Some checks are pending
ci / markdown (push) Waiting to run
Open with an executive summary that HIPAA, SOC 2, and ISO 27001 are organizational programs. Verae DataCubes supply store, communicate, timestamp, verify, and audit tools for the technical portion only. Chapters cover transit (HPKE, visible routing), rest (IPFS/Peergos hash-verified restore), receipts, EU Peergos evaluations (Cure53 2019, ROS 2024), cross-blockchain timestamping, and write-once Iceberg archive. PDF is branded with the Verae logo top-left and Verae Inc contact in the footer; last chapters are sourced bios for Garfinkel (FINRA CRD 5052743), Haber, and Lambert.
140 lines
5.3 KiB
ReStructuredText
140 lines
5.3 KiB
ReStructuredText
Peergos security evaluations in Europe
|
|
======================================
|
|
|
|
What was evaluated
|
|
------------------
|
|
|
|
The offline storage and replication system used with Verae
|
|
DataCubes is **Peergos**: an encrypted, peer-to-peer filesystem
|
|
whose blocks live on IPFS. Peergos was designed as a
|
|
trust-minimized store --- the server is treated as an adversary
|
|
for content and for most metadata --- and that design was
|
|
submitted to independent European security firms **twice**.
|
|
|
|
Those engagements are **security evaluations of the Peergos
|
|
protocol and implementation**. They are pentests, source-code
|
|
audits, and (in 2019) a cryptographic and design review. They
|
|
are **not**:
|
|
|
|
* a HIPAA certification of Peergos, Verae, or any customer;
|
|
* a SOC 2 Type I or Type II report;
|
|
* an ISO 27001 certificate of an ISMS;
|
|
* a government "certified filesystem" designation.
|
|
|
|
Both firms are **EU-based**. Both full reports are **public**.
|
|
That combination --- independent, European, public, repeat ---
|
|
is unusual and is worth attaching to a vendor-assurance file,
|
|
**labeled correctly**.
|
|
|
|
2019 --- Cure53, Berlin, Germany
|
|
--------------------------------
|
|
|
|
* **Firm:** Cure53
|
|
* **Location:** Berlin, Germany
|
|
* **When:** May--June 2019
|
|
* **Scope:** pentest, source-code audit, and
|
|
**crypto / design review**
|
|
* **Outcome:** no fundamental architectural or cryptographic
|
|
problems were identified. Issues that were identified were
|
|
fixed. Cure53 stated that the platform **passed this
|
|
evaluation**.
|
|
|
|
A crypto/design review is a stronger statement than a
|
|
black-box pentest of a web form. It is an expert reading of
|
|
whether the cryptree, the chunk encryption, the identity
|
|
model, and the threat model hang together. "Passed" in
|
|
Cure53's language is not an ISO mark. It is a specialist
|
|
firm saying: we looked at the cryptography and the
|
|
architecture, we did not find a fatal flaw, and the issues we
|
|
did find were addressed.
|
|
|
|
Primary sources (public):
|
|
|
|
* https://peergos.org/posts/security-audit
|
|
* https://cure53.de/pentest-report_peergos.pdf
|
|
* https://github.com/Peergos/Peergos/tree/master/audits
|
|
|
|
2024 --- Radically Open Security B.V., Amsterdam
|
|
------------------------------------------------
|
|
|
|
* **Firm:** Radically Open Security B.V.
|
|
* **Location:** Amsterdam, Netherlands
|
|
* **When:** September--November 2024
|
|
* **Scope:** crystal-box pentest plus code audit of the
|
|
Peergos **web UI**
|
|
* **Funding context:** the 2024 audit post states support
|
|
from **NLnet** (Netherlands) and refers to European Union
|
|
Horizon 2020 **NGI-POINTER**, grant **871528**.
|
|
* **Outcome:** **0** findings rated extreme, high, or
|
|
elevated; **2** moderate; **6** low. Peergos states that
|
|
all were fixed. There was **no data exposure** and **no
|
|
integrity compromise**. The issues were described as
|
|
mostly UI crashes.
|
|
|
|
A crystal-box (clear-box) engagement means the testers had
|
|
source. A web-UI scope is narrower than the 2019
|
|
crypto/design review: it is evidence about the interface
|
|
that humans actually click, which is where many "encrypted
|
|
backend" products fail in practice. Zero high-severity
|
|
findings, no data exposure, no integrity compromise, and a
|
|
public report are the facts. They are good facts. They are
|
|
still not a customer's Type II.
|
|
|
|
Primary sources (public):
|
|
|
|
* https://peergos.org/posts/security-audit-2024
|
|
* https://github.com/Peergos/Peergos/tree/master/audits
|
|
|
|
How to present these reports to an auditor
|
|
------------------------------------------
|
|
|
|
Correct:
|
|
|
|
"Our at-rest layer is Peergos. Peergos was independently
|
|
evaluated in Berlin in 2019 (crypto and design) and in
|
|
Amsterdam in 2024 (web UI, crystal box). Both reports are
|
|
public. We attach them as **component security
|
|
evaluations**. They are not our SOC 2, not our ISO 27001,
|
|
and not a HIPAA certification. Our own controls, our own
|
|
period of examination, and our own auditor are separate."
|
|
|
|
Incorrect:
|
|
|
|
"We are HIPAA certified because Peergos was audited in
|
|
Europe."
|
|
|
|
"Peergos is ISO 27001."
|
|
|
|
"The EU certified this filesystem."
|
|
|
|
EU funding is not a certification. NGI-POINTER grant 871528
|
|
is a research-and-innovation funding fact. It is worth
|
|
listing under "provenance." It is not a registrar's mark.
|
|
|
|
Hosted Peergos versus self-hosted DataCubes
|
|
-------------------------------------------
|
|
|
|
Peergos's hosted privacy notice has stated that peergos.net
|
|
uses servers in **Germany**. A **self-hosted** organizational
|
|
DataCube is a **different processing location**. The
|
|
customer's Record of Processing, BAA pack, and ISO scope
|
|
must name *that* location --- the customer's ns1, region, or
|
|
chosen host --- not peergos.net's Germany, unless the
|
|
customer actually uses peergos.net.
|
|
|
|
The evaluations still apply to the **protocol and
|
|
implementation**. Location of processing is an
|
|
organizational fact on top.
|
|
|
|
What "designed under funding from Cure53 / ROS" is not
|
|
------------------------------------------------------
|
|
|
|
The 2019 Cure53 work and the 2024 ROS work are **evaluations**
|
|
of a system that was designed by the Peergos authors. They
|
|
are not a claim that Cure53 or Radically Open Security
|
|
designed Peergos. The accurate statement is: the storage and
|
|
replication system was **independently audited twice in
|
|
Europe**, by Cure53 in Berlin (2019) and by Radically Open
|
|
Security in Amsterdam (2024), with public reports, and the
|
|
2019 work included a cryptographic and design review of the
|
|
architecture that Verae DataCubes rely on for data at rest.
|