Service

New fabric services from one request, applied consistently across devices

We build the orchestration that turns a fabric’s intent into device configuration: the fabric described as specs, validated before deployment, rendered for each hardware type, brought up by zero-touch provisioning and tested on a digital twin first. For a global e-commerce marketplace, the time to plan and deploy a large fabric fell from 3–4 weeks to 1–2 days.

Measured impact

What changes for your organization

  • Today

    Dozens of changes per service

    Every new service or rack in the fabric means manual changes on many devices, one command at a time.

    With intent-based orchestration

    One request per service

    Engineers change the intent, and the orchestration derives and applies the configuration on every device it touches.

  • Today

    Weeks per topology

    Fragmented tools and manual handoffs make every new fabric slow to define and deliver.

    With intent-based orchestration

    Fabrics in days

    Topologies are generated from declarative patterns, and scaling one means changing a handful of parameters.

  • Today

    Hardware tied to rework

    Swapping a switch for another vendor’s model means rewriting its configuration by hand.

    With intent-based orchestration

    Hardware you choose freely

    One configuration model with a renderer per platform, so a new hardware type means adding a renderer.

  • Today

    Configuration that drifts

    Build-time configuration goes stale within months, and hardware replacements fall back on backups.

    With intent-based orchestration

    One source of the fabric’s state

    The model is the record of the fabric, and every device gets its configuration from it at bring-up.

  • Today

    Designs proven on the floor

    A fabric design is tested only once the hardware is racked and cabled.

    With intent-based orchestration

    Designs proven before purchase

    Specs are validated and the whole topology runs as a digital twin before any device arrives.

What we deliver

  • The fabric as specs

    Layout, IP addressing, routing and hardware platforms described as layered specs, one concern at a time, with non-conforming specs rejected before deployment.

    Read the case study →
  • Topologies from patterns

    A topology combined with a connection pattern into an abstract graph, confirmed on a small fabric and scaled by raising leaf and interface counts.

  • One model, a renderer per hardware

    Every device’s configuration in one unified model, rendered at the end for each hardware type and network OS flavor.

  • Zero-touch bring-up

    Each device boots, fetches an agent and receives the configuration for its role and position in the topology, production-ready.

  • A digital twin of the fabric

    The whole topology emulated, routing policies included, so designs and changes are tested before any hardware is touched.

    Read the case study →
  • An API and room to extend

    A northbound API and a message bus, with your own logic added as services that start and stop independently of the rest.

How it runs

The scope follows your fabric, hardware and tools; the order of the work stays the same.

  1. Discovery

    Your fabrics, hardware and the way they are built today, the target model and architecture, and a plan.

  2. Model

    The fabric’s specs and the configuration model, confirmed on a small topology before anything scales.

  3. Renderers

    Configuration rendered for each hardware platform you run, validated against its constraints.

  4. Digital twin

    The topology emulated and tested, designs and changes included, before deployment.

  5. Zero-touch rollout

    Devices brought up from the model, starting with one fabric, then the next.

  6. Handover

    Your engineers extend the model and add renderers and logic themselves, with our support for as long as you want it.

Proof

Why CodiLime

  • Built with your engineers

    For a global e-commerce marketplace, our developers worked inside its network engineering team, and its engineers now extend the model.

  • Vendor-neutral by design

    One model with a renderer per platform, and one control plane over mixed legacy and new devices for a data center provider.

  • Tested before it is built

    Digital twins of a whole fabric for a marketplace, and of full customer SD-WAN networks for a global networking leader.

FAQ

Good to know

  • Do we have to run SONiC?

    Our largest fabric case runs on SONiC, and the design is vendor-neutral: each hardware platform and OS gets its own renderer from the same model. The discovery confirms the renderers your fleet needs.

  • Build or buy?

    The discovery holds your requirements against what the market offers. Where the full set describes a system nobody sells, we build it with your team, on a model your engineers can change.

  • What about the fabrics already in production?

    The model is confirmed on a small topology first, and existing fabrics move to it in stages. The digital twin shows the result before any device in production changes.

  • Do we need a source of truth first?

    The orchestration keeps its own model of the fabric, and it can sync with the source of truth or IPAM you run. Where you have none, the model takes that role for the fabric.

  • Who owns it afterwards?

    Your team. The model is written so network engineers can change it, and we stay on for new renderers, topologies and logic for as long as you want us to.

Start with a fabric orchestration assessment

Tell us how your fabrics are built and changed today. We map your topologies, hardware and tools, design the target model, and hand you a plan that starts with one fabric.