January 2018 · Reading time: 11 mins
Introduction
In today’s distributed enterprise landscape, application performance directly affects user productivity and business outcomes. As organizations transitioned from MPLS-centric architectures to SD-WAN in 2018, ensuring visibility into application performance became a critical IT requirement. Traditional WAN tools proved insufficient in a world of SaaS, cloud-native workloads, and hybrid architectures.
This post explores the technologies, metrics, and best practices that shaped application performance visibility in SD-WAN environments at the time. We'll cover how SD-WAN enables deeper visibility, which metrics matter, the role of real-time analytics, and how IT teams leveraged telemetry to deliver a superior user experience.
The Limitations of Traditional Monitoring
Legacy WAN monitoring tools were not designed for today’s traffic patterns. Most focused on link-level statistics (e.g., interface utilization, packet drops) or basic reachability (e.g., ping, traceroute). Tools like SNMP or NetFlow offered a partial view but lacked application-layer context.
Furthermore, many traditional approaches required manual configuration to track application flows. As applications increasingly moved to the cloud or adopted microservices architectures, such tools failed to keep up. Visibility gaps widened, leading to poor root cause analysis and finger-pointing between network and application teams.
SD-WAN as an Enabler of Visibility
One of SD-WAN’s most impactful features is its ability to inspect and classify application traffic at the edge. Unlike traditional routers, SD-WAN edge devices include deep packet inspection (DPI) engines and can detect thousands of applications out-of-the-box.
This visibility allows IT teams to understand not just where traffic is going, but what applications are consuming bandwidth, how they’re performing, and what transport paths they’re using. SD-WAN controllers aggregate this telemetry, offering centralized dashboards and reports.
- Application discovery without manual configuration
- Real-time traffic classification and statistics
- Path performance monitoring (latency, jitter, loss)
- Dynamic policy enforcement based on application behavior
Why Application Metrics Matter
Users don’t complain about “latency on the WAN”—they complain that Salesforce is slow, or that Teams meetings are choppy. Network teams must therefore align performance monitoring to application behavior.
Commonly tracked metrics in 2018 included:
- Latency: Round-trip delay, particularly for voice/video and transaction-heavy apps
- Jitter: Variation in packet arrival, impacting VoIP and streaming
- Packet Loss: Even small loss percentages can break real-time traffic
- MOS (Mean Opinion Score): Voice quality indicator
- Application Response Time: Measured at the edge, including DNS lookup and TCP handshakes
Real-Time Dashboards and Predictive Analytics
Leading SD-WAN vendors in 2018—Cisco Viptela, VMware VeloCloud, Silver Peak—offered real-time analytics dashboards. These provided visual insight into application performance trends, traffic spikes, anomalies, and site-to-site comparisons.
Machine learning also began to play a role. Anomaly detection algorithms identified outlier traffic patterns or deviations from baselines. Alerts triggered automated actions such as path switching or user notification.
Key features included:
- Per-application performance graphs and baselines
- Automatic detection of degraded links or applications
- Heatmaps and topology maps with drill-down
- Multi-tenant views for MSPs or large enterprise IT
Integration with Helpdesk and ITSM Tools
Another key trend was integrating SD-WAN visibility with IT operations. NOC teams leveraged webhook integrations to pass real-time alerts to systems like ServiceNow, PagerDuty, or Slack. Some platforms exposed APIs to fetch telemetry data for deeper analytics or dashboard consolidation.
This level of automation and integration significantly reduced mean-time-to-resolution (MTTR) and empowered L1/L2 support staff to triage WAN issues without escalation.
Challenges and Considerations
Despite the advantages, SD-WAN visibility wasn’t plug-and-play. Organizations needed to ensure:
- Edge devices had adequate resources (CPU, memory) to process telemetry
- Central controllers scaled to ingest and analyze traffic from all sites
- Security and privacy concerns were addressed when inspecting application payloads
- Change management accounted for policy tweaks affecting routing decisions
Additionally, visibility was only as good as the underlying classification engine. False positives or unrecognized applications reduced trust in the analytics platform.
Case Study: Retail Chain with 200+ Sites
One global retail chain implemented SD-WAN in late 2017. With centralized dashboards, they identified that 35% of WAN usage came from background Windows Update traffic during business hours. By shaping and rescheduling these flows, they reduced link saturation and improved POS application reliability.
Further, they detected that specific branches suffered recurring latency issues due to overloaded LTE backups. Visibility into real-time link health enabled proactive failover to fiber circuits.
Conclusion
In 2018, SD-WAN transformed how enterprises approached WAN monitoring and performance management. Visibility was no longer just about link status—it became a business-critical requirement tied to application outcomes and user experience.
Organizations embracing SD-WAN must invest in tools and practices that surface actionable insights. Doing so enables faster troubleshooting, smarter policy enforcement, and ultimately a more resilient and responsive enterprise network.