Using Device Model for Virtual DSP Processing Nodes - Valid Approach?

[Benni77]

Hey guys short question about use and misuse of the device driver model:
I'm designing a modular DSP pipeline system where multiple processing nodes (equalizers, compressors, filters, etc.) need to auto-register themselves and be discovered by an orchestrator component at runtime.
My question: Is it appropriate/recommended to use Zephyr's device model for purely virtual/software DSP processing nodes or even the complete pipeline, or is this considered a misuse of the framework?
The device model's capabilities to model complex DSP pipelines seems very very useful, but I want to avoid potential pitfalls or architectural mistakes.
Thanks for any insights!

I generated a super simple example of this:

/ {
    dsp_pipeline: dsp-pipeline {
        compatible = "company,dsp-pipeline-v1";
        sample-rate = <48000>;
        buffer-size = <256>;
        channels = <2>;
        status = "okay";

        input_gain: input-gain {
            compatible = "company,dsp-gain";
            gain-db = <0>;
            enabled = <1>;
            processing-order = <10>;
        };

        highpass_filter: hpf {
            compatible = "company,dsp-highpass";
            cutoff-frequency = <80>;
            q-factor = <0.707>;
            enabled = <1>;
            processing-order = <20>;
            input = <&input_gain>;
        };

        ...

};