Supporting case · Resilience & Security

GCP to Bare Metal: Cost and Control

An infrastructure migration where cost and direct network control had to be solved together.

Disclosure

This adaptation keeps the migration costs visible. The result applies to this platform and is not universal advice against cloud infrastructure.

One-sentence outcome

The migration produced a 30% infrastructure cost reduction while also restoring the network control the platform needed.

Context and stakes

The platform's traffic profile made egress costs visible at executive level, while application-layer defense work exposed where cloud abstraction stopped short of the operational control the system needed.

Baseline and constraints

The cloud default still had real strengths, and the migration carried painful trade-offs: overlapping infrastructure spend, harder Kafka operations, new SRE depth requirements, and temporary delivery friction for the team.

Exact role

Chief Technology Officer owning the migration case, the economic argument, and the operational choice to accept more infrastructure responsibility in exchange for cost and network control.

Trigger to measured change

Trigger
Egress bills and custom edge-defense requirements exposed that the cloud abstraction had become both a cost surface and a control ceiling.
Constraint
Leaving the cloud meant accepting temporary double spend, steeper SRE demands, and business impatience while the new stack stabilized.
Decision
Move from GCP to a bare-metal and hosted Kubernetes setup where network behavior, edge controls, and longer-term cost structure could be owned directly.
Measured change
The reported result is a 30% infrastructure cost reduction, with the caveat that the path there was operationally painful and not a generic prescription.

Reduced infrastructure cost by 30% after a GCP-to-bare-metal migration.

State the system scope and the migration trade-offs. Do not turn this into universal anti-cloud advice or imply that the saving came from DDoS work alone.

Context: Anonymized iGaming infrastructure with large data movement and edge-control demands.
Timeframe: Migration retrospective recorded in the phase-0 ledger on 2026-09-02.
Baseline: GCP egress and network-control limits versus the higher operational burden of owned infrastructure.
Role: Chief Technology Officer owning the migration case, the economic argument, and the operational choice to accept more infrastructure responsibility in exchange for cost and network control.
Source: Review the source record
Provenance: Source-language LinkedIn post.
Confidence: B - first-party retrospective with explicit trade-off caveat.
Disclosure: Anonymized case; the platform and organization remain unnamed.

Measured or observable result

Infrastructure cost fell by 30% after the move from GCP to bare metal. More control and lower spend came with a harder migration path.

Attribution and caveat

State the system scope and the migration trade-offs. Do not turn this into universal anti-cloud advice or imply that the saving came from DDoS work alone.

Retained capability

The organization retained direct network control, more predictable latency behavior, and an SRE team that had crossed a deeper operational threshold.