Preparing Your IP Network for Voice and Video Traffic: Bandwidth, QoS, Wi-Fi, and Network Performance

Voice and video conferencing have become essential components of modern business communications. Whether an organization uses Microsoft Teams, Zoom, Cisco collaboration systems, Polycom endpoints, or traditional Voice over IP (VoIP) telephones, the quality of these services depends heavily on the underlying network.

A network that performs well for everyday applications such as email, web browsing, and file transfers may still struggle to support real-time voice and video communications.

Unlike conventional data traffic, voice and video are particularly sensitive to network delays, inconsistent packet delivery, and congestion. Even brief disruptions can result in distorted audio, frozen video, dropped calls, or synchronization problems.

Preparing an IP network for voice and video requires more than simply purchasing additional bandwidth. Network administrators must consider traffic prioritization, bandwidth requirements, wireless performance, Internet connectivity, and continuous monitoring.

This article explains the major considerations involved in designing and maintaining a network capable of reliably supporting real-time communications.

1. Understanding How Network Performance Affects Voice and Video

Before designing a network for real-time communications, it is important to understand the factors that influence call quality.

Packet Loss

Packet loss occurs when network packets fail to reach their intended destination.

Traditional applications can often recover from packet loss by retransmitting missing information. Real-time communications are less forgiving because retransmitted information may arrive too late to be useful.

Packet loss can cause:

• Missing or distorted audio
• Robotic or broken speech
• Pixelated or frozen video
• Reduced picture quality
• Call interruptions

Some voice and video systems use packet-loss concealment techniques to reduce the impact, but these techniques cannot compensate for unlimited packet loss.

Latency

Latency is the amount of time it takes for data to travel between two points on a network.

High latency creates noticeable delays during conversations. Participants may unintentionally interrupt one another because the audio they hear is arriving late.

Latency may be introduced by routing decisions, congested network equipment, geographic distance, or additional processing along the network path.

Jitter

Jitter refers to variation in packet arrival times.

For example, a stream of audio packets might normally arrive at intervals of 20 milliseconds, followed by packets arriving at irregular intervals because of network congestion.

Voice and video systems use jitter buffers to smooth out these variations. However, excessive jitter can cause audio interruptions or require additional buffering that increases conversational delay.

Bandwidth Limitations

Bandwidth represents the amount of data a network connection can carry over a given period.

When network traffic exceeds the available capacity, packets may be delayed or discarded.

For example, if an employee begins transferring a large file over an already congested WAN connection, the additional traffic may interfere with an ongoing video conference.

The important distinction is that having a fast network connection does not automatically guarantee good voice and video quality.

The network must also transport packets consistently and prioritize time-sensitive traffic where congestion can occur.

2. Understanding the Difference Between Data Traffic and Real-Time Traffic

Most conventional network applications generate traffic in bursts.

For example, opening a website may cause the browser to download a large amount of information briefly, followed by a period of relatively little activity.

File downloads, backups, and software updates can produce similar bursts.

Voice and video communications behave differently.

A voice conversation generally produces a continuous stream of relatively small packets, while video conferencing produces an ongoing stream whose bandwidth changes according to the video content, resolution, frame rate, and compression.

When these applications share a congested connection, sudden bursts of ordinary data traffic can interfere with voice and video packets.

This is where Quality of Service becomes important.

Quality of Service (QoS) allows network equipment to identify different types of traffic and provide preferential treatment to time-sensitive applications.

For example, when a network interface becomes congested, properly configured QoS can allow voice packets to be transmitted before lower-priority file-transfer packets.

QoS does not create additional bandwidth. Instead, it controls how available network capacity is shared when competing applications need access to it.

3. Planning Bandwidth for Voice and Video Communications

One of the most important steps in preparing a network is determining how much bandwidth voice and video applications require.

Bandwidth planning should account for more than the number of devices installed. Administrators need to understand how those devices will actually be used.

Factors to Consider

Several factors influence voice and video bandwidth requirements:

• Number of endpoints: How many phones, conferencing systems, and desktop clients are deployed?
• Concurrent usage: How many devices are expected to participate in calls simultaneously?
• Call resolution: Will video systems operate at standard definition, 720p, 1080p, or higher resolutions?
• Frame rate: Will video operate at 15, 30, or 60 frames per second?
• Codec selection: Which audio and video compression technologies are being used?
• Call routing: Does traffic travel directly between endpoints, through a conferencing bridge, or through a cloud service?
• Shared applications: What other network applications compete for the same bandwidth?
• Network overhead: How much additional capacity is needed for transport overhead and operating headroom?

