps or recalled questions
Exam dumps, leaked questions, and memorized answer sets do not establish competence or guarantee a passing result. They can also direct study toward obsolete or inaccurate content. Use the official blueprint, legitimate Cisco learning resources, and your own labs. Practice explaining why an answer is correct and how you would validate the corresponding automation behavior.
Mistake: treating a successful script as proof of readiness
A script that works once may still be non-idempotent, poorly validated, dependent on hidden variables, or unsafe when state changes. Repeat the task, inspect the resulting state, alter an input, and troubleshoot a controlled failure. Readiness includes recognizing when automation should stop and what evidence is needed before a retry.
Mistake: ignoring scheduling details until the last moment
Before booking, confirm that your registration shows the current exam, review Cisco’s current price and language information, and check whether Cisco Learning Credits apply to your situation. The supplied official information does not establish a delivery method or every scheduling rule, so use Cisco’s live registration process for those details.
What should you do in the final review week?
Stop adding unrelated tools and concentrate on retrieval, comparison, and diagnosis. Re-read the current blueprint, review your error log, and rehearse the complete automation lifecycle. Your final review should answer practical questions: Which model or interface is involved? What state is intended? What evidence proves the change worked? Where would you look first when it did not?
Use short, evidence-based review sessions
For each session, choose one domain and one concrete task. Explain the task aloud or in writing, inspect a relevant artifact, and resolve one fault. Finish by recording the remaining uncertainty. This produces a useful final list of questions instead of a vague sense that more reading is needed.
Do not spend the final review trying to predict exact questions, scores, or item counts. Cisco’s official facts supplied here confirm the duration, price, languages, certification uses, and domain weights, but they do not provide those predictions. Prepare for the capabilities described by the blueprint.
Make the scheduling decision deliberately
Schedule when your checklist shows demonstrated ability across all four domains, with particular attention to Infrastructure as Code, Network Element Programmability, and Operations, each weighted at 25%. Confirm the current exam name and registration details immediately before booking. If a major topic remains red, postpone or revise the study plan rather than relying on familiarity with older DCAUTO material.
Which official sources should remain open while you study?
Use Cisco’s current exam listing for the current title, duration, price, languages, certification outcomes, concentration-exam use, and recertification information. Use the current v2.0 exam-topics PDF for the blueprint domains, percentages, technologies, protocols, tools, and workflows. Keep the legacy v1.1 PDF and DCAUI course page only when you need to identify older naming or historical course information.
Source discipline protects your preparation
Third-party summaries can lag behind a version change. When two pages disagree, check the current Cisco exam listing and current exam-topics PDF first. Do not transfer legacy claims about certification outcomes or course requirements to the current exam without current support. The legacy DCAUI page states that its prerequisite was Introducing Automation for Cisco Solutions and that it awarded 24 Continuing Education credits toward recertification; those facts belong to that legacy course context, not automatically to current DCNAUTO preparation.
Conclusion
Treat DCAUTO as a version-control and skills problem before treating it as a booking problem. Confirm that your target is Cisco’s current 300-635 DCNAUTO v2.0, organize study around the four named domains, and give equal practical attention to the three 25% domains while building the 15% foundation into every lab. Then verify models, protocols, tools, platform context, and troubleshooting through observable work. Your next action is to download the current blueprint, mark your gaps, and begin with the first red task rather than another generic question set.
Related exams
- 300-610 exam — Designing Cisco Data Center Infrastructure (DCID)
- Troubleshooting Cisco Data Center Infrastructure (300-615 DCIT)
- Implementing Cisco Application Centric Infrastructure (300-620 DCACI)
- 300-630 exam — Implementing Cisco Application Centric Infrastructure - Advanced (DCACIA)
- Implementing Cisco Data Center Core Technologies (350-601 DCCOR)
