Back to projects

NETWORK TRAFFIC ANALYSIS CASE STUDY

Network Perimeter Assessment

Investigated my home network from an external cloud VPS, combining TCP and UDP scans with packet review to assess public service exposure and interpret ambiguous results.

  • StatusCase study published
  • InvestigationSeptember 28–29, 2026
  • CategoryNetwork Security
  • PublishedNovember 2, 2026

OBJECTIVE

Could an outside system reach a service?

A service listening inside a home lab is not necessarily reachable from the internet. I wanted to assess that distinction by testing my own home network from a cloud virtual private server (VPS), outside the home network, and reviewing the traffic associated with the investigation.

The objective was to identify evidence of public service exposure, understand how multiple routers, NAT, and firewall controls could affect connectivity, and document what the evidence could support.

Evidence basis: This retrospective uses the available summary of my September 28–29, 2026 investigation. Original scan logs and packet captures have not been verified as available for this write-up. Findings are reported at summary level; no reconstructed scan output is presented as original evidence.

ENVIRONMENT AND SCOPE

An external view of a layered home network

External test system

A cloud VPS running Nmap provided the external scanning vantage point. Its address is represented as [Cloud_VPS_IP].

Assessment target

My home-network public address is represented as [Home_Public_IP]. Testing concerned my own environment.

Network controls

Multiple routers, Network Address Translation (NAT), and firewall controls formed the environment. NAT translates addresses; firewall rules control permitted traffic.

Coverage

TCP ports 1–65535 and selected UDP ports were tested. Packet captures were reviewed for connection attempts and replies.

This was a perimeter-reachability assessment. Internal access, authenticated access, application vulnerabilities, and a comprehensive IPv6 assessment were not established by the available evidence.

METHODOLOGY

Combine scan results with packet observations

The steps below summarize the work performed. They do not reconstruct exact command options, timing, packet counts, or the selected UDP port list.

  1. Test from outside the home network

    Conducted scans from the cloud VPS against the home-network public address. This tested a different access path from a connection within the local network.

  2. Expand TCP coverage

    Completed a TCP scan covering ports 1–65535 to look beyond commonly used service ports. No open TCP ports were confirmed in that assessment.

  3. Assess selected UDP ports

    Tested a subset of UDP ports and encountered open|filtered results. Preserved that uncertainty rather than treating the label as confirmation of a reachable service.

  4. Review packets and the network path

    Examined captures showing outbound TCP SYN packets without corresponding observed replies. Considered the layered router, NAT, and firewall environment without assigning an unverified cause to a particular device.

OBSERVATIONS AND INTERPRETATION

Keep the finding within the evidence

These observations come from the investigation summary, with the raw-artifact limitation disclosed above. Each interpretation has a defined boundary.

Reported observations and supported conclusions
Observation Supported interpretation Limit
TCP: ports 1–65535 scanned; no open ports confirmed No externally reachable TCP service was established from that VPS during the test window. This does not prove every port was closed or that the network was free of vulnerabilities.
UDP: selected ports returned open|filtered Inconclusive. The scan could not distinguish an open, nonresponding service from filtering. Neither exposed services nor their absence were established. Untested UDP ports remain outside scope.
Capture: outbound SYN packets; no corresponding replies observed Connection attempts were visible in the examined traffic, without an observed response. This does not prove arrival at the home router or locate a blocking device.
Environment: multiple routers, NAT, and firewall controls Several network layers were relevant to troubleshooting the access path. The available evidence did not isolate which layer caused the missing replies.

A TCP SYN requests the start of a connection. In a SYN scan, a SYN/ACK response supports an open port; a reset supports a closed port. Silence alone does not establish a closed port. Filtering, routing problems, or packet loss are possible explanations, not verified causes in this investigation. Nmap's TCP SYN scan documentation explains these response patterns.

UDP does not use the TCP handshake. An open UDP service may ignore an unsuitable probe, while a firewall may also drop it silently. That ambiguity is why open|filtered must remain inconclusive, as described in Nmap's UDP scan documentation.

SANITIZED EXAMPLES

Show the scope without exposing the network

These are illustrative command templates for a future authorized retest, not a transcript of the September investigation. Replace the bracketed placeholders before use. The UDP port list must be selected for that retest; it is not recovered here.

# Illustrative TCP scan template
sudo nmap -sS -p 1-65535 --reason "[Home_Public_IP]"

# Illustrative selected-UDP scan template
sudo nmap -sU -p "[UDP_Port_List]" --reason "[Home_Public_IP]"

-sS selects a TCP SYN scan; -sU selects UDP scanning; -p specifies the ports; and --reason asks Nmap to explain its state classifications. A public address such as [Home_Public_IP] and an internal host such as [Private_Lab_IP] represent different test targets and access paths.

Publication boundary: No real network addresses, credentials, hostnames, packet payloads, or detailed firewall configurations are included. These examples contain no fabricated results, service banners, or packet timestamps.

LIMITATIONS

What this assessment could not establish

  • Original logs and captures were not independently rechecked for this publication. Exact scan flags, retries, timeouts, selected UDP ports, and raw packet details are not reproduced.
  • Results reflect one external VPS and the September 28–29 test window. Other source networks, later configuration changes, or intermittent services could produce different observations.
  • Capture placement, filters, and time coverage were not independently revalidated. Missing replies in the examined capture cannot identify where traffic stopped.
  • Full TCP port coverage did not make this a comprehensive security audit. UDP coverage was limited, and IPv6 exposure, internal services, software weaknesses, and credentials require separate assessment.

SECURITY CONCLUSIONS

A bounded conclusion, with uncertainty preserved

Assessment conclusion: The available evidence did not establish externally reachable TCP services across the tested port range from the cloud VPS during the investigation. Selected UDP results remained inconclusive. The cause of the missing replies was not isolated.

I could report the tested TCP outcome, but could not conclude that all ports were closed, that every firewall control was effective, or that the network was secure in every respect. No specific exposed UDP service was demonstrated by an open|filtered label.

Separate observation from hypothesis

An unanswered probe is an observation. A particular router dropping that probe is a hypothesis until corroborated.

Carry uncertainty into SOC reporting

Report the vantage point, test scope, and evidence gaps. Recommend the next check without turning an ambiguous result into a confirmed incident or an assurance of safety.

PLANNED FOLLOW-UP

Make the next assessment easier to verify

These are proposed improvements, not completed work.

  1. Preserve a reproducible evidence set

    Confirm the target address and external route, record UTC timestamps and scanner version, and retain scan commands, machine-readable results, and captures privately. Publish only reviewed, sanitized excerpts.

  2. Correlate both sides of the path

    Compare synchronized captures and available firewall logs at authorized points with the internal service inventory. Look for probe arrival and reply departure before attributing a failure to a device.

  3. Resolve remaining scope gaps

    Use appropriate protocol-specific requests for selected UDP services, document unresolved ambiguity, assess IPv6 separately if applicable, and repeat external checks after relevant configuration changes.

TECHNICAL REFERENCES

How the scan states are interpreted

These primary sources explain Nmap's behavior. They do not independently verify the lab findings.

Return to featured projects