hatiyildiz f5daac52af refactor(platform): remove k8gb — replaced by PowerDNS lua-records (#171 )

PowerDNS lua-records (`ifurlup`, `pickclosest`, `ifportup`) cover everything
k8gb was doing — geo-aware response selection, health-checked failover,
weighted round-robin — at the authoritative DNS layer. Eliminates a
separate K8s controller, CRD set, and CoreDNS plugin from every Sovereign.

Changes:
- platform/k8gb/ deleted (Chart.yaml, values.yaml, blueprint.yaml never
  authored — only README existed)
- products/catalyst/bootstrap/ui/public/component-logos/k8gb.svg deleted
- componentGroups.ts: remove k8gb component (PowerDNS already there)
- componentLogos.tsx: drop logo_k8gb + k8gb map entry
- model.ts DEFAULT_COMPONENT_GROUPS spine: replace k8gb with powerdns
- StepInfrastructure.tsx: copy refers to PowerDNS lua-records, not k8gb
- provision.html: replace k8gb tile and edges with powerdns
- catalog.generated.ts regenerated (now includes bp-powerdns)
- docs sweep — every k8gb reference in PLATFORM-TECH-STACK, NAMING-
  CONVENTION, SOVEREIGN-PROVISIONING, SRE, ARCHITECTURE, GLOSSARY,
  COMPONENT-LOGOS, IMPLEMENTATION-STATUS, BUSINESS-STRATEGY,
  TECHNOLOGY-FORECAST, README, infra/hetzner/README, platform READMEs
  (cilium, external-dns, failover-controller, litmus, flux, opentofu)
  rewritten to point at PowerDNS lua-records / MULTI-REGION-DNS.md.
  Historical entries in VALIDATION-LOG.md preserved as audit trail.
- New docs/MULTI-REGION-DNS.md — canonical reference for the lua-record
  patterns (ifurlup all/pickclosest/pickfirst, ifportup, pickwhashed),
  Application Placement → lua-record selector mapping, when to add a
  second Sovereign region, operational checks.

Closes #171.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

2026-04-29 08:51:09 +02:00

9.6 KiB

Raw Blame History

Technology Forecast 2027-2030

Component technology assessment and strategic forecast for the OpenOva platform.

Status: Accepted | Updated: 2026-04-28

Overview

This document provides a forward-looking assessment of all 56 platform components, evaluating their relevance trajectory through 2027 and 2030 in the context of AI-driven development, regulatory evolution, and cloud-native ecosystem maturation.

Scoring Methodology

Each component is scored on a 0-100 scale across three time horizons:

2026 (Current): Today's relevance and essentiality
2027 (Near-term): Expected relevance in 12 months
2030 (Long-term): Expected relevance in 4 years

Factors considered: AI replacement risk, regulatory demand, ecosystem maturity, operational complexity, and community momentum.

Mandatory Components (29)

Classification basis: "Mandatory" = installed on every Sovereign — comprises the Catalyst control plane (per PLATFORM-TECH-STACK.md §2) plus per-host-cluster infrastructure (§3) plus the transient Phase-0 bootstrap component (sealed-secrets, used by the bootstrap kit until External Secrets Operator + OpenBao take over). "A La Carte" below = Application Blueprints (§4) that customers opt into per Environment.

Component	2026	2027	2030	Trend	Notes
cert-manager	95	95	95	Stable	TLS automation is evergreen
cilium	95	96	95	Stable	eBPF dominance continues
external-secrets	95	95	95	Stable	Secrets management is foundational
openbao	93	93	90	Stable	May face cloud-native alternatives
flux	92	92	88	Stable	GitOps is established pattern
seaweedfs	92	93	92	Stable	Replaces MinIO in 2026-04. Native S3 + tiered cloud archive in one component; Apache 2.0; encapsulates cloud-S3 backends behind a single in-cluster S3 endpoint
velero	92	90	88	Stable	Backup is non-negotiable
harbor	90	92	92	Rising	Supply chain security drives adoption
falco	90	92	93	Rising	Runtime security more critical with AI workloads
trivy	90	90	88	Stable	Complemented by Syft/Grype for SBOM
sigstore	90	93	95	Rising	Regulatory mandates drive adoption (EU CRA)
syft-grype	90	92	94	Rising	SBOM requirements accelerating
coraza	88	88	85	Stable	WAF remains necessary for web apps
external-dns	90	88	85	Stable	DNS synchronization is mechanical
powerdns	88	88	88	Stable	Authoritative DNS + lua-records (geo + health-checked failover) — replaces k8gb in OpenOva's Catalyst stack
grafana	88	88	85	Stable	AI may generate dashboards but metrics collection stays
kyverno	88	90	90	Rising	Policy-as-code increasingly mandated
crossplane	78	80	82	Rising	Cloud resource management via CRDs maturing
opentofu	82	80	75	Declining	Bootstrap-only; Crossplane takes over day-2
gitea	83	82	78	Stable	Self-hosted Git remains important
keda	80	82	85	Rising	Event-driven autoscaling more relevant with AI
vpa	78	78	75	Stable	Right-sizing is ongoing need
reloader	80	80	78	Stable	Simple operator, high value
failover-controller	82	82	80	Stable	Multi-region failover always needed
keycloak	85	85	85	Stable	Catalyst control-plane identity — per-Org realms in SME, per-Sovereign realm in corporate
spire	88	90	92	Rising	SPIFFE/SPIRE workload identity — 5-min rotating SVIDs, root server on mgt + per-host-cluster agent. Zero-trust mTLS substrate beneath Cilium.
nats-jetstream	90	92	92	Rising	Catalyst control-plane event spine (pub/sub + Streams + KV). Apache 2.0. Replaces Redpanda + Valkey for the control plane only; per-Organization Accounts.
sealed-secrets	75	70	60	Declining	Transient Phase-0 bootstrap-only — used to seal the initial OpenBao unseal keys before ESO + OpenBao take over day-2 secret distribution. Decommissioned after Phase 1.

