Cross-Domain, Technology-Agnostic Orchestration (CDTAO)
This section is not part of the official DEVCOR certification; however, in the spirit of growing network programmability skills, it does seem appropriate to discuss. You may skip this section if you prefer.
Most work in network IT tends to be very domain-specific. It’s not unusual to see engineers and operators focus on specific technologies—enterprise networking, route/switch, wireless, storage networks, compute, wide-area networking, MPLS, security, and so on. However, many do embrace multidomain expertise.
Often the management applications follow a similar segmentation. It is easy to appreciate, then, when management apps bring a multidomain perspective to monitoring, provisioning, and management. However, consider why you’re doing IT: there’s a lot to support a business and the apps it depends on for the services it provides. Some typical supporting technologies include DNS, server connectivity, link aggregation, routing, switching, storage, compute, virtualized workloads, authentication, databases, firewall security, content filtering security, threat mitigation security, and application hosting. I’m sure you can think of many more!
So, is your operational perspective keeping up with the scope of your IT services? If you end up using multiple tools for different domains or scale or geographical segments or security segmentation, do you have an aggregate view of the health of your IT services, or do you switch back and forth between multiple tools? Doesn’t this issue get exacerbated when you pull in other IT vendors and open-source solutions? Is this something you accept as “the way it is” or do you try to “glue” together these systems for more unified operational insight?
How do you glue these systems together?
APIs are the unifying capability that enable you to achieve that glue. Most partner-vendors, Cisco included, strive to provide the best customer experience possible with their product and service offers. However, there are many customer segments, different sizes, different areas of concentration, and constraints. I have been asked, “Why isn’t there just one management tool?” Can you imagine the size in server requirements, cost, and maintenance necessary to provide such a solution? Would the broad functions, some of which don’t apply to your circumstances, distract your focus or enable it? In a friendly recognition to Hasbro, the movie series, and the legacy Cisco management suite, we would have to call it “Cisco Optimus Prime”! Most would agree that’s a bit unrealistic. Even building an uber-modular framework to allow the specific selection of desired functions and device support would increase complexity.
So is there an answer? Most providers enable their tools with APIs. If you pick the tools and apps you need based on function, need, cost, and preference, then you can obtain a converged operational perspective by using orchestration to collect the key health indicators from the individual tools and controllers. The orchestrator’s workflow would also include activities to create dashboards and portals unifying the information into converged operational portals that direct your attention to the domain-specific management tools, as necessary.
Is this possible? It’s not provided out of the box, again due to the variety of device types and functions, but it is doable. Consider the portal developed for the CiscoLive NOC in Figure 10-14. This example represents, essentially, a mashup of key health metrics from several tools: Prime Infrastructure, DNA Center, vCenter, Prime Network Registrar, Hyperflex, and so on.
Figure 10.14 NOC Dashboard
So what does the technology-agnostic part of Cross-Domain, Technology-Agnostic Orchestration (CDTAO) entail? It’s a wonderful concept to glue together network IT services in a cross-domain perspective. What about some out-of-the-box thinking that also brings in non-networking IT? From Figure 10-14, you can observe collaboration, digital signage, and NetApp storage. What other network-connected technology (think IoT) can be accessed and operational insight retrieved?
What industry do you work in?
■ Healthcare: Pull in network-connected systems, such as blood-pressure cuffs, pulse ox monitors, and crash carts.
■ Financial: Pull in ATM (cash, not legacy networking!), vault/deposit box status.
■ Retail: Fork lifts, credit card and point-of-sale terminals.
■ Education: Digital projector status, teacher location/availability, bus/parking lot, camera status.
If you add “business care-abouts” to the network IT perspectives, does that allow you to see contribution and impact of the supporting infrastructure to the broader company? Sure, it does!