We’ve lastly arrived on the third and closing installment of this riveting weblog collection. Whereas some could also be unhappy on the disappearance of additional sleep they acquired from studying this (my prose is often an ideal remedy for insomnia), on this weblog, we’ll be masking the shiny new DCNAUTO specialization and applied sciences close to and pricey to my coronary heart. Similar to the weblog on AUTOCOR and Enauto 2.0I hope that this may assist make clear the rationale and intent for the big remodeling of the examination matters to help and help you in your research.
A fork within the highway
The unique DCAUTO examination had the “shortest” listing of examination matters (based mostly solely on my unscientific evaluation of the quantity of textual content on a PDF), however that doesn’t imply that the examination was easy. It coated a broad set of applied sciences with disparate terminology and spanned a number of (sometimes) separate groups (server/compute groups often are separate from the datacenter networking groups).
However even if you happen to had been in a corporation that had ACI and UCS, more often than not you’re employed with just one expertise or the opposite, not each. This complication was solely exacerbated by the truth that the Unified Computing System (UCS) Supervisor Platform Emulator (UCSM-PE) couldn’t be related to Cisco Intersight; solely sure builds which had been out there solely to particular groups like Cisco DevNet for his or her Sandbox might accomplish that.
This lead to an enormous inner resolution: How do we offer an automation certification that focuses on the datacenter, covers the community expertise out there at this time, contains platforms and units, and covers the evolving realities within the datacenter (like Kubernetes and containers)? We had some robust decisions to make, however the result’s the Dcnaut 2.0 (be aware the “N” for networking)
Give it to me straight, what has been faraway from DCAUTO?
Based mostly on this picture, you may see that a big chunk of the unique blueprint has been eliminated/modified ultimately(the highlighted sections). In some instances, the matters had been eliminated for a similar as they had been in ENCOR 2.0; matters like Git, fundamental APIs, or Python digital environments had been eliminated as a result of both (a) they’re assumed data (b) coated within the core examination or (c) will be changed with different applied sciences which will work higher with bigger workflows (e.g. growth within a container with mapped volumes can substitute digital environments inside Python).
Inside area 2.0, we eliminated most of the particular API and SDK duties as they pertain to ACI. Whereas these two strategies of automation are nonetheless legitimate, a lot of the event and integration effort throughout the datacenter has been targeted on Infrastructure as Code (IaC) instruments. With the ability to automate platforms and applied sciences with instruments which have multi-platform assist is necessary as a result of these datacenters are more and more heterogeneous. So understanding the right way to use these instruments throughout the community infrastructure turns into a crucial talent.
Area 3.0 acquired a light-weight contact of adjustments, principally targeted on refining and trimming down superfluous device-centric automation and app-hosting strategies. Whereas these capabilities are nonetheless built-in to our huge datacenter switching portfolio, we tried to give attention to the most typical use-cases and applied sciences. Keep in mind, the main target of the brand new blueprints is to create practicality and applicability into exams, so we needed to trim away a few of the esoteric or much less used options and performance.
And also you dropped compute?!
Sure.
I assume you’ll be searching for a purpose on this one, too. Consider me, it wasn’t a straightforward resolution. We went backwards and forwards on this and there have been sturdy arguments to each side, however finally, as a rule, the compute and server groups are utterly totally different than community infrastructure groups, and the practitioners inside these groups had vastly totally different skillsets, making the crossover to be that rather more tough.
Moderately than weakening the depth of the take a look at (and the sensible functions gained from it) to assist added breadth, we determined to drop the compute automation utterly. I can already hear the sighs of aid from community automation of us, however I do know there are a couple of of us that can miss the inclusion of Intersight and the UCSM APIs (my former compute Developer Advocate counterpart included).
Sufficient about what was dropped, what do we have to research?
Inside the datacenter, there are a couple of key applied sciences that we selected to give attention to. As with the AUTOCOR and Enauto 2.0, reference the highest paragraph of the examination matters listing to get an understanding of the in-scope platforms. These platforms shouldn’t come as a shock, nevertheless it’s useful to set context round your research.
Infrastructure as Code (IaC)
The datacenter have to be:
- Agile
- Multivendor
- Even multicloud
This implies click-ops or particular person automations for various platforms gained’t at all times be accepted. The unifying issue to all of that is one thing like Ansible or Terraform, whereby the syntax throughout platforms and clouds is identical and the one distinction is the modules/collections or suppliers in use.
