Implementing Cisco Application Centric Infrastructure - Advanced (600-660 DCACIA) Exam Guide
The 600-660 DCACIA exam is identified by Cisco as Implementing Cisco Application Centric Infrastructure – Advanced v1.0 and evaluates advanced work with switches in ACI mode, including implementation, management, configuration, and troubleshooting. This guide is most useful to candidates deciding whether to study the published DCACIA blueprint or move to Cisco’s current replacement path. Because Cisco later announced the retirement of the related 300-630 DCACIA exam and specialist certification, confirm the active exam route before scheduling or purchasing preparation resources.
Check the exam’s current status before you study
Do not treat the published 600-660 blueprint as proof that a current appointment is available. Cisco’s exam-topics document identifies 600-660 DCACIA, while a later Cisco announcement says the 300-630 DCACIA exam and its associated specialist certification reached end of life on May 20, 2025, with May 20, 2025 identified as the last test date. Verify the current certification and scheduling information with Cisco before committing to this exam path.
The source material uses two different identifiers: the 600-660 code in the official exam-topics document and the 300-630 code in the retirement announcement and training document. That distinction matters. The facts supplied here support the published 600-660 blueprint and its 90-minute duration, but they do not establish that a 600-660 appointment remains available today.
Cisco recommended the 300-620 DCACI exam for learners interested in the retired DCACIA exam because 300-620 was updated to include key concepts and topics from it. A candidate starting now should therefore make a status decision first: follow an officially available 300-620 route, or use the DCACIA blueprint for a historical or employer-specific objective only after confirming eligibility and scheduling directly with Cisco.
What the exam validates
The published DCACIA objective is advanced operational knowledge of Cisco switches in ACI mode. The exam covers configuration, implementation, management, and troubleshooting, so preparation must connect policy intent with fabric behavior rather than stop at terminology or isolated menu navigation.
This is an implementation-and-diagnosis exam, not merely a product vocabulary test. You should be able to reason from an application requirement to an ACI policy design, predict how traffic should move, identify where a dependency exists, and isolate a failure when the observed behavior differs from the intended design.
The official document also cautions that its topics are general content guidelines and that related topics may appear on a specific exam delivery. Treat the blueprint as a boundary for study, not as a promise that every question will use the exact wording of a listed bullet. Build understanding around the listed technologies and their relationships.
Who should use this guide
This guide suits network engineers, data-center administrators, consultants, and Cisco professionals who already understand core ACI concepts and need to work through advanced connectivity, policy, multi-pod, and multi-site decisions. It is not an efficient starting point for someone who has not yet learned tenants, VRFs, bridge domains, EPGs, contracts, and fabric roles.
Experienced candidates should still test their foundations. Advanced troubleshooting often fails because the engineer misidentifies the object that owns a behavior: a contract may govern communication, a bridge-domain setting may influence forwarding, and an external network object may change how routes enter or leave the fabric.
Use the guide differently depending on your starting point. If you have production ACI experience, organize revision around failure isolation and design trade-offs. If you have mostly studied theory, prioritize controlled configuration exercises and packet-path tracing before attempting broad practice sets.
How the published blueprint divides study time
The published DCACIA v1.0 blueprint places the greatest listed emphasis on Advanced ACI policies and integrations at 25%, while ACI packet forwarding, Multipod, and Multisite each account for 20% and Traditional network with ACI accounts for 15%. Use those labeled domains to allocate study effort, but do not interpret the percentages as a pass-score formula.
Advanced ACI policies and integrations account for 25% of the published blueprint. This domain should receive the largest single study block. Its listed subjects include Layer 3 out transit routing, common tenants, VRF route leaking, contracts, and Layer 4–Layer 7 policy-based routing. Study each feature as a design mechanism and as a possible troubleshooting boundary.
ACI packet forwarding accounts for 20% of the published DCACIA v1.0 blueprint. The listed topics include VXLAN leaf-to-leaf forwarding, server NIC teaming, and endpoint-learning optimizations. Your preparation should explain not only what each feature does, but also what evidence would show that the expected forwarding or learning behavior is not occurring.
