AV over IP has advanced the audiovisual world tremendously, but behind its apparent simplicity lies a complex ecosystem in which AV technology and network engineering are inextricably intertwined. Many installations work flawlessly—until they stop. And when problems do arise, they rarely lie with encoders or decoders, but rather in the layers below: design choices, timing, configuration, cabling, or management. That is precisely why AVoIP requires a structured approach in which protocols, infrastructure, and operational discipline align perfectly. In this article, we explore this layering and show why a stable AVoIP system begins long before connecting the first AV endpoints.
AV over IP is rarely just an AV problem; it is network engineering with AV requirements on top.
AV over IP often appears simple: you connect an encoder and a decoder, audio and video travel over Ethernet, and the system does what it should. Until it doesn’t. Because when an AVoIP installation falters, the cause is rarely the AV equipment itself. In practice, the issue is usually found deeper in the chain: in network design, timing, configuration, cabling, or management. That aligns closely with broader IT outage data: the Uptime Institute’s 2025 analysis highlights the increasing complexity of modern architectures and reports that the vast majority of human‑error‑related outages stem from ignored or inadequate procedures. (Uptime Institute)
AV over IP is not a technology—it’s an ecosystem!
A common mistake is treating AV over IP as if it is a single technology. In reality, it is an umbrella term for multiple ecosystems, each with its own requirements for multicast, QoS, timing, and switching.
Audinate describes Dante as AV over IP built on common IT standards that can run on conventional switches and cabling. Crestron describes DM NVX as multicast‑driven, with IGMP snooping, an IGMP querier, and sufficient nonblocking bandwidth as hard design prerequisites. Q‑SYS requires PTP to receive the highest priority, audio directly underneath it, and video below that. Milan, meanwhile, is built on Ethernet hardware with AVB functions such as gPTP and SRP to guarantee interoperability and predictability.
This also means there is no universal checklist. A Dante network is set up differently than a SMPTE ST 2110 environment, and a multicast‑heavy DM NVX design demands different choices than a Milan/AVB installation. Anyone who does not first define the protocol stack ends up building on assumptions rather than on design decisions.
The AVoIP pyramid
A stable AVoIP installation is built from the ground up. In practice, the pyramid looks like this:
- Protocol and system requirements
- Switch platform
- Network design
- Physical layer: cabling and optics
- Configuration and validation
- Redundancy
- Monitoring, documentation, and change control
In this model, documentation is not only the finishing touch at the top, but also the cement binding all layers together.
1. First protocol choice, then hardware
The real foundation begins even before choosing a switch. You must first determine what latency is acceptable, how large the system must be able to grow, how critical failover is, and which protocol or product family will lead the design.
This seems obvious, yet many later problems originate right here: a design that works fine for standard Dante traffic is not automatically suitable for multicast video, for Q‑SYS prioritization, or for SMPTE/Milan scenarios.
The same applies in the NETGEAR world. Current AV profiles in NETGEAR’s AV Switches include Dante, Q‑SYS, AVB, SMPTE, Crestron DigitalMedia, NDI, ,IPMX and various video/audio combinations—but some profiles are model‑dependent. AVB is not available on all series, and the SMPTE ST 2110 profile applies only to specific M4350 models. This perfectly underscores the point of this layer: define requirements first, choose the platform second. (NETGEAR Knowledge Base)
2. The switch platform: the real foundation
Not every managed switch is automatically suitable for AVoIP. A datasheet listing “multicast support” or “IGMP” says very little about how predictably the switch behaves under load.
Crestron explicitly advises nonblocking Layer 3 switches for DM NVX, with sufficient bandwidth from edge to core and correctly configured uplinks. Audinate is clear that Dante can run on ordinary Ethernet switches but recommends managed gigabit switches with QoS, four queues with strict priority, and the ability to disable EEE for larger, more reliable networks. (Crestron Docs)
NETGEAR positions its AV Switches exactly at this intersection: the M4250 and M4350 series are designed with input from AV professionals, offering a dedicated AV interface and features such as Auto‑LAG, Auto‑Trunk, and IGMP Plus to simplify AVoIP deployment. The nuance is important: these features do not replace design knowledge, but they do reduce the chance of incorrect base configurations.
3. Network design: where things often truly go wrong
A weak design cannot be fully repaired later with better hardware or extra tuning. Most “it used to work” failures originate at this level.
DM NVX documentation emphasizes VLAN or MPLS segregation, correct uplink configuration, IGMP snooping, and a single querier. Q‑SYS adds requirements for multicast management and the placement of PTP, audio, and video streams in the correct queues. Crestron further advises involving network professionals early in more advanced topologies. (Crestron Docs)
NETGEAR can help here—primarily as an accelerator and standardizer. Engage and the AV UI allow profile‑driven configuration, while ProAV documentation provides examples for core/edge or spine/leaf topologies, uplinks, redundancy, and troubleshooting. This lowers the barrier for implementation but does not replace the fundamental rule: multicast, QoS, uplinks, and timing must still be designed deliberately.
