Routing is the network-layer function that selects a next hop for an IP packet and forwards it toward its destination using routing tables and routing policies.
In practice, routing decides “where to send this packet next” until it reaches the destination subnet.
When to use routing
Use routing when you need to:
- Connect multiple IP subnets (e.g., VLAN-to-VLAN, VPC-to-VPC, site-to-site).
- Provide Internet connectivity or interconnect separate networks (LAN ↔ WAN).
- Support redundancy with multiple paths (failover, ECMP).
- Control traffic paths with policies (e.g., send some traffic via a security appliance).
- Scale a network using dynamic routing protocols (OSPF/IS-IS in enterprises, BGP in ISPs/Internet).
When not to use routing
Routing may be unnecessary or counterproductive when:
- All devices are in the same Layer 2 segment/subnet (switching/bridging is enough).
- You only need name resolution or service discovery (DNS/SD) rather than path selection.
- You’re trying to fix application latency that is actually caused by server performance, not network path.
- You can solve the problem with static routes and the network is small and stable (dynamic routing adds complexity).
- Your security model requires no direct network reachability (prefer proxies, gateways, or zero-trust access patterns).
Signals you need routing (symptoms)
You likely need routing (or your routing is wrong) if you see:
- Hosts can reach their local subnet but not other subnets/VLANs.
- Cloud workloads in different VPCs/VNETs can’t communicate without a transit.
- Traceroute stops at a gateway or shows unexpected hops.
- Traffic works in one direction but fails in the other (asymmetric routing).
- Intermittent reachability during failures (slow convergence, flaps).
- Stateful firewalls/load balancers fail because return traffic takes a different path.
Key components
Routers (what they do)
Routers receive packets, inspect IP headers, and forward packets between networks. They may also apply policies (ACLs, QoS, PBR) and participate in routing protocols.
Routing tables (what decisions are based on)
A routing table is a list of destination prefixes and how to reach them.
Typical fields:
- Destination prefix (e.g., 203.0.113.0/24)
- Next hop (router IP or directly connected)
- Outgoing interface
- Metric / preference (cost)
- Route source (static, OSPF, BGP, etc.)
Routing algorithms and protocols (what changes routes)
Algorithms (the logic)
- Distance Vector (e.g., RIP): learns routes from neighbors; simple metrics like hop count.
- Link-State (e.g., OSPF, IS-IS): builds a topology map; computes shortest paths (Dijkstra).
- Path Vector (BGP): exchanges reachable prefixes plus AS-path; supports policy-based decisions.
Protocols (the communication)
- IGPs (inside one organization/AS): OSPF, IS-IS, RIP.
- EGP (between autonomous systems): BGP (the Internet’s routing protocol).
The Role of Routers
Routers are the backbone of routing in computer networks. They are responsible for receiving, analyzing, and forwarding data packets between different networks or subnets. A router typically consists of hardware components, such as a processor, memory, and network interfaces, as well as software components, including an operating system and routing protocols.
Routers have multiple interfaces, each connected to a different network segment. These interfaces can be physical ports, such as Ethernet or serial ports, or logical interfaces, such as VLANs (Virtual Local Area Networks). Routers use these interfaces to send and receive data packets and make forwarding decisions based on the information contained in the packets’ headers.
Routing Tables
At the core of a router’s functionality are routing tables. A routing table is a data structure that contains information about the network topology and the best paths to reach different network destinations. Each entry in a routing table typically includes the following information:
- Network address: The IP address or prefix of a network or subnet.
- Next hop: The IP address of the next router or gateway to which packets destined for the specified network should be forwarded.
- Metric: A value that represents the cost or preference of the path, used to determine the best route when multiple paths exist.
When a router receives a data packet, it examines the packet’s destination IP address and consults its routing table to determine the best path for forwarding the packet. This process, known as a routing table lookup, involves finding the most specific match between the destination address and the network addresses in the routing table.
