Solution Detail
Flow Hunter Network Traffic Analysis + BestPerf Network Quality Measurement
For ongoing dedicated-link operations: identify which links, directions, periods and applications consume bandwidth, how long capacity will last, and whether to expand or reschedule traffic first.
Main Test and Analysis Tools
Products
2
Scenarios
12
Evidence
PCAP
Why
Many link issues involve sustained peak-hour utilization, brief bursts, backups competing with daytime work, abnormal cross-cloud synchronization or individual hosts creating congestion. Average interface utilization can hide minute-scale bursts, directional imbalance, five-minute peaks, busy-hour 95th percentiles and actual traffic sources.
Bandwidth monitoring must also align ingress and egress at CE/PE handoffs, enterprise and carrier billing methods, cloud/data center directional bottlenecks, QoS treatment and backup schedules. Correlate bandwidth curves, sessions, application classes and active capacity measurements to answer these questions.
Flow Hunter continuously observes real traffic and identifies top applications, hosts, sessions, protocols, VLANs, subnets, bursts and causes. BestPerf schedules available-capacity, TCP goodput and load-quality tests. Together they explain high utilization by direction, time, application, capacity threshold and expansion trigger.
Topology
How
Inventory MPLS VPN, internet, cloud, data center, branch and cross-region circuits. Record CE/PE handoffs, committed and burst bandwidth, CIR/PIR and provider billing methods. Define ingress, egress and inter-site directions consistently to prevent enterprise/provider utilization disputes.
Mirror both directions at CE/PE handoffs, cloud entrances, core exits or data center links into Flow Hunter. Track averages, five-minute peaks, busy-hour 95th percentiles, daily/weekly/monthly peaks and trends. Mark thresholds such as 70%, 80% and 90% to distinguish sustained shortages, bursts, directional congestion and scheduled workloads.
Break bandwidth curves down by application, host, session, protocol, port, VLAN, subnet and QoS queue. For each congestion window, identify conferencing, object storage, database synchronization, distribution, backups, log uploads, artifact repositories, SaaS or scanning traffic and its direction.
Align bursts with the business calendar. Check backups, cross-cloud synchronization, warehouse extraction, log uploads, registry synchronization, video uploads and batch jobs for production interference. Consider rate limits, rescheduling, queue changes, routing or concurrency controls; assess expansion for sustained demand that cannot be managed.
Schedule measurements at both endpoints during quiet periods, peaks, changes and suspected bottlenecks. Measure TCP goodput, directional differences, quality under load and available capacity. If throughput drops before interfaces saturate, investigate provider limits, cloud policies, queues, endpoint bottlenecks, TCP windows or link quality.
Combine busy-hour 95th percentiles, five-minute peaks, congestion counts and durations, top applications, QoS use, active tests and historical trends. Report current thresholds, expected expansion timing, backup/QoS adjustments, billing reviews and before/after expansion results.
Product Stack
Tool Combination
Flow Hunter
Capture passively through mirrored ports or TAPs for L2-L7 analysis, session investigation, protocol-stage diagnosis and PCAP export.
BestPerf
Actively measure RTT, directional loss, jitter, micro-outages, TCP goodput and quality under load to establish link baselines.
Result
See both directions, five-minute peaks, busy-hour 95th percentiles, bursts, top applications, hosts, sessions and QoS queues rather than average interface utilization alone.
Distinguish sustained growth from a few poorly scheduled jobs before buying bandwidth. Trends, utilization thresholds and expansion triggers support budgets, procurement and supplier discussions.
Compare bursts and throughput drops with real traffic, active tests and billing methods to distinguish provider limits, cloud policies, enterprise congestion, queue settings, endpoint limitations and application issues.
BestPerf distinguishes low utilization, degraded link quality, endpoint bottlenecks and reduced TCP goodput, avoiding the assumption that every slowdown needs more bandwidth.
Deliver more than throughput or delay figures: include topology, parameters, timelines, quality curves, key sessions, anomalies, PCAP extracts and retest results. Development, testing, operations, vendors and customers can discuss the same evidence rather than relying on screenshots, verbal descriptions and repeated debate.
References
Next Step
NetAccura can create a test topology, parameter table and POC plan based on your links, models, application protocols, available PCAPs and SLA goals.