Bandwidth Requirements Vary by Application

Different communication systems have different bandwidth requirements.

The original Polycom training material included the following illustrative figures:

These figures reflect particular older Polycom technologies and should not be treated as universal requirements for modern conferencing systems.

Modern bandwidth requirements depend on the endpoint, codec, meeting layout, supported features, and service provider.

Always consult the current specifications for the system being deployed.

Example: Calculating Bandwidth Requirements

Consider an office with eight video conferencing endpoints and twelve VoIP telephone channels.

For this example, assume:

• Eight video conferences operating simultaneously
• Each video conference requiring 1.5 Mbps
• Twelve simultaneous G.729 voice channels
• Each G.729 channel using an estimated 35 Kbps, as illustrated in the original presentation

The estimated video bandwidth would be:

8 × 1.5 Mbps = 12 Mbps

If a 20% planning allowance is applied to the video traffic:

12 Mbps × 1.20 = 14.4 Mbps

Voice bandwidth would be:

12 × 35 Kbps = 420 Kbps

The combined estimated demand would therefore be:

14.4 Mbps + 0.42 Mbps = 14.82 Mbps

This is an illustrative calculation, not a recommended universal bandwidth requirement. Actual capacity must be calculated for the relevant network links and traffic directions, with the correct codec rates and overhead.

Additional capacity would also be needed for other network applications.

Importantly, bandwidth requirements should be evaluated at every potentially congested connection, not simply against the organization’s advertised Internet speed.

Understanding Concurrency

Concurrency is the number of devices or sessions using the network at the same time.

For example, an organization may have 50 conferencing endpoints but rarely use more than 10 simultaneously.

Designing around realistic concurrency can help avoid unnecessary costs.

However, administrators should also account for unusual periods of heavy usage, such as company-wide meetings, training sessions, and peak business hours.

4. Quality of Service: Prioritizing Voice and Video Traffic

Quality of Service is one of the most important components of a reliable real-time communications network.

The original Polycom presentation identifies five major components of a comprehensive QoS design:

1. Traffic classification
2. Trust boundaries
3. WAN QoS
4. LAN QoS
5. Wireless QoS

Each component plays a different role in ensuring that voice and video packets receive appropriate treatment throughout the network.

Traffic Classification

Traffic classification is the process of identifying packets and assigning them to appropriate service categories.

A video conferencing system may generate several different kinds of traffic:

Media Traffic: The actual audio and video streams transmitted during a conference.

Signaling Traffic: Messages used to register devices, establish calls, control sessions, and terminate connections.

Management Traffic: Communications used to monitor, configure, and administer conferencing equipment.

These traffic types should not automatically receive identical priority.

For example, a management web interface generally does not require the same network priority as a real-time audio stream.

Differentiated Services Code Point (DSCP)

DSCP is a value carried in the IP packet header that allows network equipment to classify traffic for differentiated handling.

The presentation provides the following example classifications:

These markings represent an example QoS design rather than a mandatory classification for every environment.

Actual DSCP policies should align with the organization’s communication platform, network equipment, and service-provider requirements.

Why Classification Matters

Consider a conferencing endpoint that sends video media, signaling messages, and administrative traffic.

Assigning the same high-priority classification to every packet originating from that device would incorrectly prioritize traffic that does not require real-time treatment.

A properly designed policy distinguishes between those different traffic types.

This helps preserve high-priority network capacity for applications that genuinely require it.

5. Understanding QoS Trust Boundaries

A trust boundary determines whether network equipment accepts the priority markings assigned to incoming packets.

For example, a properly configured conferencing device may mark its outgoing video packets as AF41.

The switch receiving those packets must determine whether to preserve the marking or replace it according to local policy.

A network should not blindly trust every packet’s classification.

Otherwise, an incorrectly configured or unauthorized device could mark ordinary traffic as high priority.

Common approaches include:

• Trusting classifications from known, managed communication devices
• Applying trust policies to designated network interfaces or VLANs
• Examining traffic types and assigning classifications according to defined rules
• Rewriting incorrect or untrusted markings at the network boundary

The original training material recommends marking traffic at the endpoint when possible and establishing a clear trust boundary.

However, marking packets and giving them priority are two separate operations.

Even when DSCP markings are correct, switches and routers must still be configured with the appropriate queuing and scheduling policies.

6. Configuring QoS on the LAN

Organizations sometimes focus heavily on WAN performance while overlooking congestion inside the local network.

Although LAN connections are commonly faster than WAN connections, local network congestion can still affect real-time communications.