Multipod accounts for 20% of the published blueprint. The domain includes the IPN, inter-pod packet flow, firewall and load-balancer design, and service graphs. Practice separating fabric-wide policy problems from inter-pod transport problems and from service-insertion design errors.
Multisite accounts for 20% of the published blueprint. The listed subjects cover Multi-Site Orchestrator, ISN, stretched-component options, and communication across sites. Focus on scope, inter-site dependencies, and the consequences of choosing to stretch or keep components local.
Traditional network with ACI accounts for 15% of the published blueprint. Prepare to explain how ACI interacts with conventional network designs, where responsibilities sit between the fabric and external infrastructure, and how an apparently correct ACI policy can still fail because of an adjacent traditional-network dependency.
The five published domain percentages total the full blueprint allocation, but they do not tell you which individual questions will appear or how a particular delivery will be structured. They are most useful for prioritizing revision and identifying an underprepared domain, not for predicting a result.
Build the foundation before advanced policy work
Start with a dependency map of ACI objects and traffic direction. Before studying route leaking or service graphs, confirm that you can describe the relationship among tenants, VRFs, bridge domains, EPGs, contracts, external connectivity, and endpoint learning. Without that map, advanced features become memorized exceptions instead of explainable behavior.
Create a one-page diagram for each traffic case you study. Mark the source EPG, destination EPG, VRF boundary, contract direction, bridge-domain settings, external connection, and expected forwarding path. Then annotate the evidence you would collect if the traffic failed. This turns a conceptual topic into a repeatable diagnostic process.
Use active recall rather than rereading. Close the documentation and answer questions such as: Which object expresses the communication relationship? Where would a route be learned? Which boundary is being crossed? What must be true for the endpoint to be selected as a destination? Review the source only after recording your reasoning.
Keep a distinction between configuration intent and operational state. A policy can exist in the configuration while a fault remains in endpoint learning, route propagation, transport reachability, or service insertion. Your notes should have separate columns for intended policy, expected state, observed state, and next verification step.
Study ACI packet forwarding through packet paths
Learn packet forwarding as a sequence, not a collection of feature names. Trace how a frame or packet is classified, how the source and destination endpoints are identified, how the fabric selects a path, and where encapsulation or policy enforcement affects the journey. Then repeat the exercise for a failure at each stage.
For VXLAN leaf-to-leaf forwarding, draw the ingress leaf, egress leaf, endpoint locations, and the relevant fabric transport. Explain what information the fabric needs before it can forward toward the remote endpoint. If the destination is unknown or learned incorrectly, identify how that condition would appear in operational evidence rather than assuming the underlay is at fault.
Server NIC teaming deserves a separate worksheet. Compare the server’s teaming behavior with the ACI-facing design and note which assumptions must agree. A teaming mismatch can resemble a fabric forwarding problem, so include host-side validation in your troubleshooting sequence instead of checking only leaf configuration.
Endpoint-learning optimizations should be studied as behavior changes with diagnostic consequences. Record what the optimization is intended to improve, what endpoint information it influences, and which verification step would distinguish an optimization issue from a cabling, VLAN, policy, or transport issue.
A useful lab exercise is to begin with a working endpoint pair, document the normal path, and then change one dependency at a time. After each change, predict the symptom before checking it. The value is not the particular fault; it is learning to connect a controlled change with observable forwarding evidence.
Master advanced policies by designing the boundary
Advanced policy questions become easier when you identify the boundary being crossed. For every design, state whether the traffic crosses EPGs, VRFs, tenants, the ACI fabric edge, or a Layer 4–Layer 7 service. Then specify the policy object responsible for permitting, routing, leaking, or steering that traffic.
Layer 3 out transit routing requires more than knowing the feature name. Practice identifying the external routing relationship, the VRF context, the route direction, and the policy that allows the intended application path. Include failure cases where the ACI policy is present but the external peer, route, or next-hop expectation is wrong.
