Three-tier Rakuten kernel proposal — 8.10 preferred, 8.6 sustaining, $800k-$1M PS for full 8.6

May 4, 2026 at 10:52 PMstrategyhigh

Situation

Push Rakuten to migrate RLC 8.6 to 8.10. Three-tier proposal: (1) preferred — full support on 8.10 with CIQ vendor coordination to accelerate hardware recertification; (2) alternative — sustaining support on 8.6 with no new patches/backports (security risk on Rakuten); (3) PS engagement — $800k-$1M/year to fund two dedicated kernel engineers for full 8.6 support, framed explicitly as Professional Services cost not mainline engineering. June renewal is the forcing function. The original handshake-pricing deal with Tarek is void.

Reasoning

8.6 has a unique kernel architecture; CVE backports arent feasible without dedicated headcount — the cost has to land on Rakuten, not on CIQs kernel team. The handshake deal with Tarek is gone, so theres no political reason to preserve legacy pricing — restart from real economics. Three tiers do two jobs: (a) make migration to 8.10 the cheap, easy choice; (b) if Rakuten still wants 8.6, the price is the price — not a discount. The sustaining-support middle tier is a face-saver that lets Rakuten say no to migration without forcing them to the $1M tier and protects CIQ from we-are-paying-you pressure for backports. Same lever as Everfox: large customer wants custom work, must fund the engineering, not extract it.

Additional Context

Direct precedents: 2026-01-30 LTS roll-forward policy, 2026-04-27 Route Non-Differentiating FTE Classes to Partners. Same session as Everfox $2M floor (D4) — common lever. Hardware certification is Rakutens true blocker; CIQ offering vendor coordination removes their stated reason for staying on 8.6.

Observed Evidence

Peter explicitly framed the three options in Rakuten Proposal Overview meeting; reinforced same framing to Greg in 1:1. Direct quote: This is a professional services cost, not a mainline engineering expense.

Matching Patterns

60%
Route Non-Differentiating FTE Classes to Partners(legacy specialized burden routed to paid lane)

Confidence Breakdown

30/35
Evidence
25/30
Pattern
19/20
Source
13/15
Corroboration

Reasoning Depth Analysis

Org Signal:Legacy support is priced at the cost of dedicated engineers. Goodwill discounts have an end date.
Who Affected:Nathan (kernel team protected from 8.6 absorption), Suzanne (selling clear options vs fuzzy we-figure-it-out), Ramesh (proposal owner), every other large customer asking for legacy version preservation
Precedent:Every legacy-version request from a large customer gets the three-tier shape: migrate, sustain, or fund.
Consequences:Rakuten may push back hard or threaten REL move. Acceptable vs absorbing 8.6 indefinitely with no funding.
Timing:June renewal forces the conversation; better to set the negotiation frame than react to Rakutens counter.

Source

reflection

AI Confidence

87%

Related Context

🎥
Rakuten Proposal Overview 5/4

fathom

Three-Tiered Proposal: Preferred Full support for RLC 8.10. Alternative Sustaining support for RLC 8.6 (no new patches/backports). PS Engagement Full 8.6 support via dedicated team. Cost $800k-$1M/year to fund two high-paid kernel engineers. Rationale: This is a professional services cost, not a mainline engineering expense.

🎥
Peter <> Greg 1:1 5/4

fathom

Peter proposed a migration plan to 8.10 or a new contract requiring funding for two full-time engineers. The original low-cost contract was a personal handshake deal with Tarek, now void.

Outcome

No outcome recorded yet.

Decision ID: 93af04af-e959-4b25-a472-ba27df414a90