And when needed, the NETGEAR ProAV Design team is always available via ProAVdesign@netgear.com. (NETGEAR Downloads)
4. The physical layer: suitability matters more than “bigger is better”
The physical layer is often underestimated, yet it is responsible for many instabilities: poorly terminated connectors, unreliable patching, marginal optics, or cabling that meets the standard on paper but not in practice.
In professional environments—such as theaters, studios, conference venues, and installations with heavy electromagnetic load from dimmers, motor controllers, power distribution, or LED drivers—a reliable physical infrastructure is essential for predictable AVoIP performance.
It is not about over‑engineering everything, but about matching the physical layer to the environment. In EMC‑heavy environments, shielded cabling such as S/FTP or SF/UTP can be a wise choice, provided it is installed and grounded correctly and consistently.
Electrical interference, crosstalk, and line contamination can manifest as jitter, audio dropouts, video flashes, PTP instability, or drifting clocks.
The same applies to fiber: choose compatible and reliable modules, check for errors, and take the physical layer seriously during validation. NETGEAR’s ProAV configuration guide explicitly includes port status, error checking, fiber module status, cable tests, and logging. (NETGEAR Downloads)
The lesson is simple: an unstable physical layer often disguises itself as a network or endpoint problem.
5. Configuration and validation: the birthplace of mysterious failures
Many AVoIP issues turn out to be configuration errors rather than hardware faults: incorrect QoS, a missing querier, a wrong trunk/access model, or “green” Ethernet features disrupting timing.
Audinate explicitly advises disabling EEE Power Savings on all Dante ports because it can cause poor synchronization and intermittent dropouts. Q‑SYS requires PTP in the highest priority queue, audio below it, and video below that. DM NVX demands IGMP snooping on all switches and at least one querier. (Dante)
This is also where NETGEAR AV provides the most tangible benefit. Engage can apply preconfigured or vendor-specific network profiles, update firmware during onboarding, show devices in the topology, run tests, and display device status. This makes configuration less artisanal and more reproducible—reducing the risk of inconsistency and environmental noise. (NETGEAR Downloads)
6. Redundancy: only valuable when truly independent
Redundancy is useful only when it is designed to be independent. A second cable in the same fault domain is not real redundancy.
Milan explicitly specifies seamless network redundancy, and NETGEAR’s ProAV guide covers LAGs and MLAGs as separate design steps for redundancy and additional bandwidth. The principle remains the same: only when paths, power, uplinks, and failover logic are truly independent does availability increase meaningfully. (Avnu Alliance)
Redundancy belongs higher in the pyramid than basic design and configuration. The network must first be stable; only then can availability be improved for environments where downtime is unacceptable.
7. Monitoring, documentation, and change control
The top layer is where a well-delivered system becomes a well‑managed system. Uptime data shows that procedures, discipline, and change control play a major role in outages. The same applies to AVoIP: a system remains stable not because it once worked, but because changes are traceable, configurations are stored, and deviations are detected quickly. (Uptime Institute)
NETGEAR Engage fits this layer perfectly. Engage allows onboarding and management of devices across sites, centralized configuration storage, firmware updates, profile templates, topology and neighbor visibility, testing, and device status monitoring. This strengthens the operational layer: fewer ad‑hoc notes, more reproducibility. (NETGEAR Downloads)
Where NETGEAR AV fits into this story
NETGEAR AV fits well into this pyramid—when positioned correctly. Not as a miracle tool that “prevents all failures forever,” but as a platform that makes AVoIP design and operations more consistent.
The combination of an AV‑focused UI, profile‑driven configuration, centralized storage, topology visibility, and protocol‑specific templates helps particularly in the layers where AVoIP projects most frequently stumble: switch selection, configuration, validation, and operations.
Conclusion
The AVoIP pyramid is not a marketing trick; it is a practical mental model. A stable AV‑over‑IP installation does not begin with the encoder or decoder, but with the layers beneath: the chosen protocol, the switch platform, the network design, the physical layer, and disciplined configuration and management.
Official documentation from Dante, DM NVX, Q‑SYS, Milan, and NETGEAR consistently shows the same pattern: failures rarely originate from “AV alone,” but from a mismatch between protocol, infrastructure, and operational execution.
No platform eliminates failures entirely. But with the right switch platform, a thoughtful design, a reliable physical layer, validated configurations, and strict change control, the chances of dropouts, jitter, and unexplained issues become demonstrably smaller. And that is where the true value of a well‑designed AVoIP network lies.
Eric Lindeman, NETGEAR ProAV Staff Systems Engineer Benelux
For more information about NETGEAR AV Switching, please contact the NETGEAR Pro AV Design Team via email: ProAVdesign@netgear.com
If you’d like to delve deeper into AV over IP switching, I invite you to check out our Online Academy via the link: https://academy.netgear.com/
On our training portal, you can find both AV and IT-related training courses. These courses are free to attend after registration, and at the end of each course, you can take an exam to earn a certificate.
