DFMEA - a “how-to” within the Networking Industry

Written By Nate Ross

What Is DFMEA in an ISP Environment?

  • Design Failure Mode and Effects Analysis (DFMEA) is a way to look at every step of a circuit design and ask:

    1. “How could this fail?”

    2. “What happens if it fails?”

    3. “How can we detect or prevent it?”

    By doing this early—before circuits are fully installed and turned up—you can avoid outages, costly rework, and unhappy customers.

    • Reference: “FMEA | ASQ” (https://asq.org/quality-resources/fmea)

Why Is DFMEA Important for an ISP Delivering Circuits?

  1. Proactive Problem Prevention

    • ISPs can identify issues like fiber patch cable mishandling, incorrect transceiver types (e.g., 10Gb SFP+ mismatch), or single points of failure in the route before a large enterprise experiences downtime.

    • Reference: “Network Design Best Practices | Cisco” (https://www.cisco.com/c/en/us/solutions/enterprise-networks/branch-solutions/design-guides.html)

  2. Cost Savings

    • Fixing circuit design flaws after a customer is live can be very expensive. DFMEA helps catch problems before deployment, which is almost always cheaper.

    • Reference: “Corrective vs. Preventive Action | iSixSigma” (https://www.isixsigma.com/dictionary/corrective-and-preventive-action-capa/)

  3. Higher Reliability & Customer Satisfaction

    • Fewer circuit failures mean better uptime, which increases trust and satisfaction from large enterprise customers.

    • Reference: “High Availability and Redundancy | Cisco” (https://www.cisco.com/c/en/us/support/docs/high-availability/)

Key DFMEA Sections

A DFMEA is typically shown in a spreadsheet or table. It contains:

  1. Design Function: What each part of the circuit or system is meant to do.

  2. Potential Failure Mode: Ways it could fail (e.g., fiber break, power loss, configuration error).

  3. Potential Effects: Consequences if that failure happens (e.g., total outage, degraded performance).

  4. Possible Causes: Reasons for that failure (e.g., poor splicing, incorrect port configuration, single path with no redundancy).

  5. Current Controls: Existing methods to prevent or detect the problem (e.g., SLA monitoring tools, redundant paths).

  6. Severity (S): How bad the effect is if this failure occurs (on a scale of 1 to 10).

  7. Occurrence (O): How often you expect this failure to happen (1 to 10).

  8. Detection (D): How likely your current controls are to catch this issue before it becomes a problem (1 to 10).

  9. Risk Priority Number (RPN): Calculated as S × O × D. Higher RPN = higher priority to fix.

  10. Recommended Actions: Specific improvements or design changes to reduce the risk.

  • Reference: “FMEA Fundamentals | iSixSigma” (https://www.isixsigma.com/tools-templates/fmea/fmea-fourth-edition-fundamentals/)

Step-by-Step: DFMEA for New ISP Circuits

1. Identify the Scope

  • Focus on 1Gb and 10Gb circuits from the customer’s premise equipment (CPE) to the ISP’s backbone and peering routers. Include all connections: the local fiber patch, the main fiber run, the intermediate wiring centers, and the final peer exchange point.

  • Reference: “Circuit Design and Implementation | Cisco” (https://www.cisco.com/c/en/us/support/docs/optical-networking/)

2. List All Functions

  • For each circuit path, outline the main functions. Examples:

    • Physical Connectivity: Provide a stable 1Gb/10Gb link.

    • Signal Integrity: Maintain error-free data transfer.

    • Peering & Routing: Exchange routing information reliably with peer providers.

  • Reference: “Layer 1 (Physical Layer) Design Considerations | Fluke Networks” (https://www.flukenetworks.com/knowledge-base/cabling)

3. Identify Potential Failure Modes

  • Brainstorm how each function might fail. Examples:

    • Fiber Patch Cable Damage: Kinked or pinched cable causing high loss.

    • SFP Module Failure: Transceiver mismatch or defective optics.

    • No Redundant Path: Single route that, if cut, leads to outage.

    • Power Loss at Wiring Center: UPS/generator failure.

  • Reference: “Common Fiber Failures | Corning” (https://www.corning.com/optical-communications/worldwide/en/home.html)

4. Determine Potential Effects

  • For each failure mode, ask: “What’s the worst thing that can happen?” Examples:

    • Fiber Patch Cable Damage: Complete circuit outage or high packet loss.

    • SFP Module Failure: Circuit flaps or link does not come up at all.

    • No Redundant Path: Customer site goes offline if fiber is cut.

  • Reference: “Impact of Single Points of Failure | Cisco” (https://www.cisco.com/c/en/us/solutions/enterprise-networks/high-availability.html)

5. Look at Possible Causes

  • Dive deeper into the root causes. Examples:

    • Physical Wear & Tear: Cable jackets rubbing in cable trays.

    • Incorrect Transceiver Type: 10Gb LR used where only SR optics are compatible.

    • Underground Fiber Cut: Construction crews hitting buried lines.

  • Reference: “Root Cause Analysis | iSixSigma” (https://www.isixsigma.com/tools-templates/cause-effect/dmaic-approach-root-cause-analysis/)

6. Assess Current Controls

  • Document what measures are in place today to prevent or detect issues:

    • Network Monitoring Tools (e.g., real-time link monitoring, SNMP traps).

    • Dual-Homed Circuits (redundant paths with diverse routing).

    • Regular Maintenance & Testing (fiber characterization, cleaning connectors).