For common tenants, focus on why shared policy or shared services are being used and what scope that choice creates. Draw the consuming EPGs and the common service relationship. Then ask whether the design accidentally broadens access or makes troubleshooting ambiguous by hiding the true policy owner.
VRF route leaking should be approached as a controlled inter-VRF relationship. Write down the source VRF, destination VRF, route direction, permitted prefixes, and contract requirements. Do not reduce route leaking to a single configuration step; the exam-relevant reasoning is whether the complete path is both routable and authorized.
Contracts require precise direction and subject interpretation. For each contract exercise, label provider, consumer, filters, subjects, and expected traffic. Test both the intended flow and a similar flow that should remain denied. This prevents the common mistake of treating the presence of a contract as proof that every required packet is permitted.
Layer 4–Layer 7 policy-based routing should be studied with service intent and return traffic in mind. Diagram where the service is inserted, which traffic is classified, what the next hop or service function should be, and how the response returns. A one-way diagram is insufficient for a stateful service design.
Prepare Multipod as a transport and service-design problem
Multipod preparation should begin with the IPN and then move upward to inter-pod packet flow and service placement. Keep the IPN transport separate from tenant policy in your diagrams. When a flow fails between pods, first determine whether the fabric has the required transport reachability before attributing the symptom to contracts or application policy.
Draw a packet path that crosses pods and label each meaningful handoff. Identify the source and destination pod, the inter-pod transport, the relevant policy scope, and the expected return path. Repeat the exercise for local-to-remote and remote-to-local traffic so that asymmetric assumptions are exposed.
Firewall and load-balancer design requires placement reasoning. For each design, record whether the device is connected as a service, where traffic enters and exits, and which policy controls the insertion. Ask what happens if the device is reachable but the service graph classification is wrong, or if the policy is correct but the device’s own routing is incomplete.
Service graphs should be learned as traffic choreography. Describe the service node, the connectors, the traffic direction, and the relationship between the graph and the contract. A strong study note includes both the intended sequence and the verification evidence for each transition.
Avoid a common Multipod mistake: treating every cross-pod problem as an IPN problem. The IPN may be healthy while policy scope, endpoint placement, service insertion, or return routing is incorrect. Your troubleshooting order should eliminate broad transport causes before narrowing to application-specific policy.
Prepare Multisite around scope and communication
Multisite study is strongest when every object is assigned a location and scope. Mark what is local to a site, what is coordinated through Multi-Site Orchestrator, what depends on the ISN, and what is intentionally stretched. This prevents the vague assumption that a policy created centrally behaves identically in every site.
Include Multi-Site Orchestrator in an operational workflow: define the desired policy relationship, identify the sites involved, determine how the policy is deployed, and verify the resulting site-local state. The important preparation decision is knowing which layer owns the intent and which layer must be checked when behavior diverges.
The ISN should be tied to inter-site communication rather than memorized as an acronym. Diagram the connection between sites and list the dependencies required for the intended traffic. Then create a fault-isolation table distinguishing orchestration, inter-site transport, local fabric policy, and endpoint or external-network problems.
Stretched-component options require explicit trade-off notes. For each component, record why it might be stretched, what remains site-local, and what new dependency or failure domain the choice creates. The goal is to reason about design consequences, not to select a setting from memory without understanding the topology.
Practice a communication scenario in both directions. Identify the source and destination site, the policy relationship, the route or transport dependency, and the expected enforcement point. When you review the result, ask whether the issue is lack of policy, lack of reachability, incorrect scope, or an unsupported assumption about stretching.
Connect ACI to a traditional network
Traditional network with ACI is best studied at the boundary between fabric policy and external network responsibility. For every external-connectivity exercise, identify who owns addressing, routing, VLAN or interface configuration, security enforcement, and return traffic. This prevents a correct ACI configuration from being mistaken for a complete end-to-end implementation.
Build a responsibility matrix with columns for ACI, the external switch or router, security devices, load balancers, and servers. Add the expected control or data-plane evidence for each. When a test fails, use the matrix to choose the next owner to verify instead of repeatedly changing ACI policy.
