商店現已上線

我們打造用於網站、AI 智慧體和 LLM 訓練的 Workstation AI 機器——線上配置併購買。

獲取同行價格
AI 產品
AI 實驗室OpenAI 智慧體Claude 代理Grok BotWorkstation AI 原生 CRM營銷所有產品
AI 解決方案
應用型 AIAI 智慧體生成式 AIAI 自動化AI 技能提升中小企業 AI 套餐企業 AI 實驗室AI 工作站NVIDIA 硬體與軟體NVIDIA 加速計算私有 AIGPU 私有云GPU 叢集邊緣 AI行業 AI
AI 服務
平臺現代化數字工程資料基礎與 AI自主運營AI 諮詢DevOps 自動化網路安全軟體開發智慧體構建MLOps 搭建GPU 租賃GPU 私有云
關於我們
公司我們的故事合作伙伴客戶案例
AI 資料中心
AI 資料中心GPU 私有云GPU 叢集GPU 租賃
AI 洞察Jev 系列基準測試實驗室模型選擇
技術說明
文件
Linux 作業系統UbuntuRed Hat Enterprise LinuxSUSE LinuxAlpine LinuxDebian、Rocky 與 AlmaLinux
AI 雲平臺AWS 與 Amazon BedrockGoogle Cloud 與 Vertex AIMicrosoft Azure 與 AI FoundryGPU 雲與算力市場
私有 AI 資料中心面向律師事務所面向會計師事務所本地部署平臺全面安全的雲
工具Ring PromoterJobshout
新聞
部落格
聯絡我們登入

面向現代企業的 AI 工作站、AI 多智慧體軟體、GPU 基礎設施和智慧代理解決方案。

聯絡我們

AI 解決方案

應用型 AIAI 智慧體生成式 AIAI 自動化AI 技能提升中小企業 AI 套餐企業 AI 實驗室行業 AI

產品

所有產品Workstation AI native CRM 與 ERP營銷OpenAI 智慧體Ring PromoterJobshout

公司

關於我們公司AI 資料中心為什麼選擇Workstation合作伙伴客戶案例定價聯絡TenthMatrix

資源

AI 洞察技術說明文件Linux 作業系統AI 雲平臺私有 AI 資料中心部落格搜尋網站地圖
英國總部
77-79 Marlowes, Hemel Hempstead HP1 1LF路線指引 — 從 M25 外環倫敦 20 號出口駛出公司編號: 11641870週一至週五:上午 9:00 - 下午 6:00 GMT
+44 7515 356 146
比利時辦公室
Workstation SRL, Rue Vanderkindere 34, 1180 Uccle, BrusselsBE 0751.518.683週一至週五:上午 9:00 - 下午 6:00 CET
+32 492 45 67 46
印度辦公室
#159 Sector 9, Pocket 1, DDA Flats, 110077 Dwarka, New Delhi
+91 98881 98841
台灣辦公室
3F, No. 2-4, Aisi Rd., Ren'ai Dist., Keelung City 200, Taiwan基隆市仁愛區愛四路2之4號3樓週一至週五:上午 9:00 - 下午 6:00 臺灣時間(GMT+8)
+886 926 892 701+32 492 45 67 46

© 2026 Workstation AI。保留所有權利。

隱私資料刪除Cookie服務條款網站地圖

NVIDIA、NVIDIA 標誌、RTX、RTX PRO、Blackwell、CUDA、NVLink、MGX、Omniverse、Spectrum-X、Quantum、BlueField、DGX 與 HGX 為 NVIDIA Corporation 在美國及其他國家的商標及/或註冊商標。AMD、AMD Arrow 標誌、EPYC、Instinct、Ryzen 與 Threadripper 為 Advanced Micro Devices, Inc. 的商標。Dell 與 PowerEdge 為 Dell Inc. 或其子公司的商標。Intel 與 Xeon 為 Intel Corporation 或其子公司的商標。AMAX、AceleMax 與 LiquidMax 為 AMAX Engineering Corporation 的商標。所有其他產品名稱、標誌與品牌均為其各自所有者之財產。我們使用它們僅為識別我們所描述、供應或支援的產品;使用這些名稱並不表示其所有者為我們背書。

首頁 / 技術說明 / Technology
DevOpsSecuritySREKubernetesCompliance

WSL Vault Technical Deep Dive: Zero-Trust Secrets, Dual Control, MFA & Multi-Region HA

Technical deep dive: envelope hierarchy, dual-control workflows, MFA, leases, active/active multi-region HA, feature tours, and presentation chapters

2026年10月7日Technology5 min read

Long-form technical companion to the compliance blog. How WSL Vault makes secret management a breeze for operators while enforcing cryptographic tenant isolation, dual-control shared access, MFA, dynamic leases, and active/active multi-region HA. Product: Workstation WSL Vault · wslvault.org · GitHub.