  • Reference: “Network Monitoring | SolarWinds” (https://documentation.solarwinds.com/en/success_center)

7. Assign Severity (S), Occurrence (O), Detection (D)

  • Use 1–10 scales (1 = Low, 10 = High). For example:

    • Severity: 9 if it brings down an entire data center connection.

    • Occurrence: 3 if it rarely happens (e.g., fiber breaks in a secure data center).

    • Detection: 7 if it is hard to detect before it happens (no advanced alert).

  • Reference: “Risk Priority Number (RPN) in FMEA | iSixSigma” (https://www.isixsigma.com/tools-templates/fmea/fmea-rpn/)

8. Calculate RPN

  • Multiply S × O × D to get RPN.

  • Decide on a threshold; for instance, any RPN above 100 requires a design revision.

  • Reference: “RPN Thresholds | iSixSigma” (https://www.isixsigma.com/tools-templates/fmea/fmea-rpn/)

9. Recommend Actions

  • For any high RPN items, list concrete steps to reduce risk. Examples:

    • Add a Second Fiber Path: Achieve route diversity.

    • Specify Higher-Grade Transceivers: Use more reliable modules with advanced diagnostic monitoring.

    • Improve Power Backup: Ensure generator and UPS are tested regularly.

  • Reference: “Network Redundancy & Resiliency | Cisco” (https://www.cisco.com/c/en/us/support/docs/availability/high-availability/)

10. Implement, Test, and Track

  • Update the design with the recommended actions.

  • Conduct thorough lab or pilot tests to confirm improvements.

  • Keep the DFMEA document current whenever new circuits or changes are introduced.

  • Reference: “Continuous Improvement | ASQ” (https://asq.org/quality-resources/continuous-improvement)

Sample DFMEA Table (High Granularity Example)

Below is a simplified example table with explicit references to interfaces, circuits, and components. Adjust the Severity, Occurrence, and Detection scales to your ISP’s standards.



Example DFMEA Chart

How to Interpret RPN: Here, any RPN over 100 might require immediate design revisions. For instance, the first row (RPN 160) might lead you to invest in bend-insensitive fiber or re-route cables to avoid tight bends.

Practical Benefits of DFMEA for ISPs

  1. Reduced Outages: Pinpointing risky single points of failure (e.g., fiber paths sharing the same conduit) helps the ISP reroute or add true physical diversity.

  2. Better Customer Satisfaction: Enterprises get more reliable circuits, leading to positive feedback and potentially more business.

  3. Efficient Use of Resources: Prioritizing highest RPN items first ensures the engineering team spends time on the most critical design flaws.

  4. Living Document: Update the DFMEA whenever new sites or hardware changes occur, keeping it relevant over time.

  • Reference: “Continuous Improvement | ASQ” (https://asq.org/quality-resources/continuous-improvement)

Final Thoughts & Next Steps

  • 1. Document Completion: Fill out a DFMEA table for each segment of the circuit path.

  • 2. Set RPN Thresholds: Decide on the RPN cutoff for immediate design changes (e.g., 100).

  • 3. Implement Changes & Test: Update the design to address high-risk items. Run pilot tests.

  • 4. Ongoing Updates: Revise the DFMEA as circuits evolve or new fiber routes are added.

By following these steps, an ISP can deliver robust, dependable circuits to large enterprises, identifying and managing risks before they cause real-world failures.

References (Linked)

  1. “FMEA | ASQ” (https://asq.org/quality-resources/fmea)

  2. “Network Design Best Practices | Cisco” (https://www.cisco.com/c/en/us/solutions/enterprise-networks/branch-solutions/design-guides.html)

  3. “Corrective vs. Preventive Action | iSixSigma” (https://www.isixsigma.com/dictionary/corrective-and-preventive-action-capa/)

  4. “High Availability and Redundancy | Cisco” (https://www.cisco.com/c/en/us/support/docs/high-availability/)

  5. “FMEA Fundamentals | iSixSigma” (https://www.isixsigma.com/tools-templates/fmea/fmea-fourth-edition-fundamentals/)

  6. “Circuit Design and Implementation | Cisco” (https://www.cisco.com/c/en/us/support/docs/optical-networking/)

  7. “Layer 1 (Physical Layer) Design Considerations | Fluke Networks” (https://www.flukenetworks.com/knowledge-base/cabling)

  8. “Common Fiber Failures | Corning” (https://www.corning.com/optical-communications/worldwide/en/home.html)

  9. “Impact of Single Points of Failure | Cisco” (https://www.cisco.com/c/en/us/solutions/enterprise-networks/high-availability.html)

  10. “Root Cause Analysis | iSixSigma” (https://www.isixsigma.com/tools-templates/cause-effect/dmaic-approach-root-cause-analysis/)

  11. “Network Monitoring | SolarWinds” (https://documentation.solarwinds.com/en/success_center)

  12. “Risk Priority Number (RPN) in FMEA | iSixSigma” (https://www.isixsigma.com/tools-templates/fmea/fmea-rpn/)

  13. “RPN Thresholds | iSixSigma” (https://www.isixsigma.com/tools-templates/fmea/fmea-rpn/)

  14. “Network Redundancy & Resiliency | Cisco” (https://www.cisco.com/c/en/us/support/docs/availability/high-availability/)

  15. “Continuous Improvement | ASQ” (https://asq.org/quality-resources/continuous-improvement)

Nate Ross
Next

How to Achieve Continuous Improvement