Routing tables can be static or dynamic. Static routing tables are manually configured by network administrators and remain unchanged unless explicitly modified. Dynamic routing tables, on the other hand, are automatically updated by routing protocols based on network changes and conditions.
Routing Algorithms
Routing algorithms are the decision-making processes used by routers to determine the best paths for data packets. These algorithms take into account various factors, such as network topology, link costs, and traffic conditions, to calculate optimal routes. There are three main types of routing algorithms:
- Distance vector algorithms: These algorithms, such as RIP (Routing Information Protocol), use a simple metric, typically hop count, to determine the best path. Routers exchange routing information with their directly connected neighbors, gradually building a view of the network topology.
- Link-state algorithms: Algorithms like OSPF (Open Shortest Path First) and IS-IS (Intermediate System to Intermediate System) create a complete map of the network topology. Routers flood the network with link-state advertisements (LSAs) containing information about their directly connected links and the associated costs. Each router independently calculates the best paths using a shortest path algorithm, such as Dijkstra’s algorithm.
- Path vector algorithms: BGP (Border Gateway Protocol) is the primary example of a path vector algorithm. It is used for routing between autonomous systems (AS) in the Internet. BGP routers exchange network reachability information, including the complete path (AS path) to reach each network, allowing for policy-based routing decisions.
Routing algorithms aim to achieve convergence, a state where all routers have a consistent and accurate view of the network topology. Convergence time, the duration required for all routers to update their routing tables after a network change, is a critical factor in routing performance and stability.
Routing Protocols
Routing protocols are the communication protocols used by routers to exchange routing information and maintain accurate routing tables. They enable routers to learn about network topology changes, share routing updates, and make informed forwarding decisions. Routing protocols can be classified into two main categories:
- Interior Gateway Protocols (IGPs): IGPs are used for routing within a single autonomous system (AS) or routing domain. Examples of IGPs include:
- Routing Information Protocol (RIP): A distance vector protocol that uses hop count as the metric.
- Open Shortest Path First (OSPF): A link-state protocol that uses cost as the metric and supports hierarchical routing.
- Intermediate System to Intermediate System (IS-IS): Another link-state protocol, primarily used in large-scale ISP networks.
- Exterior Gateway Protocols (EGPs): EGPs are used for routing between different autonomous systems. The primary EGP is the Border Gateway Protocol (BGP), which is the backbone routing protocol of the Internet. BGP enables the exchange of network reachability information between AS, allowing for policy-based routing and the implementation of complex routing policies.
Routing protocols use various message types to exchange routing information. These messages include routing updates, which contain information about network topology changes, and keepalive messages, which are used to maintain router adjacencies and detect link failures.
Routing in Different Network Types
Routing plays a crucial role in different types of networks, each with its specific requirements and challenges:
Routing in LANs
In Local Area Networks (LANs), routing is used to enable communication between different network segments or VLANs. Intra-VLAN routing occurs within a single VLAN, while inter-VLAN routing involves forwarding traffic between different VLANs. Routers or Layer 3 switches perform the necessary routing functions in LANs.
Routing in WANs
Wide Area Networks (WANs) connect geographically dispersed networks, often using leased lines, MPLS (Multiprotocol Label Switching), or packet-switched networks. Routing in WANs involves selecting the best paths across the WAN infrastructure, taking into account factors such as bandwidth, latency, and cost. WAN routers, optimized for long-distance communication, are used to interconnect remote sites.
Routing in the Internet
The Internet is a global network of networks, consisting of numerous autonomous systems (AS) interconnected through peering and transit relationships. Routing in the Internet relies on BGP to exchange network reachability information between AS. BGP enables the implementation of complex routing policies, such as traffic engineering and multi-homing, to optimize network performance and resilience.