WSL Vault zero-trust tenant isolation

Presentation (~20 min) — youtu.be/O6q3_H0k2M8 · Silent tour: wslvault.org/#tour

Thesis. A secrets platform is only “easy” if the safe path is the fast path — and only trustworthy if one compromised credential cannot empty the building. WSL Vault combines Vault-compatible engines with a per-tenant envelope hierarchy, MFA-backed identity, real leases, and active/active regions so operators move quickly while blast radius stays one tenant wide.

1. Why shared root power fails audits

In many HashiCorp Vault deployments, a root token (or equivalent highly privileged unseal/admin path) remains the organisational backstop. That credential’s blast radius is the entire secrets graph: every namespace, every environment, every team’s API keys and database passwords. If it lands in the wrong hands, you do not have an incident ticket — you have a breach narrative.

WSL Vault’s answer is structural: identity is a signed, per-tenant token; material is sealed under a tenant KEK; cross-tenant requests fail closed at the crypto layer. Root custody, when needed for unseal, can be Shamir-split so no single operator holds the whole key.

Root token blast radius versus WSL Vault tenant isolation

2. Envelope encryption — the reason isolation is real

WSLVault envelope encryption key hierarchy

Each secret is sealed with a data encryption key (DEK). The DEK is wrapped under a tenant KEK. Tenant KEKs are wrapped under a root KEK held by the crypto service (KMS, HSM, or Shamir quorum). AES-256-GCM provides authenticated encryption. A database dump yields ciphertext wrapped under keys the database never sees.

That is why “steal the server” does not equal “steal the secrets” — and why Team A’s token cannot unwrap Team B’s DEKs even if an upper layer mislabels a path.

3. Tenants for apps, teams, and environments

Model the organisation the way auditors already think:

  • App tenants — Shop vs OpsAPI vs edge proxies each keep separate sealed material.
  • Team tenants — Platform, payments, data science never share unwrap rights by accident.
  • Environment tenants — INT / TEST / ACC / PROD as separate crypto worlds so a staging token cannot open production.

Isolation is not a WHERE clause. A cross-tenant read returns not found. That cryptographic refusal is what you show in a control description.

4. Ceremony workflow — oath and dual-tenant control

Dual-tenant ceremony and control gate

  1. Issue — At ceremony close, each tenant custodian receives a signed per-tenant token (and MFA enrolment).
  2. Oath — Custodians accept personal responsibility: never share the token, never store it in chat or git, rotate on departure.
  3. Day-to-day — Team A operates only inside Tenant A. Sprawl drops because the vault is the fastest place to put a secret.
  4. Wide audience / shared path — For secrets that must be obtained across tenants (joint release, shared break-glass, dual approval), policy requires both Tenant A and Tenant B tokens (plus MFA). Neither side alone succeeds.
  5. Audit — Both presentations and the stated purpose land on the per-tenant hash chain.

That dual-control gate is how you translate zero trust into an operating procedure regulators recognise.

5. Making secrets a breeze — engines and workflows

Feature tour (aligned with wslvault.org/#features):

Surface Operator experience Security property
KV v2Versioned put/get, check-and-set, soft-delete/destroySealed before disk under tenant hierarchy
TransitEncrypt / decrypt / sign / verify in one callPlaintext keys never leave the vault
PKIPrivate CA, roles, short-lived certsCA keys envelope-encrypted
LeasesIssue, renew, list, revokeRevoke stops use immediately
Identity + MFATOTP, recovery codes, OIDC/LDAP/mTLS/workloadSigned per-tenant token — never a trusted header alone
$ wslvault login --key wslv_…
✓ authenticated as tenant payments  region=manchester  mfa=ok

$ wslvault kv put prod/db/creds password=••••
✓ sealed with dek 01a0…93a7  version=1

$ wslvault kv get prod/db/creds --region london
# same secret, peer region decrypt
password = ••••   replication_lag=14ms

$ wslvault lease revoke lease_…
✓ credential dead — not a soft flag

MFA, lease, and multi-region HA workflow

6. MFA and identity detail

Humans enrol self-service TOTP with recovery codes. API keys and JWTs sit under fine-grained policies. Workloads authenticate with mTLS or cloud-workload patterns so apps never borrow a human root. Identity always resolves to a tenant-scoped principal before any unwrap.

