Open core ↔ Enterprise

DimensionOpen coreEnterprise

Tenancy

Single-tenant / self-hosted

Multi-tenant SaaS or dedicated Ent account

Multi-step durability

Queues + cron + Durable Objects

Workflows (hours/days, step state, human gates)

Audit

Analytics Engine + structured logs

Logpush → R2 / SIEM, Tail Workers, long retention

Security rings

Five-layer model

Layers 0–5+ (Bot Mgmt, API Shield, Access, AI Gateway)

Real-time

REST default; optional WS DO

Hibernatable WS, event-driven risk

AI

Workers AI + basic RAG

AI Gateway, cost control, institutional volume

Support

Community / self-serve

Commercial SLAs, runbooks, higher limits

Capability pillars

Six levers that amplify the open core without reintroducing regional VMs.

01Multi-tenancyUpcoming

Workers for Platforms

True multi-tenant isolation: Dispatch Worker routes by tenant claim / subdomain to isolated User Workers per fund, strategy book, or customer.

▸Namespaces, tags for metering and billing
▸Per-tenant DOs, Queues, D1/R2 paths, secrets
▸Synchronous first deploy — routable when HTTP 200
▸Hosted SaaS or dedicated Enterprise account

Docs ↗

02Durable executionUpcoming

Cloudflare Workflows

Long-running, step-persisted trade and compliance processes — hours to days — with retries, pause-for-external-events, and human approval gates.

▸Full trade lifecycle: validate → risk → execute → reconcile → report
▸Atomic kill-switch + flatten sequences
▸Compliance reports with approval gates
▸Post-trade reconciliation across exchanges and storage

Docs ↗

03ComplianceUpcoming

Logpush · Traces · R2 Audit

Regulator-grade observability: Workers Trace Events, HTTP, WAF, Queues, DOs pushed immutably to R2 or SIEM with long retention.

▸One-click R2 Logpush + custom fields
▸Tail Workers for sample / filter / enrich
▸End-to-end traces across Bindings, DOs, Queues, Workflows
▸Analytics Engine for metrics; Logpush for audit

Docs ↗

04Layers 0–5+Upcoming

Extended Security Stack

Open-core five-layer model plus Enterprise Bot Management, API Shield, Zero Trust Access, mTLS, and AI guardrails.

▸Bot scores, JA3/JA4, verified bots on signal ingress
▸API Shield: schema, JWT, mTLS, sequence mitigation
▸Access SSO, SCIM, device posture for ops surfaces
▸AI Gateway: sanitization, rate limit, cost control

Docs ↗

05Real-timeUpcoming

Hibernatable WebSockets

Persistent exchange connectivity via Durable Object hibernatable WebSockets — live fills, partials, and books without burning CPU while idle.

▸ExchangeConnectionManager DO pattern
▸Event-driven risk beside (or instead of) cron
▸Cheap keep-alive while DO sleeps
▸Pairs with Workflows for durable stream handling

Docs ↗

06Limits · AI · DataUpcoming

Institutional Scale

Higher CPU, memory, subrequests, D1/R2/Queue/Vectorize quotas, AI Gateway at volume, and data residency controls on R2.

▸Custom limit increases via account team
▸AI Gateway logging, caching, multi-provider failover
▸R2 jurisdiction + lifecycle for immutable audit
▸Hyperdrive for hybrid historical data if needed

Docs ↗

Security layers 0–5+

0

Bot Management

Scores, verified bots, JA3/JA4, early drop of toxic signal sources

1

WAF + API Shield

Schema validation, JWT, mTLS, 100+ rate-limit rules

2

Zero Trust Access

SSO, SCIM, device posture on dashboard and mgmt APIs

3

Internal + mTLS

Service Binding auth + per-tenant claims, optional mTLS

4

DO + crypto controls

Idempotency mutex, signed audit records, fail-closed isolation

5

AI Guardrails

Gateway sanitization, rate limits, cost ceilings on LLM paths

Non-negotiable invariants

01Edge-native composition — Service Bindings, DOs, Smart Placement stay the mesh
02Evolutionary — open-core workers become templates and system services, not a fork-from-scratch
03Isolation before convenience — tenant leakage fails closed at dispatch
04Audit by default — security events are immutable appends
05Open core stays useful — institutions scale up; individuals are not degraded

Documentation

Enterprise hub ↗

Commercial layer overview and document index

Architecture ↗

Topology, Workflows, WfP, scale levers

Multi-tenancy ↗

Dispatch, User Workers, isolation primitives

Observability & audit ↗

Logpush, traces, Tail Workers, R2

Security & compliance ↗

Extended 5-layer model + Ent controls