Have you ever asked what is network orchestration? It’s a pretty fancy title. All it really is though is a technique to push commands down into a network to change its configuration. Sounds simple.

However, don’t let that description imply that it’s easy. In some cases, it can be relatively easy – if this happens, then do that (IFTTT). If a new customer order arrives in a queue, then push a command (and associated attributes) to the network to allow the order to be activated. If the performance of a given network link goes outside expected bounds, then apply some sort of traffic shaping policy.

That’s the theory. In practice, it tends to be more complicated than IFTTT. More often, it includes a workflow with conditional decisions or other activities that need to be performed in a prescribed order. A service activation needs steps such as credit checks, resource availability checks, billing activations, customer notifications and more. Not only that, but the processes often access multiple adjacent systems, have different policies / configurations to apply to different equipment makes/models and need careful sequencing of activities, especially across multi-domain networks. It then also needs a rollback sequence to cope with any unexpected failures or constraints along the way.

Modern network orchestration can improve your top and bottom lines. But how?

These processes are known as orchestration plans. However, the orchestration plan doesn’t just coordinate and order the flow sequence, It also coordinates the interfaces / APIs that need to be invoked, as well as the parameters that are passed across each interface.

Speaking of process instances, network orchestration solutions generally need to operate in two distinct modes. First is design-time mode. This is when network operators define the framework for orchestration. They need to define and design the orchestration plans, with all the sequences, variables, interfaces and conditionals described above. It’s like designing the factory with all its conveyor belts, machines and robots.

Then once the orchestration plan has been designed, tested and commissioned, the network orchestration tool is used in run-time mode. That is, it then moves into production and runs the orchestration plan over and over for each instance of a process (e.g. a customer order type). This is like a factory producing many widgets. In the case of network orchestration, it probably produces many different variants depending on the sophistication of the orchestration plan and the conditionals within it.

Network orchestration. Unleashing a wave of efficiency gains in your organization

It’s the repeatable nature of run-time mode that allows a network operator to achieve great efficiency, stamping out “products” as if coming off an automated conveyor belt.

But we should point out that not every step in an orchestration plan is automated. Ideally each step will be, but often there are manual activities to be performed. Two neighbours might order the exact same service off a carrier within a minute of each other. One might have their service activated almost immediately because the operator already has infrastructure installed to their premises and it’s just a case of performing automated activation tasks. The second neighbour might have to wait weeks for their service to be activated because their premises doesn’t have an existing lead-in cable or network termination device yet. The orchestrator has to initiate physical build / provisioning activities by the field service team. The same orchestration plan was invoked for both neighbours, but different paths needed to be followed by each, resulting in different activation times.

Are you losing control of your service activations? Your network orchestrator could be to blame

For each process instance, there can be many different activities performed as it moves through the orchestration plan from start to finish. There can be many different process instances in-flight at any one time. In parallel, there can be many different orchestration plans being invoked. Combined, there can be many thousands of process steps underway at any point in time. The orchestration solution is responsible for coordinating every single one of these activities, optimising the streams to ensure they’re flowing efficiently, without lapsing into a fall-out state or falling into jeopardy (i.e. when a process takes longer than the approved duration or SLA to complete).

As the name implies, a network orchestrator is like the conductor of an orchestra. It needs to coordinate all of the individual resources (devices, interfaces, applications, connections, services, etc) to be applied at exactly the right time in exactly the right order to ensure happy customers.


SunVizion Service Order Management includes a network orchestration function that manages processes such as Order to Activate (O2A) or Quote to Cash (Q2C) that are responsible for generating revenues for network operators. If your current O2A or Q2C workflows are problematic, with fall-outs, jeopardy events or sub-optimal ready-for-service (RFS) metrics, then consider whether a new orchestrator can deliver top and bottom-line benefits to your organisation.


We’d also like to highlight that network orchestration isn’t just used for service activations. They’re also often used as the feedback mechanism for closed-loop network optimization activities. They can be really effective network automation tools. For example, if a network service goes outside expected thresholds, a control signal can be issued to the orchestrator to implement control policies into the network to bring it back within required tolerances.