The DCNAUTO examination displays this, as 25% of the examination falls throughout the IaC area. This requires you to be accustomed to the instruments and management options in addition to the platforms coated by the blueprint.
On-box automation and programmability
With the dimensions and scale of contemporary datacenter networks, platforms are sometimes used to handle the material. Nonetheless, there could also be both particular community automation options or day 0 provisioning that dictate a “box-by-box” course of. Due to this, we’ve included particular examination matters to validate a learner’s data round these “community aspect” automation duties in Area 3.
When it comes to particular community aspect programmability, we’ve included:
- NETCONF assist, as YANG fashions similar to OpenConfig are utilized in giant, doubtlessly multi-vendor or web-scale datacenters, because it normalizes configuration throughout quite a lot of units
- Familiarity with NETCONF and ncclient, which can be utilized to ship XML-structured payloads to a tool through code written in Python
- Understanding the day-0 provisioning of a tool exterior of the usage of a controller, and the on-box programmability strategies out there throughout the Nexus platform
- Data round NXAPI and the circulation of making bespoke templates (which might then be utilized as coverage) inside Nexus Dashboard rounds out the area
Operations (together with Linux Networking!)
One of many bigger shifts (throughout all new CCNP-Automation exams) has been the give attention to operational features of an automation resolution. In spite of everything, what good is deploying a change with out understanding the influence of that change on the community? That is no totally different throughout the datacenter and a few would argue that it’s extra necessary; datacenters are finely tuned devices to maneuver information in a short time from place to put. If it doesn’t work, it’s typically costing giant sums of cash.
On this examination, we’ve not a lot “eliminated” matters, however shifted them in complexity. The unique DCAUTO examination had parts that touched on model-driven telemetry and understanding subscriptions to information., together with next-generation protocols like gNMI and gRPC. We additionally embrace digital twins and pyATS validation, as we’ve got in different exams. To not be forgotten, we additionally cowl the power to retrieve well being data through Python towards units as effectively.
Lastly, we additionally added the requirement to troubleshoot packet flows from Linux-based hosts working containers. Everyone knows that containers are the brand new VMs, however the hosts working these containers don’t use the identical instruments and terminology as a Kind-1 Hypervisor; we should perceive how Linux networking works and the way it’s configured.
This contains how interfaces, subinterfaces, and bonded interfaces are created, in addition to how customary bridges are outlined and the connection between digital Ethernet (veth) interfaces on the host degree and interfaces outlined throughout the container runtime. These expertise are not non-obligatory and we felt it necessary to know them effectively sufficient to repair them once they break.
We needed to toss in some AI, too
Similar to with the remainder of the skilled automation specializations, some AI wanted to be included throughout the examination matters listing; it’s being talked about all over the place and our certifications needs to be no totally different.
- Understanding the safety implications of utilizing AI throughout the datacenter is necessary to guard the huge quantities and worth of that information. Right here there may very well be unintended penalties round information publicity and as a vector for exfiltration.
- As agentic AI turns into mainstream, understanding how these brokers join to numerous platforms, units, and controllers is a baseband process; one thing that everybody ought to perceive.
- With the prevalence of automation and orchestration throughout the datacenter, describing and understanding how generative AI can be utilized to speed up prototyping and iteration over community automation options will not be an non-obligatory talent. It ought to validated for any automation skilled.
Bringing all of it collectively
By way of this weblog, and the earlier ones on the AUTOCOR and ENAUTO 2.0, I hope you’ve gained just a little bit extra perception into the certification and the particular exams (each core and focus). This isn’t simply associated to the exams and matters themselves, but in addition the mindset shift and totally different strategy in creating the examination matters listing, shifting from software program engineers which might be studying “community” to community engineers which might be studying “automation.” It sounds delicate, however the consequence will be fairly totally different. By way of this distinction, we hope that you simply discover that the brand new exams align to your automation work in a way more impactful method.
As at all times, joyful studying! If in case you have any questions, please contact me on X (@qsnyder) or by way of the Cisco Studying Community message boards.
Efficient February 3, 2026, the 300-635 DCNAUTO examination might be up to date to v2.0 and renamed, “Automating Cisco Information Middle Networking Options v2.0.”
Join Cisco U. | Be a part of theCisco Studying Community at this time totally free.
Study with Cisco
X | Threads | Fb | LinkedIn | Instagram | YouTube
Use #CiscoU and#CiscoCert to hitch the dialog.
