Skip to content

Bulkhead Isolation

Partition shared resources or failure domains into bounded compartments so local failure stays contained instead of spreading through coupling.

Solution archetype #
128
Problem family
Fragility, Failure & Continuity Risk
Problem subfamily
Fault Containment & Bounded Service Loss

The Diagnostic Story

Symptom: One part of the system runs into trouble — overloaded, compromised, or failing — and everything connected to it starts failing too. The problem is not that parts fail; it is that failure travels freely through shared capacity, shared state, or shared pathways. What should be a local incident becomes a system-wide crisis because there is no seam to stop propagation. Unaffected parts get pulled down not by their own faults but by what they share with the failing part.

Pivot: Partition the system into bounded compartments so that shared capacity, state, and flow can only cross compartment lines through controlled interfaces. Each compartment can fail, saturate, or degrade without consuming the capacity or corrupting the state of its neighbors.

Resolution: A failure stays local: the damaged compartment can degrade, recover, or be bypassed while unaffected compartments retain enough capacity and function to keep operating. The system trades some pooling efficiency and coordination simplicity for a bounded blast radius and independent recoverability.

Reach for this when you hear…

[platform SRE] “The recommendations service melted down and took checkout with it because they're all sharing the same thread pool — we need separate resource budgets for each workload.”

[public health] “If we cohort the exposed residents separately instead of mixing them with the general population, we contain the spread even if our testing is slow.”

[financial risk] “The whole point of ring-fencing that subsidiary is so if it blows up, the losses stay inside the fence and don't take the parent institution with them.”

When This Archetype Applies

Complete catalog groundingAt least one sufficient condition set is fully represented by existing primes or domain-specific abstractions.

Shared resources, state, or dependencies allow a local failure, overload, contamination, or exhaustion event to propagate into otherwise viable parts of the system.

Show the applicability expression

Applicability expression1 distinct condition

Shared-path failure propagation
Algebraic1

groundedpartly groundedopen

1 condition, all required.

1Required in every casenumbered 1–1

These hold no matter which pattern applies.

1

Shared-path failure propagation · grounded

A local failure can propagate beyond its origin through shared resources, state, capacity, or dependencies.

Other requirements and context (1)

Why these sit outside the expression

Solution feasibilityit describes whether the intervention can work, not whether the diagnostic problem exists.

  • Solution feasibilityThe system can be divided into bounded failure domains without destroying its protected function.

1 of 1 conditions grounded.

Read the methodologyDownload the trigger-logic data

Abstractions this archetype builds on — directly (a source ingredient) or as a related pattern. Links follow the typed catalog namespace.

Built directly on (6)

  • Boundary: Defines system limits.
  • Containment: Holding a hazard, process, or agent within a deliberately maintained perimeter to prevent its spread or uncontrolled interaction with the surroundings.
  • Fault Tolerance: Continue operating under failure.
  • Modularity: Breaks systems into smaller units.
  • Resource Partitioning
  • Selective Coupling

Also references 8 related abstractions

  • Constraint: Limits possibilities to guide outcomes.
  • Coupling: Interdependence among subsystems.
  • Flow: Structured movement of energy, matter, or information.
  • Hierarchy: Organizes elements into levels or ranks.
  • Resilience: Absorb shocks and adapt.
  • Resource Management: Allocation of finite assets.
  • Topology: Studies properties preserved under deformation.
  • Trade-offs: Balancing competing priorities.

Variants

Narrower or domain-specific specializations that share this archetype's core structure. Recognized variants are established; candidate variants are provisional.

Blast Radius Reduction · risk or failure variant · recognized

A framing of bulkhead isolation focused on limiting how far a failure can propagate.

Editorial Notes

Problem Classification

Classification: Fragility, Failure & Continuity RiskFault Containment & Bounded Service Loss

Problem kernel: local faults can spread across shared failure domains

Rationale: Coupled resources lack compartments, so a defect in one zone can cross boundaries and remove the wider function.

Independent corroboration: The earliest necessary condition in the frozen evidence is: Partition shared resources or failure domains into bounded compartments so local failure stays contained instead of spreading through coupling. That is a fault containment and bounded service loss problem because A local fault or partner failure can eliminate the whole function because failure domains, alternate activation, and essential degraded service are absent.

Review outcome: Independent reviewer agreement; high confidence.