Study designs in which the external network supplies a dependency that ACI cannot correct by itself. Examples include an incorrect route advertisement, a missing return path, a mismatched interface expectation, or a security device that does not permit the selected flow. The exact symptom is less important than the discipline of checking both sides of the boundary.
A useful exercise is to explain the same application flow twice: first from the application toward the external service, then from the response back to the application. Mark where policy is enforced in each direction. This exposes asymmetric routing and incomplete return-path analysis, two recurring causes of misleading troubleshooting conclusions.
Use a troubleshooting workflow instead of feature guessing
When an ACI scenario fails, begin by defining the expected flow and the first point where reality differs. Check scope, endpoint identity, policy relationship, routing, transport, and service insertion in an order that narrows the fault. Avoid changing several objects at once, because that destroys the evidence needed to identify the cause.
A practical sequence is: confirm the source and destination endpoints; verify their attachment and learning state; confirm the relevant EPG, bridge domain, and VRF; validate contract or service-policy requirements; inspect routing and external reachability; then examine inter-pod or inter-site transport when the topology requires it. Adapt the sequence to the failure rather than applying it mechanically.
For every topic, make a three-part note: symptom, likely boundaries, and discriminating evidence. For example, “remote endpoint unreachable” is a symptom, not a diagnosis. Candidate boundaries might include endpoint learning, leaf-to-leaf forwarding, IPN transport, route exchange, contract policy, or service insertion. The evidence column tells you which check separates them.
Practice explaining why a proposed fix is appropriate. If a question suggests adding a contract, ask whether the traffic is already blocked by routing or endpoint learning. If it suggests changing transport, ask whether local policy is the actual failure. This habit reduces attractive but unsupported fixes in scenario-based questions.
Do not prepare from exam dumps or leaked-question claims. They cannot establish understanding, may be inaccurate or unauthorized, and do not guarantee a passing result. Use the official blueprint, authoritative Cisco documentation, controlled practice, and your own troubleshooting explanations instead.
A practical study roadmap
A staged roadmap works better than treating all five domains as one reading assignment. First establish ACI object relationships and basic packet paths, then study the highest-weight policy domain, then work through Multipod and Multisite, and finish with mixed troubleshooting and traditional-network boundary cases.
Stage one should produce a working foundation. Review tenants, VRFs, bridge domains, EPGs, contracts, external connectivity, endpoint learning, and the distinction between configuration intent and operational state. Draw traffic flows until you can identify the likely policy and forwarding boundaries without consulting notes.
Stage two should concentrate on Advanced ACI policies and integrations, the 25% domain. Work through Layer 3 out transit routing, common tenants, VRF route leaking, contracts, and Layer 4–Layer 7 policy-based routing. For each topic, write one design explanation, one failure scenario, and one verification plan.
Stage three should cover ACI packet forwarding, Multipod, and Multisite as separate study blocks. For packet forwarding, trace VXLAN leaf-to-leaf behavior, server NIC teaming, and endpoint-learning optimizations. For Multipod, map the IPN, inter-pod flow, service graphs, firewalls, and load balancers. For Multisite, map Multi-Site Orchestrator, the ISN, stretched components, and site-to-site communication.
Stage four should integrate the 15% Traditional network with ACI domain into end-to-end cases. Use a responsibility matrix and trace forward and return traffic. Then combine two domains in each exercise, such as a cross-pod service graph or a multisite flow that depends on external routing.
Stage five should be an evidence-based readiness review. Without notes, choose a traffic scenario, draw its path, identify the policy objects, list dependencies, and describe a troubleshooting order. Any step that depends on memorized wording rather than an explanation should return to the relevant domain study block.
How to use practice questions responsibly
Practice questions are useful only when they test reasoning against the blueprint. After answering, explain why the selected option fits the topology and why each alternative fails. A score alone does not show whether you understand endpoint learning, contract direction, route leaking, IPN transport, or multisite scope.
Classify every missed question by cause: missing concept, incorrect object relationship, overlooked topology detail, policy-direction error, routing mistake, or rushed reading. Keep an error log with the corrected reasoning and the evidence that would confirm it in a real environment. Review patterns in the log rather than repeatedly retaking the same set.