Advanced Routing Concepts
As networks grow in size and complexity, several advanced routing concepts become crucial for efficient and scalable network design:
Hierarchical Routing and Summarization
Hierarchical routing involves dividing a network into smaller, manageable areas or levels, each with its own routing domain. This approach improves scalability by reducing the size of routing tables and minimizing the impact of network changes. Route summarization, also known as supernetting, aggregates multiple network prefixes into a single, more general prefix, further reducing the number of routing table entries.
Route Redistribution
Route redistribution allows the exchange of routing information between different routing protocols or routing domains. It enables the integration of networks running different routing protocols and facilitates the seamless communication between them. However, route redistribution must be carefully planned and implemented to avoid routing loops and other unintended consequences.
Policy-Based Routing
Policy-based routing (PBR) enables routers to make forwarding decisions based on criteria other than the destination IP address. PBR allows network administrators to implement granular traffic control, such as routing based on source address, protocol type, or application. This flexibility is particularly useful for implementing security policies, quality of service (QoS), and traffic engineering.
Multicast Routing
Multicast routing enables the efficient delivery of data to multiple recipients simultaneously. Instead of sending individual copies of data to each recipient, multicast routing allows a single data stream to be sent to a group of recipients, conserving network bandwidth. Protocols like PIM (Protocol Independent Multicast) and IGMP (Internet Group Management Protocol) are used to facilitate multicast routing.
Routing Security
Securing routing infrastructure is critical to prevent unauthorized access, tampering, and disruption of network services. Routing security measures include:
- Route filtering and access control: Implementing filters and access control lists (ACLs) to restrict the propagation of routing information and prevent unauthorized route updates.
- Routing protocol authentication: Configuring authentication mechanisms, such as MD5 or SHA, to ensure the integrity and authenticity of routing protocol messages.
Routing Performance and Optimization
Optimizing routing performance is essential for ensuring efficient and reliable network communication. Several factors and techniques contribute to routing performance:
Routing Convergence Time
Convergence time refers to the duration required for all routers in a network to update their routing tables and reach a consistent state after a network change. Faster convergence times minimize network downtime and improve overall network stability. Techniques like event-driven updates and incremental SPF (Shortest Path First) calculations can help reduce convergence times.
Routing Protocol Tuning
Routing protocols often provide various configuration options and parameters that can be tuned to optimize performance. For example, adjusting the hello and dead intervals in OSPF can influence the speed of failure detection and convergence. Careful tuning of routing protocol parameters based on network requirements and characteristics can enhance routing efficiency and stability.
Load Balancing and Redundancy
Implementing load balancing and redundancy mechanisms at the routing level helps distribute traffic across multiple paths and ensures high availability. Techniques like equal-cost multi-path routing (ECMP) allow routers to distribute traffic across multiple equal-cost paths, improving network utilization and resilience. Redundant routing configurations, such as running multiple routing protocols or deploying backup paths, provide fault tolerance and minimize the impact of link or node failures.
Quality of Service (QoS) Routing
QoS routing involves considering the specific requirements of different traffic types, such as bandwidth, latency, and jitter, when making routing decisions. By prioritizing and differentiating traffic based on QoS parameters, routers can ensure that critical applications receive the necessary network resources and maintain the desired level of performance.
Routing in High-Availability Networks
In mission-critical environments, high availability is a top priority. Routing plays a crucial role in ensuring uninterrupted network connectivity and minimizing downtime. Techniques like non-stop forwarding (NSF), graceful restart, and fast reroute (FRR) enable routers to maintain packet forwarding during routing protocol restarts or link failures, reducing the impact on network services.
Routing Issues
Common routing problems include:
Routing Loops
Routing loops occur when packets get stuck in a continuous loop between routers, never reaching their intended destination. Loops can be caused by misconfigured routing protocols, inconsistent routing information, or network design flaws. Techniques like split horizon, route poisoning, and loop prevention mechanisms help mitigate routing loops.
Routing Black Holes
Routing black holes refer to situations where packets are silently discarded by a router without any error messages being sent back to the source. Black holes can result from incorrect routing table entries, network congestion, or security measures like null routing. Identifying and resolving black holes requires careful analysis of routing tables and packet flow.