That combination — MFA for people, short leases for machines — is what keeps “convenient” from becoming “copied into a wiki.”

7. Multi-region HA without a shared live database

WSLVault multi-region replication and audit

Each region is a complete stack: own PostgreSQL, own services, own public hostname. Regions share key material, not one hot database. Agents pull encrypted changes over the public edge. Clients read the same sealed secret from the peer with low replication lag. Failover is a PoP / traffic decision — operators do not rebuild custody under pressure.

Every operation joins a per-tenant signed hash chain. Records cannot be silently altered without breaking the chain — the evidence pack auditors ask for.

8. Deploy path — Kubernetes, Helm, GitOps

Ship with the Helm chart and operator; scrape Prometheus; drive from Argo CD / GitOps. Clients: CLI plus Go, Python, Rust, TypeScript SDKs, and an MCP server for agent tooling. Getting started: GETTING-STARTED.md.

9. Guided tours (videos and site)

Tour Where What you see
Silent console (~2 min)wslvault.org/#tourTenants, KV, transit, MFA, regions, audit
Features#featuresEngines behind one API
Architecture#architectureActive/active region mesh + CLI demo
Security model#securityAudit, Shamir, tenant isolation, no plaintext
Full presentationyoutu.be/O6q3_H0k2M8~20 min deep walkthrough

Presentation chapters (YouTube)

Time Topic
0:00Why this matters (steal the server)
1:30What WSLVault is
3:00Envelope encryption and key hierarchy
5:30Multi-tenancy that actually isolates
8:00Console walkthrough (secrets, transit, PKI)
11:00Identity, policies, leases and MFA
14:00Regions, HA and audit
17:00Deploy on Kubernetes
19:00Where to go next

10. Related Workstation content

  • Compliance blog (this article’s companion)
  • Intro blog · Intro technical brief
  • AI agents blog · AI agents article
  • Product page · wslvault.org · GitHub

Published by Workstation — secrets custody you can explain to a regulator and still enjoy as an engineer.

分享這篇文章

更多 Technology 內容

AMD Instinct vs NVIDIA for AI Infrastructure: Evaluate, Port and Run Both (2026)

AMD Instinct vs NVIDIA for AI Infrastructure: Evaluate, Port and Run Both (2026)

From Lisa Su's 2024 interview to Helios: the 2026 platforms layer by layer, Instinct product lines, memory arithmetic, the CUDA-to-ROCm software map, a porting workflow with commands, dual-vendor Kubernetes, fair benchmarking, a decision workflow and day-two operations

閱讀更多
Private AI for Financial Data: Residency, Routing and Cost on AWS Bedrock, Microsoft Azure and Your Own GPUs

Private AI for Financial Data: Residency, Routing and Cost on AWS Bedrock, Microsoft Azure and Your Own GPUs

Billing models and idle cost, per-call cost at three volumes, Bedrock in-region vs EU profiles, Converse API with prompt caching and IAM region enforcement, Azure deployment types and Claude on Foundry, UK and EU providers, a residency-aware provider registry, cost-cutting workflow, a golden-set evaluation harness and GPU break-even

閱讀更多
NVKE and NVKR Technote: Building a Managed GPU Kubernetes Service from Bare Metal, with Reproducible, Verifiable Builds

NVKE and NVKR Technote: Building a Managed GPU Kubernetes Service from Bare Metal, with Reproducible, Verifiable Builds

Architecture, components and pinned versions, the NVKR Definition-Run-Release-Verify lifecycle, the full nvkectl workflow from racked servers to a customer cluster, node attestation, day-two operations, every preview limitation with a mitigation, a comparison with other approaches and a staged evaluation plan

閱讀更多