Managing a utility or infrastructure network means dealing with data that is spread across large geographic areas, connected through complex relationships, and constantly changing. At the heart of making sense of that data is routing: the ability to trace paths, understand connectivity, and analyze how your network actually behaves. Choosing the right routing model is one of the most impactful decisions you can make in your GIS setup, and it directly shapes the quality of your spatial analysis capabilities and outcomes.

Whether you work in water distribution, gas transmission, electricity grids, or telecommunications, the routing model you select determines what questions you can answer—and how accurately you can answer them. This guide walks you through the key questions you need to ask before making that choice.

What is a routing model in network analysis? #

A routing model in network analysis is a structured framework that defines how connectivity, flow, and traversal are calculated across a network dataset. It establishes the rules for how data moves through your network, which paths are valid, and what constraints apply at each node or edge. In short, it turns your geographic data into a connected, queryable system.

Without a routing model, your network is just a collection of lines and points on a map. With one, those lines and points become a living representation of your infrastructure, capable of answering operational questions such as: Where does this pipe lead? Which assets are affected if this valve closes? What is the shortest path to reach this fault location?

Routing models typically define three core components: network elements (edges and junctions), connectivity rules that determine how those elements relate to each other, and attributes that influence traversal, such as flow direction, capacity, or impedance. Getting these three components right is what makes your routing model useful in practice.

What types of routing models exist for utility networks? #

The main types of routing models for utility networks are geometric networks, utility networks, and general-purpose network datasets. Each serves a different purpose and suits different infrastructure types. Geometric networks are designed for flow-based systems, utility networks offer a more modern, rule-driven approach, and network datasets are built primarily for transportation-style routing.

Geometric networks #

Geometric networks have been a standard in GIS for utility management for many years. They model flow direction and connectivity across infrastructure such as water pipes or electrical cables, making them well suited for tracing upstream and downstream relationships. They work well when your primary need is understanding how resources move through a connected system.

Utility networks #

Utility networks represent a more advanced model that supports tiered structures, containment relationships, and more granular connectivity rules. They allow you to model complex infrastructure hierarchies, such as a cable running inside a conduit inside a duct bank—something a geometric network cannot represent cleanly. They also support subnetworks, which makes them useful for managing distinct operational zones within a larger grid.

Network datasets #

Network datasets are best suited for transportation and logistics use cases, such as routing service vehicles or calculating travel times. They are less appropriate for modeling the physical and logical behavior of utility infrastructure, where flow direction and containment matter more than travel cost.

What’s the difference between geometric and utility network routing? #

The core difference between geometric and utility network routing is structural complexity. A geometric network models simple, connected flow across a flat topology, while a utility network supports hierarchical containment, tiered subnetworks, and more sophisticated connectivity rules. Utility networks are designed for the realities of modern infrastructure, while geometric networks handle simpler flow-tracing tasks.

In a geometric network, every feature exists at a single level. A pipe connects to another pipe through a junction, and the system traces connectivity based on those simple relationships. This works well for straightforward distribution networks, but it struggles when your infrastructure has layered physical structures or when you need to model logical relationships separately from physical ones.

A utility network separates the physical layer from the logical layer. You can model the physical containment of cables inside conduits while simultaneously managing the logical connectivity of the circuits those cables carry. This distinction is particularly useful for telecommunications and electricity providers, where the physical path and the logical signal path often differ.

The trade-off is complexity. Utility networks require more upfront data modeling and schema design. If your infrastructure is relatively straightforward and your team is already comfortable with geometric networks, the migration cost may not be justified by the added capability.

How does network topology affect routing model selection? #

Network topology directly determines which routing model can accurately represent your infrastructure. Topology defines how features connect, whether connections are planar or non-planar, and whether flow has a defined direction. A routing model that does not match your network’s topological structure will produce incorrect traces and unreliable analysis results.

Planar networks, where all connections occur at intersections and features cannot cross without connecting, are common in road networks and simple pipe networks. Non-planar networks, where cables or pipes can cross without connecting, are common in telecommunications and electricity transmission. Your routing model needs to reflect this distinction; otherwise, the system will assume connectivity where none exists.

Flow direction is another topological factor. In water and gas networks, flow follows pressure gradients and can change direction based on operating conditions. Your routing model needs to support dynamic flow direction if your network operates this way. A model that assumes fixed flow direction will give you incorrect upstream and downstream traces during abnormal operating conditions.

Connectivity rules also play a topological role. If your network has strict rules about which asset types can connect to which, your routing model should enforce those rules. This prevents data quality issues from corrupting your analysis and ensures that traces reflect real operational constraints.

Which routing model is best for your infrastructure type? #

The best routing model for your infrastructure depends on three factors: the complexity of your connectivity rules, whether you need to separate physical and logical layers, and the maturity of your existing GIS data. There is no single right answer, but matching these factors to the model’s capabilities will point you in the right direction.

• Water distribution networks: Geometric networks and utility networks can both work well. If you need to model pressure zones and subnetworks separately, a utility network gives you more control. For simpler tracing and outage-impact analysis, a geometric network is often sufficient.
• Gas transmission networks: The need to track asset condition, model flow under varying pressure scenarios, and integrate sensor data makes utility networks a stronger choice. The ability to integrate separate data sources and apply rules adds real operational value.
• Electricity distribution: Utility networks are well suited here because they support the separation of physical cables from logical circuits, and they handle the tiered substation hierarchy that electricity grids typically require.
• Telecommunications: The combination of physical infrastructure (conduits, cables) and logical services (circuits, connections) makes utility networks the preferred model. The containment and association relationships they support map directly onto how telecommunications infrastructure is actually built.
• Government and multi-utility: When you manage multiple infrastructure types on a single platform, a utility network’s flexibility and rule-based architecture provide the consistency you need to maintain data quality across diverse asset types.

How do you implement a routing model in an existing GIS environment? #

Implementing a routing model in an existing GIS environment involves four main steps: auditing your current data quality, defining your connectivity rules and schema, migrating or restructuring your data to match the new model, and validating the results through network traces and analysis. The process requires careful planning, but it does not have to disrupt your day-to-day operations if you approach it incrementally.

Start with a data quality audit #

Before you define any routing rules, assess the state of your existing spatial data. Gaps in connectivity, inconsistent attribute values, and misaligned geometries will all undermine your routing model, regardless of how well it is designed. Automated data quality tools can help you identify and prioritize these issues before migration begins.

Define your schema and connectivity rules #

Work with your network operators and GIS team to document the rules that govern your real-world infrastructure. Which asset types can connect to which? What flow directions apply under normal and emergency conditions? What subnetworks or zones need to be modeled separately? These decisions form the foundation of your routing model schema.

Migrate and validate incrementally #

Rather than migrating your entire network at once, consider a phased approach by region or asset type. This lets you validate routing behavior against known operational scenarios before committing to a full rollout. Running parallel traces in your old and new models on the same section of the network is a practical way to catch discrepancies early.

Integrate with your broader spatial analysis workflow #

A routing model delivers its full value when it connects to your wider spatial analysis capabilities. Our tools at Spatial Eye are built to support native data access, meaning you can connect to your network data directly, apply routing and topology functions, and produce reports and visualizations without needing to extract or duplicate data. This keeps your routing model aligned with your operational data at all times, so the insights you generate reflect the current state of your network rather than a snapshot from weeks ago. Contact our team to get started.