TeleDynamics Think Tank

Minimize jitter, latency and other UC issues through proper routing

Written by Daniel Noworatzky | Jan 15, 2025 3:17:00 PM

Routing is a vital part of network operations, and employing it correctly is of utmost importance for any enterprise network.

In the context of networking, routing is the process by which data packets are directed to their final destination as they travel from node to node.  There are many aspects of routing design that can affect general network traffic patterns and particular network services. UC and VoIP services can be adversely affected if routing design does not take into account their special characteristics.

This article examines various aspects of routing implementation to see how it can potentially affect real-time services. We also look at what specific design aspects you should consider to ensure the proper operation of such services.

What is routing?

A data network consists of network segments separated by routers or network nodes. These nodes are programmed to receive each data packet, read the destination IP address in its header, and decide which port from which to forward the packet towards its destination. This process is called routing.

Routing occurs at each network node encountered along the path of a packet from its source to its destination. Each router or network node makes the routing decision individually until the packet reaches its destination.

How routes are learned

Routers learn about the appropriate exit interfaces for particular destinations either by administrators configuring static routing rules within each router or setting up a dynamic routing protocol.

Statically configured routing rules benefit small networks where efficiency and simplicity outweigh the need for scalability and redundancy. For larger networks, routing protocols are preferable.

When protocols are implemented, routers advertise their routes to each other automatically. Then, they determine the best exit interface for each destination using specific algorithms.

If, for whatever reason, a router or a link fails, dynamic routing protocols can automatically determine alternative paths to the intended destination, resulting in a “self-healing” network infrastructure.

Routing design concepts

Various routing design concepts can be employed in a network that are beneficial for overall network operation, ensuring bandwidth usage efficiency and redundancy.

There are also some routing phenomena that one must be aware of that are associated with certain behaviors; here are a couple of the most significant for VoIP and UC.

Equal-cost multi-path (ECMP) routing

This technique allows you to configure the routing to balance traffic load across two or more paths at an equal cost. This provides both network bandwidth use efficiency and redundancy in the event of a link or node failure.

Using ECMP routing provides more available throughput between source and destination as well as a more reliable network. This arrangement is typically desirable, especially for networks that serve mission-critical and high-bandwidth services and applications.

Asymmetric routing

The fact that a packet is able to reach its destination doesn’t guarantee that a return path is available; if it is, it doesn’t mean that the return path will be the same as the forward path. When the return path differs from the forward path, this is called asymmetric routing.

The diagram below shows how the green forward path differs from the blue return path.

Asymmetric routing can occur for several reasons, including misconfigured routing settings, atypical network metrics, or intentional design choices.

It may arise when routing tables are not properly synchronized, cost metrics favor different paths for inbound and outbound traffic, or load balancing techniques distribute traffic across multiple routes.

VoIP and UC requirements

Because they work in real-time, VoIP and UC applications have traffic routing requirements that are not shared by more traditional data types.

Network phenomena that affect these time-sensitive data streams include jitter, latency, packet loss, and out-of-order packet arrivals.

Network planners must be aware that routing operation and design can impact these aspects of data transmission, thus directly affecting the quality of VoIP and UC applications on enterprise networks.

Effects of routing on VoIP and UC

Although ECMP delivers important benefits to network operations, the multiple paths available to traffic may increase jitter and out-of-order packet arrival. The differing latencies that the multiple paths may introduce can cause unpredictable packet arrival times, resulting in degradation in quality and, in extreme cases, even dropped calls.

Similarly, asymmetric routing may also cause problems with real-time traffic. Asymmetric routing may not impact more traditional data traffic to as great a degree, but it will affect real-time services.

When the path for sending packets is different from the path for receiving packets, this can cause increased jitter, latency, and packet reordering, all of which degrade the quality of VoIP and UC.

For example, a packet might take a low-latency direct route to the destination but return via a high-latency backup link.

This can be disruptive in a voice or video call where interactions between remote participants could become highly out of sync.

ECMP and asymmetric routing are among the most prominent routing-related mechanisms that can affect time-sensitive communications. However, additional routing and related network operations that can and do affect UC and VoIP services include the following:

  • Policy-based routing (PBR): This routing technique makes routing decisions based on more parameters than just the lowest cost to the destination. If routing decisions are made based on policies (e.g., application type or source IP) rather than just the optimal path, VoIP and UC packets may take suboptimal routes, increasing latency and jitter. When applying PBR, you must ensure that policies prioritize time-sensitive traffic.

  • Dynamic routing protocol stability: The “self-healing” characteristic of dynamic routing protocols can cause problems with network stability if misconfigurations or network instability result in too frequent route changes or “route flapping.” Such a situation can lead to temporary outages, increased latency, or jitter, which are all problematic with real-time communications.

    Although such phenomena
    are harmful to all types of data, VoIP and UC are much more susceptible to the detrimental effects of this instability and are often the services that are most affected. You must implement techniques such as route dampening and proper stability mechanisms to deal with such situations.

  • Traffic engineering (TE): This is a technique used with many routing protocols that aims to optimize traffic patterns in extensively large networks based not only on lowest cost paths but also on a multitude of constraints. These constraints allow for more granular control over traffic patterns to ensure the absolute best and most efficient usage of network resources.

    However, if TE constraints are too restrictive, VoIP and UC traffic may be forced onto suboptimal paths, resulting in degraded service quality. Appropriately balancing TE objectives with real-time traffic needs will ensure sufficient bandwidth and low-latency paths.

Routing-aware design

The abovementioned issues are just some of the ways that routing can affect network services of all types, especially time-sensitive services such as UC and VoIP.

When designing enterprise networks and deploying these real-time services, it is important to be aware of the underlying routing design that will serve them. Ensuring a well-thought-out, “routing-aware” approach to the deployment of VoIP and UC services is essential.

Conclusion

Routing plays an important role in the performance and reliability of enterprise networks, particularly for time-sensitive services like VoIP and UC. Proper routing design and implementation must be deployed, taking into consideration the network services that will be delivered.

This will minimize issues such as jitter, latency, and packet reordering, which can significantly degrade real-time communication quality. By understanding the unique requirements of these services and carefully managing routing mechanisms, network planners can create robust and efficient networks that support both traditional and real-time applications effectively.

You may also like: 

Layer 2 or 3? Choose the right switch for optimal network performance

AI-driven network infrastructure: the future of UC systems

Prevent network failures with effective power management