What is OSCP pen test

Messages

[Update] Our preliminary information on the new test from Stiftung Warentest

Created on September 26, 2016, 5:30 p.m. | Category: blog

Update: We now know that we emerged as one of two test winners from the Stiftung Warentest test. Even though some of our encryption functions and security technologies that we support have not been tested. After all, our crypto mail storage was apparently considered retrospectively and mentioned in the article. Nevertheless, we are pleased with the test victory.
Regardless of our offer, we are very relieved that Stiftung Warentest has decided not to explicitly prefer compromised PGP solutions. This became apparent during the test phase, which we monitored critically, and was also documented in the documents.

Dear Posteo customers,

As a precautionary measure, we want to inform you in advance about a new test by Stiftung Warentest, which will take place on September 28th. will appear.

You may be familiar with the previous history: In the last test in January 2015, Posteo was one of two test winners. However, the test contained numerous errors and misrepresentations - including about Posteo, so that Stiftung Warentest stopped selling its magazine and published almost a complete page of corrections in the following magazine.

After the problems with the last test, we followed the test phase critically this time. The test leader is the same.
As a tested provider, we do not yet know how the comparison test will turn out this time.

But what we already know:
Numerous encryption options and security technologies that you benefit from at Posteo are apparently not part of the e-mail provider test. As a precautionary measure, we will draw your attention to this in advance. Because consumers generally assume that the most important features of the respective products are included in a rating in an independent comparison.

The following functions were apparently not taken into account in the grading:

- the TLS shipping guarantee
- the Posteo crypto mail storage
- the Posteo key directory
- the worldwide automatic key search
- the input encryption with S / MIME
- all other offers related to S / MIME

This emerges from the final test program and the supplier advance information from 08/30/2016 as well as from the test supplements submitted by Stiftung Warentest from 09/16/2016.

There has also been no assessment of important quality features of HTTPS encryption between the browser and the provider. Functions such as HSTS, OCSP, Certificate Transparency or HPKP, which we support, are important protective measures to strengthen browser access security.

We criticized the widespread neglect of these functions and technologies - but they were not subsequently included in the research program.

Encrypted transport routes only subsequently included in the test

Even one of the most important technologies was almost not taken into account in the test: TLS encryption of the e-mail transport routes between the providers.
We had criticized that even this most important everyday encryption should not be checked. A high-quality transport route encryption between the providers is very important for the consumer in everyday communication.
The Stiftung Warentest had only tested how the providers encrypt to the user with TLS. We have referred the testers to a new technical guideline from the BSI (Federal Office for Information Security) on secure e-mail transport. On September 12, 2016, the test program was subsequently expanded accordingly: STARTTLS between the providers and support for the DANE security technology are now also part of the test.



Without wanting to evaluate this, we would also like to point out in advance that sustainability concepts, social management / commitment (corporate social / environmental responsibility) were also not examined.

Error in the provider advance information

We have not received any feedback on the content of a gross measurement or reporting error from Stiftung Warentest and therefore do not know whether this was rectified before publication. In the pre-provider information, the test result incorrectly states that Posteo would not offer PGP in the webmailer.



Therefore we state as a precaution: This is wrong.
That's right: Posteo has been offering PGP in webmail for two years. Now even with its own key directory with a worldwide, automatic key search, which also does not appear in the test documents.

Furthermore, it is surprisingly noted in the provider advance information that our data protection officer cannot be found on the website. Here we state: This is also wrong.
It is correct: In our data protection declaration as well as in the imprint, we provide the contact information of our data protection officer so that they can be easily found.

We ask for your understanding that we will point out these errors prior to publication due to the lack of specific feedback from the testers. In any case, we want to prevent the spread of false facts. The test magazine of Stiftung Warentest alone has a circulation of more than 400,000 copies - and test content is distributed more widely on the day of publication via the news agencies.

Will Stiftung Warentest recommend proprietary encryption solutions that can be compromised?

The Stiftung Warentest will evaluate offers of the providers around the end-to-end encryption PGP.
In this area, among other things due to the supplier advance information sent to us, we have to expect that Stiftung Warentest will represent a position that is technically highly controversial.

There are indications that the solutions that all leading security experts advise against will receive the best rating: namely PGP solutions in which the private key of the user (which decrypts the content) is exposed to possible access by the provider. Many sides warn against such solutions - and we, too, strictly reject them.

A new study by the BSI (Federal Office for Information Security), which was only published in August, warns:
“With end-to-end encryption, it is important that the encryption software is under the control of the user and has therefore been installed locally by the user. This means that the encryption software must not be delivered through the web application. " (Page 9) And elsewhere it says: “A browser extension must ensure that the web application is not able to access the plain text of the data to be encrypted. This is only possible if all critical steps are taken through the extension and not through the web application. This primarily includes the cryptographic functions. " (Page 9)

Solutions with keys in the provider's sphere of influence compromise end-to-end encryption - making it worthless per se. The user has no control over them. Instead, he has to trust the implementation of his provider. In its study, the BSI emphasizes that “with end-to-end encryption, however, it is important that the provider does not have to be trusted”. (Page 11)
And the trade journal c’t also gives a clear verdict on such procedures: “The alternative mail service […] breaks with the end-to-end principle by managing the private keys of its users on the server side. The integrity of communication cannot be guaranteed in this way, as the provider clearly sacrifices too much security for a little more comfort. " (Source: c’t 2015, issue 13, page 139)

We share the technical assessment of the BSI and the leading security experts - and we have implemented PGP accordingly in our webmail. We sent the new BSI study to Stiftung Warentest in August. The developer of GnuPG, Werner Koch, also worked on it.
In the pre-provider information from 08/30/2016, the solution recommended by the experts is apparently referred to as a solution “only with plug-in”. It was awarded 0.5 points there. PGP solutions that are integrated directly into the web interface of the provider, on the other hand, are given 1 point. The lack of PGP in the web interface is marked 0.

Governments want to be able to decipher messages

We will only finally find out on Wednesday whether the Stiftung Warentest actually maintains this assessment in its publication. If so, we consider it dangerous: If the key is in the provider's sphere of influence (private key on the provider's servers / in smartphone or web apps), the private key can be compromised. There is also a political level here: At the moment, end-to-end encryption is still a valid way of protecting yourself against unauthorized readers. Current plans of European governments show, however, that providers could soon be legally obliged to hand over keys, to decrypt messages or even to build a back door for the state. This can then also be misused by unauthorized persons, such as criminals and secret services.
If Stiftung Warentest is currently in favor of the practice of private keys on the provider's servers, for possibly “comfort” reasons, it is promoting precisely this political development. If providers are devalued by Stiftung Warentest because of real end-to-end encryption, some may decide to offer compromised solutions until the next test in order not to be rated worse. We won't do this at Posteo.

We also have criticism of a number of other points of the supplier's advance information. But here we have to wait for the publication.
If you have any questions, you can always contact us by email.

Best wishes
Your Posteo team