Two situations deserve particular attention.

Downlink Congestion

Downlink congestion can occur when traffic moves from a high-speed network connection onto a slower interface.

For example, a switch receives traffic through a 1 Gbps uplink but must forward it through a 100 Mbps access port.

If incoming traffic temporarily exceeds the access port’s available capacity, packets can accumulate in the output queue.

Without appropriate QoS, voice and video packets may experience delays or packet loss.

Uplink Congestion

Uplink congestion occurs when multiple devices collectively attempt to send more traffic than an uplink can accommodate.

For example, several workstations may simultaneously initiate large file transfers while conferencing endpoints are using the same switch.

Even when individual access ports are operating below their limits, the shared uplink may become congested.

This condition is often associated with network oversubscription.

Oversubscription occurs when the combined potential traffic from connected devices exceeds the capacity of a shared connection.

Properly configured LAN QoS can help protect real-time communications during temporary congestion.

However, persistent congestion should also be addressed through capacity planning and network design.

7. Preparing Wireless Networks for Voice and Video

Wireless networking introduces additional considerations because connected devices share radio-frequency capacity.

A Wi-Fi network that performs adequately for general browsing may experience problems when supporting numerous simultaneous voice or video calls.

The Polycom presentation identifies five important considerations for wireless voice and video deployments.

Signal Strength

Reliable wireless communications require adequate signal quality in both directions.

It is not enough for an access point to transmit a strong signal to a device.

The device must also be able to transmit reliably back to the access point.

This is particularly important for battery-powered phones and mobile devices, which may have different transmission capabilities.

Poor bidirectional connectivity can contribute to dropped calls and inconsistent audio.

Wireless Interference

Wireless signals can be affected by physical obstructions and competing radio-frequency activity.

Potential sources include:

• Concrete walls
• Metal structures
• Mirrored or treated glass
• Nearby wireless networks
• Microwave ovens
• Bluetooth devices
• Other equipment operating in overlapping frequency bands

Interference can increase retransmissions, reduce usable wireless capacity, and produce inconsistent packet delivery.

Access Point Density

Providing adequate wireless coverage does not necessarily mean providing adequate wireless capacity.

A single access point may cover a large room while struggling to support numerous simultaneous video conferences.

Higher-density environments may require additional access points with carefully planned channel assignments and transmit power levels.

However, installing more access points without proper planning can introduce additional interference.

Frequency and Channel Planning

Traditional 2.4 GHz Wi-Fi deployments commonly use channels 1, 6, and 11 to reduce channel overlap.

The 5 GHz band generally provides more channel-planning flexibility and additional capacity.

The original material also discusses band steering, which encourages compatible devices to use a preferred frequency band.

Effective channel planning should account for neighboring access points, interference, device capabilities, and the operating environment.

Wireless Multimedia (WMM)

Wireless Multimedia, commonly called WMM, extends traffic prioritization to Wi-Fi networks.

WMM uses different access categories to provide preferential access to the shared wireless medium.

The categories are:

Voice and video traffic can receive preferential channel access compared with ordinary background data.

WMM does not guarantee perfect performance, but it is an important component of wireless QoS.

Wireless design should also consider access-point uplink capacity and the wired or wireless backhaul connecting access points to the rest of the network.

8. Designing WAN Connectivity for Real-Time Communications

Wide Area Networks connect geographically separated offices and services.

Voice and video traffic crossing a WAN can be affected by the capacity, latency, and congestion of the underlying network.

The original presentation emphasizes the advantages of consistent QoS policies across an enterprise WAN.

Working With Multiple WAN Providers

Organizations with international offices may use different service providers in different regions.

This can reduce costs or improve local availability.

However, multiple providers introduce additional complexity.

Potential challenges include:

• Inconsistent QoS classifications
• Different traffic-prioritization policies
• Congestion at interconnection points
• Increased operational complexity
• Difficulty identifying which provider is responsible for a network problem

For example, traffic between offices in Europe and Asia might travel through a central network location in New York.

Even when the individual regional connections have sufficient capacity, the central interconnection may become a bottleneck.

The presentation historically favors a unified MPLS-based WAN for consistency. Modern organizations may use different WAN architectures, so the appropriate design should be determined by current service requirements rather than treating MPLS as mandatory.

The underlying principle remains important: network administrators must understand the complete path that voice and video traffic follows.

9. Managing Voice and Video Traffic Over the Internet

Internet-based conferencing introduces challenges because organizations typically do not control every network device between the user and the conferencing service.