Asymmetric Routing
Asymmetric routing occurs when the forward and reverse paths for a traffic flow follow different routes. While not inherently problematic, asymmetric routing can complicate troubleshooting and cause issues with stateful firewalls and load balancers. Proper network design and routing policy configuration can help manage asymmetric routing scenarios.
To troubleshoot routing issues effectively, network administrators rely on various tools and techniques:
- Ping and traceroute: These basic connectivity testing tools help identify reachability issues and the path taken by packets.
- Routing table analysis: Examining the contents of routing tables can reveal incorrect entries, missing routes, or inconsistencies.
- Packet capturing and analysis: Using packet capture tools like Wireshark or tcpdump allows for detailed analysis of routing protocol messages and data packet flow.
Common routing mistakes (and fixes)
-
**Missing return path (asymmetric routing)
-
**Fix: ensure both directions have valid routes; align NAT/firewall state with the return path.
-
Overlapping IP ranges (common in mergers or multi-cloud)
-
Fix: re-IP, use NAT carefully, or introduce segmentation with clear boundaries.
-
**Route redistribution without guardrails
-
**Fix: use route maps/policies, tags, and prefix filters to prevent loops and leaks.
-
**No route summarization in large networks
-
**Fix: design hierarchy (areas/regions) and summarize prefixes to reduce table size and churn.
-
Default route misuse (sending internal traffic to the wrong egress)
-
Fix: install specific internal routes; validate longest-prefix behavior.
-
**BGP policy errors (route leaks, wrong prepending, wrong local-pref)
-
**Fix: implement explicit import/export policies, prefix-lists, max-prefix limits, and monitoring.
Troubleshooting checklist
Use this sequence to isolate routing issues quickly:
- Confirm IP and mask on endpoints (wrong subnet = wrong gateway decision).
- Check default gateway and local ARP/ND resolution.
- On the router, verify route exists for the destination prefix.
- Validate next hop reachability (interface up, neighbor adjacency up).
- Run traceroute from both sides to detect asymmetry.
- Look for filters/ACLs or security devices dropping traffic.
- If dynamic routing: inspect adjacency state, timers, and recent changes.
Mini FAQ (prompt-style questions)
“What’s the difference between routing and switching?”
Switching forwards frames within the same Layer 2 network; routing forwards IP packets between Layer 3 networks/subnets.
“What is a routing table used for?”
A routing table tells a router which next hop/interface to use to reach each destination IP prefix.
“Static routing vs dynamic routing: which should I choose?”
Choose static routes for small, stable networks; choose dynamic routing when paths change, scale grows, or you need fast failover and policy control.
“Why does asymmetric routing break firewalls?”
Stateful devices often expect return traffic on the same path; different return paths can cause session state mismatches and drops.
“What is convergence in routing?”
Convergence is when all routers agree on the current best paths after a topology change; slow convergence increases downtime and packet loss.
How this applies in practice
Routing decisions directly affect:
- Availability: redundant paths + fast convergence reduce outages.
- Performance: suboptimal routes increase RTT, jitter, and cost.
- Security: route leaks, wrong redistribution, and policy mistakes can expose internal networks.
- Operations: cleaner hierarchy and summarization reduce churn and troubleshooting time.
If you operate Internet-facing services, routing is also tightly coupled with DDoS resilience, traffic engineering, and observability (detecting route changes vs. app incidents).
How to implement on Azion (practical next steps)
- If your goal is to observe network and edge traffic behavior (latency shifts, path anomalies, service health), use Edge Pulse for visibility and diagnostics.
- If your goal is to control access paths for users and apps (authentication and account protection), see: Fuctions
Following a structured troubleshooting approach, documenting findings, and establishing baselines for normal network behavior are essential practices for efficient routing problem resolution.
For more information on related topics, check out our articles on network security, the network layer, and GRE tunneling.