What is intent in this context?

It is a high-level policy or objective guiding network operation rather than a list of settings for one device.

What is an autonomic service agent?

It is a component that implements an autonomic function, wholly or as part of a distributed function.

Can it coexist with manual management?

Yes. Practical systems can retain operator guidance, overrides and conventional functions.

AI Glossary / AI and machine learning / Intermediate

Autonomic networking

Definition

Autonomic networking is an approach in which network functions manage aspects of their own configuration, optimization, protection or recovery. They adapt to changing conditions within high-level objectives, reducing dependence on detailed manual administration.

What it means

A script can repeat a fixed sequence without human intervention. An autonomic function goes further by discovering relevant conditions and adapting its operation. The distinction concerns behavior, not whether the product uses a fashionable AI model.

People still define goals, operating limits and accountability. The network does not decide independently what an organization should value. A high-level intent might describe an operational objective, while distributed functions work out relevant details.

For business readers, the promise is less repetitive administration and faster handling of selected changes. The useful questions are which functions are self-managing, what information they use and how operators can understand or override their actions.

How it works

Network elements discover information about themselves and their environment. Autonomic service agents implement particular functions, sometimes cooperating across devices. They use shared intent and local knowledge to determine suitable behavior.

A function observes conditions, decides whether an adjustment is needed and acts within its authority. Monitoring checks whether the desired state has been reached. Secure communication and trust between participating components are important because false information or unauthorized instructions can affect many devices.

Different functions can interact. A change intended to improve one objective may undermine another, so coordination and conflict handling matter. Practical deployments can combine autonomic and conventional management rather than replacing every operator action at once. A claimed level of autonomy should identify the specific functions and exceptions covered.

Three practical examples

A network function discovers a device's role and derives appropriate settings from an approved high-level policy.
A service monitors changing conditions and adjusts an authorized parameter within defined bounds.
An operations team introduces self-management for one repetitive task while retaining manual handling of unusual cases.

Think of it this way

Think of a thermostat that works towards a chosen temperature while responding to changing room conditions. The user supplies an objective rather than issuing every switching command.

The analogy is limited because a network contains many interacting functions and objectives. It illustrates the difference between a fixed command sequence and a feedback process that adapts to observations.

Worked example

Distinguishing automatic from autonomic behavior

In an invented office rollout, Script A assigns the same fixed settings to every new device. When a device has a different role, an administrator edits the script and reruns it.

Function B discovers the device's approved role and derives settings from the applicable intent. When the environment changes within its supported conditions, it updates the settings and verifies the result. B illustrates a more autonomic pattern; A is useful automation but relies on manual adaptation.

Neither example establishes full network autonomy. Function B might still need help for an unknown role or conflicting policy. A meaningful evaluation counts those exceptions and checks whether its derived settings actually satisfy the intended requirements.

Illustrative case study

A fictional service provider automates a recurring configuration task. Early reports describe the network as self-managing, but operators still supply device-specific values for every installation. The team clarifies that the first stage is scripted automation.

It later adds discovery and policy-based derivation for a limited device family. Operators review unsupported configurations and can override the function. The service measures configuration errors, operator effort and recovery from interrupted setup.

The case illustrates a gradual, measurable transition. Using the term autonomic becomes more informative when it is tied to a specific function and its demonstrated ability to adapt.

Common misconceptions and mistakes

Autonomic means nobody remains responsible. People still set objectives and govern authority.
Every unattended script is autonomic. Fixed automation may require manual changes when conditions differ.
Autonomic networking necessarily uses machine learning. Self-management can be implemented through several techniques.

Business use

Autonomic networking can be relevant to large or frequently changing environments where manual detail becomes difficult to manage. Begin with well-understood functions and clear success measures.

Ask suppliers to explain discovery, intent, exception handling and operator visibility. Include the cost of maintaining policies and diagnosing interactions between functions. A reduction in manual commands is useful only if the overall service remains understandable and dependable. Evaluate actual operating behavior rather than a broad autonomy label.

Check your understanding

A setup script runs without supervision but must be rewritten whenever the device environment changes. Does unattended execution alone make it autonomic?

Answer: No. Autonomic behavior includes adaptation using relevant knowledge or intent, not merely automatic execution.

Related terms

Autonomic computing
Feedback loop
intent-based networking
autonomic service agent
self-configuration
self-healing
network automation

Learn next

Read intent-based networking and feedback loop. Select one repeated network task and identify which decisions could be derived from policy, which require observation and which need operator judgment.

Feedback loop

Sources

RFC 7575, Autonomic Networking: Definitions and Design Goals: https://www.rfc-editor.org/info/rfc7575/. RFC 8993, A Reference Model for Autonomic Networking: https://www.rfc-editor.org/info/rfc8993/. These primary technical documents describe definitions and architecture; RFC 7575 is informational rather than an Internet Standards Track specification. The rollout and provider examples are fictional.

Previous: Autonomic computing Next: Autonomous agent