An organization’s internal QoS policies may prioritize traffic correctly, but those policies do not guarantee that an external Internet provider will apply the same treatment.

Understanding the Network Path

When troubleshooting Internet-based communications, it is important to identify:

• Where signaling traffic travels
• Where audio and video media streams travel
• Which services or conferencing infrastructure handle the connection
• Whether traffic passes through firewalls, VPNs, or other network devices
• Which network connections may become congested

Signaling and media do not necessarily use the same network path.

Understanding these differences can prevent administrators from troubleshooting the wrong connection.

Session Border Controllers

A Session Border Controller, or SBC, is a network component commonly used to manage and secure real-time communication sessions between different networks.

Depending on the deployment, an SBC may perform signaling control, media handling, security enforcement, and interoperability functions.

If media passes through an SBC, its bandwidth and session capacity must be included in network planning.

An SBC with insufficient capacity may create problems even when the surrounding network has sufficient bandwidth.

Internet Performance Considerations

Organizations relying on Internet-based collaboration should evaluate:

• Available Internet bandwidth
• Congestion during peak periods
• Latency and jitter toward conferencing services
• Packet loss along the communication path
• Service-provider routing and interconnections
• Geographic location of conferencing infrastructure

In geographically distributed organizations, the location of media services can influence network latency.

For example, routing participants through distant infrastructure may introduce unnecessary delays compared with appropriately located services.

10. Monitoring Voice and Video Network Performance

Designing a network correctly is only part of the process.

Organizations must also monitor performance continuously to identify developing problems and verify that the network continues to meet application requirements.

The original Polycom presentation emphasizes path-based network monitoring.

Device Monitoring vs. Path Monitoring

Traditional network monitoring frequently focuses on individual devices.

Administrators may check whether a switch is online, whether an interface has errors, or whether a router is experiencing high CPU utilization.

These measurements are useful, but they do not always explain why a video call performed poorly.

A communication session depends on the entire network path between its endpoints or media services.

Path-based monitoring examines the performance experienced across that route.

For example, every switch and router between two offices might appear operational, yet a congested interface or inconsistent QoS policy could still degrade video quality.

What Should Be Monitored?

Important measurements include:

Monitoring should ideally provide historical records.

Without historical data, a technician may have difficulty investigating a call that failed the previous afternoon after network conditions have returned to normal.

Continuous monitoring also helps identify congestion before users begin reporting problems.

Hop-by-Hop Troubleshooting

Path-based diagnostics can help narrow down where a problem occurs.

For example, an investigation might reveal that packet loss begins between the third and fourth routed hops along a network path.

This provides a starting point for identifying the responsible network segment.

However, administrators should interpret hop-by-hop measurements carefully. Intermediate devices may limit diagnostic traffic, and reported packet loss at one hop does not always mean that application traffic is being lost there.

Effective troubleshooting combines path measurements with interface statistics, routing information, QoS configurations, and actual call-quality data.

11. Common Voice and Video Problems and Where to Investigate

The following table provides practical starting points for troubleshooting common issues.

These are diagnostic starting points rather than definitive causes. Similar symptoms can also arise from endpoint hardware, software, media services, or application configuration.

12. A Practical Network Readiness Checklist

Before deploying or expanding a voice and video conferencing environment, administrators should confirm the following.

Bandwidth and Capacity

• Document the number and types of communication endpoints.
• Estimate simultaneous voice and video usage.
• Calculate bandwidth requirements for critical network links.
• Account for other applications sharing those connections.
• Verify actual circuit performance and available capacity.

Quality of Service

• Establish appropriate classifications for voice, video, signaling, and management traffic.
• Define trust boundaries.
• Verify that switches and routers implement the intended prioritization.
• Check whether DSCP markings are preserved or changed across relevant network segments.
• Review QoS policies with WAN service providers where applicable.

Wireless Networking

• Validate wireless signal quality in both directions.
• Evaluate interference and channel utilization.
• Confirm appropriate access-point density.
• Configure and verify WMM prioritization where supported.
• Ensure that access-point backhaul connections have sufficient capacity.

Internet and WAN Connectivity

• Document signaling and media paths.
• Identify potential network bottlenecks.
• Verify firewall and connectivity requirements.
• Assess the capacity of conferencing infrastructure and session border controllers.
• Test connectivity to important remote offices and external services.

Monitoring and Support

• Establish baseline measurements for latency, jitter, and packet loss.
• Monitor bandwidth utilization and QoS behavior.
• Retain historical performance information.
• Configure alerts for abnormal network conditions.
• Establish a troubleshooting process involving both the network and collaboration support teams.