Note on OpenTelemetry

OpenTelemetry is mandatory but has no separate platform directory - it is deployed as part of the observability stack via Grafana Alloy configuration.

A La Carte Components (28)

Component	2026	2027	2030	Trend	Notes
vllm	95	96	95	Stable	LLM inference is essential for AI
kserve	88	90	90	Rising	Model serving standardizing
milvus	88	90	90	Rising	Vector search increasingly critical
cnpg	90	90	90	Stable	PostgreSQL operator is proven
opensearch	50	55	60	Rising	Application Blueprint — opt-in for SIEM (paired with ClickHouse + bp-specter)
valkey	85	85	85	Stable	Caching is evergreen
nemo-guardrails	90	92	93	Rising	AI safety regulations expanding
langfuse	90	92	90	Rising	LLM observability maturing
llm-gateway	87	88	85	Stable	Claude Code access proxy
anthropic-adapter	85	85	80	Stable	API translation layer
bge	82	82	78	Stable	Embeddings are commoditizing
knative	75	75	72	Stable	Serverless for model serving
librechat	75	75	70	Stable	Chat UI may be AI-generated
ferretdb	75	78	80	Rising	MongoDB compatibility without SSPL
strimzi	72	72	70	Stable	Event streaming is established
debezium	70	70	68	Stable	CDC tied to streaming adoption
temporal	68	72	75	Rising	Saga orchestration gaining relevance
flink	60	65	70	Rising	Stream processing for real-time analytics
clickhouse	55	60	65	Rising	OLAP analytics with SIEM workloads
iceberg	50	55	65	Rising	Open table format adoption accelerating
stalwart	70	70	68	Stable	Self-hosted email is niche but needed
stunner	68	70	72	Rising	WebRTC gaining enterprise adoption
livekit	72	75	78	Rising	Video/audio infrastructure growing
matrix	70	72	75	Rising	Self-hosted chat gaining traction
neo4j	65	68	70	Rising	Knowledge graphs for RAG
litmus	72	75	78	Rising	Chaos engineering for compliance proof
openmeter	55	58	60	Stable	Usage metering is niche
guacamole	70	72	75	Rising	Clientless remote-desktop gateway with Keycloak SSO + session recording; replaces VPN+native-client distribution for auditable RDP/VNC/SSH access in regulated environments

Product Impact Analysis

OpenOva Cortex (AI Hub)

Trend: Strong growth. AI infrastructure components are all stable or rising. Addition of NeMo Guardrails and LangFuse addresses critical gaps in AI safety and observability. Removal of Airflow, SearXNG, and LangServe simplifies the stack without losing capability (AI generates custom code).

OpenOva Fingate (Open Banking)

Trend: Stable. Core components (Keycloak, OpenMeter) remain solid. Removal of Lago (billing is customer-specific) simplifies the product. PSD2/DORA regulatory pressure continues to drive demand.

OpenOva Fabric (Data & Integration)

Trend: Rising. Merging Titan + Fuse into Fabric creates a stronger product. Strimzi, Flink, Temporal, and ClickHouse are all stable-to-rising. Removal of declining components (Airflow, Trino, Superset, Camel, Dapr, RabbitMQ, ActiveMQ) makes the product leaner and more focused.

OpenOva Relay (Communication)

Trend: Rising. Self-hosted communication (email, video, chat) is a growing enterprise need driven by data sovereignty and compliance. All four components (Stalwart, LiveKit, STUNner, Matrix) are stable-to-rising.

OpenOva Specter (AIOps)

Trend: Strong growth. AI-powered operations become more capable as LLM quality improves. SIEM/SOAR integration (Falco + OpenSearch + ClickHouse) provides rich context for autonomous incident response.

Strategic Recommendations

Components to Watch (potential additions 2027+)

Component	Score	Rationale
Ray	80	Distributed AI compute (training, batch)
MLflow	78	Model registry + experiment tracking
OpenCost	75	FinOps for K8s/GPU cost visibility
Flagger	72	Progressive delivery (canary, blue-green)

Risks to Monitor

Risk	Impact	Mitigation
eBPF kernel API changes	Cilium compatibility	Track kernel LTS versions
OpenSearch license change	SIEM backend	Fork-ready posture
GPU supply constraints	Cortex deployment	Multi-GPU vendor support
EU CRA enforcement timeline	Supply chain security	Sigstore + Syft/Grype already in stack

Removed Components (Rationale)

Component	Score (2026)	Why Removed
Backstage	45	Replaced by Catalyst console (the platform's own developer-facing UI)
MongoDB	72	Replaced by FerretDB on CNPG (no SSPL, simpler DR)
Airflow	33	Replaced by Flink + OTel (AI generates workflows)
Superset	40	AI-generated visualizations replace dashboard tools
Trino	38	ClickHouse + CNPG direct queries sufficient
LangServe	73	Custom RAG behind KServe (AI generates integration code)
SearXNG	40	LLM Gateway tool registry replaces meta-search
Camel K	20	AI generates integration code directly
Dapr	30	Sidecar overhead unnecessary; Kafka + custom code
RabbitMQ	25	Kafka covers event streaming
ActiveMQ	12	JMS legacy, no modern use case
Vitess	15	MySQL sharding is niche
Lago	58	Billing is customer-specific, not platform concern

Part of OpenOva

9.6 KiB Raw Blame History