Avoid memorizing answer positions or isolated phrases. Cisco states that the listed topics are general content guidelines and that related topics may also appear on a specific exam delivery. Preparation should therefore support transfer to a new scenario, not recognition of a copied question.
Use timed practice only after the concepts are stable. The published 600-660 DCACIA duration is 90 minutes, so eventually practice making a decision, recording the rationale, and moving on without spending the entire session on one uncertain item. Do not turn that duration into a claim about a current appointment until Cisco confirms the applicable exam route.
Delivery and training details that are actually evidenced
The official exam-topics document gives the published 600-660 DCACIA duration as 90 minutes. The supplied evidence does not establish a current delivery method, language list, price, question count, passing score, prerequisite, or scheduling availability, so confirm those details through Cisco rather than relying on catalogue pages or third-party summaries.
Cisco’s DCACIA training document states that the training prepared candidates for the 300-630 DCACIA v1.0 exam and provided 40 Continuing Education credits toward recertification. That fact belongs to the training document and should not be assumed to describe the 600-660 identifier or a currently available course without checking the current Cisco offering.
The retirement announcement changes the practical decision for new candidates. Since Cisco identified May 20, 2025 as the last test date for 300-630 DCACIA and recommended 300-620 DCACI as the updated route, a candidate should verify the exam code, certification objective, and recertification treatment before booking anything.
Common preparation mistakes to avoid
The most expensive mistake is studying a retired route without checking the official status. Resolve the identifier and availability question first. A well-prepared candidate can still lose time by following an old training plan when Cisco has directed learners toward an updated exam.
Another mistake is treating blueprint percentages as a question-by-question prediction. The published percentages identify domain emphasis: Advanced ACI policies and integrations at 25%, ACI packet forwarding at 20%, Multipod at 20%, Multisite at 20%, and Traditional network with ACI at 15%. They do not reveal exact question wording or guarantee a result.
Do not study features in isolation. Route leaking without contracts, service graphs without traffic direction, Multisite without scope, or endpoint learning without packet-path context produces fragile knowledge. Pair each feature with a topology, an expected flow, a failure symptom, and a verification method.
Do not troubleshoot only inside ACI. Multipod, Multisite, and traditional-network scenarios include transport, external routing, service devices, and server dependencies. A diagnosis that never checks the adjacent system is incomplete even when the ACI configuration appears plausible.
Finally, do not confuse a configured object with a working service. Readiness requires a complete path: endpoints, policy, routing, transport, service insertion where applicable, and return traffic. Make that complete-path check the final step in every lab and practice review.
Your next actions
First, confirm with Cisco whether your target is an active exam or a historical DCACIA objective. If you are starting after the announced retirement of 300-630, investigate the recommended 300-620 DCACI route before buying DCACIA-specific training or scheduling materials.
Next, download the official blueprint and turn its five labeled domains into a study tracker. Give the largest single block to Advanced ACI policies and integrations at 25%, then schedule dedicated work for ACI packet forwarding, Multipod, Multisite, and Traditional network with ACI according to their published domain weights.
Then build a small evidence notebook. For each listed topic, record the design purpose, dependencies, expected packet path, likely failure boundaries, and verification approach. Use diagrams and controlled exercises where possible, and mark gaps that you can describe only by memorized terminology.
Finally, perform a status and readiness check before scheduling: correct exam code, current Cisco availability, current objectives, current delivery details, and demonstrated ability to explain mixed-domain troubleshooting scenarios. If the official route has changed, carry forward the transferable ACI skills while revising against the current Cisco blueprint.
Conclusion
The published 600-660 DCACIA material supports a focused study plan built around packet forwarding, advanced policy, Multipod, Multisite, and traditional-network integration. However, the later retirement notice means status verification is part of preparation, not an administrative afterthought. Confirm the active Cisco path first, then study by traffic behavior and troubleshooting evidence rather than by memorized feature names. That approach preserves the useful ACI knowledge in the published blueprint while reducing the risk of preparing for an unavailable or outdated exam route.
