Wednesday, August 1, 2018

QoS in Enterprise WANs: Revisiting Design Priorities in 2018

August 2018 • 8 min read

Introduction

Enterprise WANs remain under scrutiny in 2018 as application demands surge and cloud adoption reshapes traffic patterns. Quality of Service (QoS), long considered a must-have for voice and video, now needs a reexamination in the age of encrypted traffic, hybrid WANs, and SaaS.

The Changing Landscape of WAN Traffic

In 2018, the composition of WAN traffic is vastly different from a decade ago. SaaS, IaaS, and encrypted web traffic dominate link usage. This change reduces the effectiveness of traditional QoS classifications, which relied on clear-text application identifiers and port-based heuristics.

Application Awareness and Encrypted Flows

Deep Packet Inspection (DPI) tools struggle with TLS 1.3 and QUIC. Modern QoS policies must adapt using metadata, flow behavior, and integration with application APIs or traffic tagging. Without visibility, blindly trusting DSCP marks poses risks.

Policy Models: From Static to Dynamic

Static QoS policies—crafted per site or per app—fail in dynamic cloud environments. Enterprises move towards intent-based models, where application needs (latency sensitivity, bandwidth guarantees) define treatment. SD-WAN solutions enhance this with real-time telemetry and orchestration.

SD-WAN and QoS Synergy

SD-WAN platforms disrupt traditional QoS thinking. They perform per-packet steering, detect brownouts, and enforce policy centrally. QoS is no longer just queuing—it’s about routing decisions, traffic duplication, and failover logic embedded in overlays.

Last Mile Realities

QoS effectiveness remains highest on constrained links. Broadband circuits often lack enforceable SLAs, and upstream shaping is essential to prevent bufferbloat. Vendors offer CPE-based QoS enforcement, but success depends on accurate traffic classification.

Validation and Monitoring

QoS doesn’t end with policy deployment. Enterprises require telemetry—packet loss, jitter, MOS scores—to validate performance. Active testing (e.g., synthetic voice tests) and passive metrics (e.g., flow health) guide optimization.

Is QoS Still Worth It?

In some scenarios—such as high-capacity DIA circuits or Internet-only WANs—QoS may add little value. Enterprises must assess risk: What happens to business-critical traffic during congestion? If impact is minimal, complexity may not be justified.

Best Practices for 2018

  • Classify traffic using modern tools—NBAR2, flow analytics, application IDs
  • Align QoS classes with business priorities, not technical protocols
  • Use SD-WAN policy engines to simplify enforcement
  • Validate policies with real metrics—not assumptions
  • Continuously review classification accuracy

Final Thoughts

QoS is not dead—but it evolves. In 2018, it must align with application-centric networking, adapt to encrypted traffic, and integrate with SD-WAN. Enterprises should evaluate whether their current QoS models still serve their goals—or merely add complexity.


Eduardo Wnorowski is a network infrastructure consultant and Director.
With over 23 years of experience in IT and consulting, he helps organizations maintain stable and secure environments through proactive auditing, optimization, and strategic guidance.
LinkedIn Profile

AI-Augmented Network Management: Architecture Shifts in 2025

August, 2025 · 9 min read As enterprises grapple with increasingly complex network topologies and operational environments, 2025 mar...