BSI Status of quantum computer development - Entwicklungsstand Quantencomputer

2,745 views
Skip to first unread message

John Mattsson

unread,
Nov 14, 2023, 7:25:45 AM11/14/23
to pqc-...@list.nist.gov

Hi,

German
Bundesamt für Sicherheit in der Informationstechnik (BSI) recently published an update to the excellent report “Status of quantum computer development - Entwicklungsstand Quantencomputer”. I think this is very likely the best and most up to date overview of quantum computer and quantum algorithm development focusing on cryptography. A lot of details but not so many conclusions expect that development of any CRQC will take at least one decade and more likely two.


https://www.bsi.bund.de/SharedDocs/Downloads/DE/BSI/Publikationen/Studien/Quantencomputer/Entwicklungstand_QC_V_2_0.pdf

 

Two decades is not a lot of time and BSI encourages rapid migration to post-quantum cryptography: ”From the BSI's point of view, the question of "if" or "when" there will be quantum computers is no longer paramount. First post-quantum algorithms have been selected by NIST for standardisation and post-quantum cryptography will be used by default. Therefore, the migration to post-quantum cryptography should be pushed forward.

 

https://www.bsi.bund.de/EN/Themen/Unternehmen-und-Organisationen/Informationen-und-Empfehlungen/Quantentechnologien-und-Post-Quanten-Kryptografie/quantentechnologien-und-post-quanten-kryptografie_node.html

 

Cheers,

John Preuß Mattsson

Wrenna Robson

unread,
Nov 14, 2023, 7:37:14 AM11/14/23
to John Mattsson, pqc-forum
Am I correct in saying that BSI still recommend Classic McEliece and FrodoKEM for post-quantum key agreement? Is that intended to complement or contradict NIST's selection of standards? I've been a little confused in the past as to where things stand here and I would welcome a clarification.

Best,

Wrenna

--
You received this message because you are subscribed to the Google Groups "pqc-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pqc-forum+...@list.nist.gov.
To view this discussion on the web visit https://groups.google.com/a/list.nist.gov/d/msgid/pqc-forum/GVXPR07MB967848982DB9EAE80CB71C6689B2A%40GVXPR07MB9678.eurprd07.prod.outlook.com.

John Mattsson

unread,
Nov 14, 2023, 8:11:09 AM11/14/23
to Wrenna Robson, pqc-forum

Hi Wrenna,

 

Yes, BSI started recommending FrodoKEM and Classic McEliece in 2020 and that is still the recommendation:

 

“Recommended Mechanisms: The key exchange mechanisms FrodoKEM-976 and FrodoKEM-1344 ([5, Section 2.5]) as well as Classic McEliece with the parameters mceliece460896, mceliece6688128 and mceliece8192128 as well as their corresponding variants mceliece460896f, mceliece6688128f and mceliece8192128f [3, Section 7] are assessed to be cryptographically suitable to protect confidential information on a long-term basis at the security level aimed at in this Technical Guideline. This is a very conservative assessment that includes a significant margin of security with respect to future cryptanalytic advances. It is possible that in future revisions of this guideline other parameter choices and PQC mechanisms may also be deemed technically suitable.

 

FrodoKEM will not be standardised as part of NIST’s PQC project. This is mainly due to considerations of the efficiency of the mechanism, there are currently no doubts about its security [2]. Classic McEliece was included in the fourth round of the NIST project and could possibly be standardised at the end of the project. The BSI therefore maintains the recommendation of FrodoKEM and Classic McEliece as PQC mechanisms with a high security margin against future attacks. More details can be found in the BSI-guide “Quantum-safe cryptography” [37].

 

In Chapter 6, the hash-based signature mechanisms XMSS and LMS as well as their multi-tree variants, which are considered quantum computer-resistant according to current knowledge, are recommended.

 

At this time, no further post-quantum mechanisms are recommended in this Technical Guideline. About a possible adoption of the mechanisms selected for standardisation by NIST in July 2022 (see [2]) into the Technical Guideline will only be decided after publication of the draft standards.”

 

https://www.bsi.bund.de/SharedDocs/Downloads/EN/BSI/Publications/TechGuidelines/TG02102/BSI-TR-02102-1.pdf

 

As stated in the January 2023 edition of BSI TR-02102-1, BSI will make a decision on possible adoption on the mechanism standardized by NIST. I don’t know what that decision will be, but I hope that BSI aligns with France, The Netherlands, and the UK and recommend ML-KEM and ML-DSA. The Netherland also recommended FrodoKEM and Classic McEliece in 2022 but are now recommending ML-KEM and ML-DSA (FrodoKEM and Classic McEliece are no longer recommended). I hope that with open-access NIST algorithms available, BSI stops recommending non-standardized and paywalled algorithms, which clearly are a cybersecurity risk. If any European country would like to continue recommending FrodoKEM they should drive for publication as an RFC.

 

Cheers,

John Preuß Mattsson

Simon Hoerder

unread,
Nov 14, 2023, 8:11:30 AM11/14/23
to pqc-forum
Hi Wrenna,

At the recent PKI Consortium conference on PQC the BSI representative said that an update to TR-02102 is expected for 2024/2025 when the NIST standards are published in their final form. FrodoKEM and Classic McEliece will remain in TR-02102, ML-KEM, ML-DSA and SLH-DSA are expected additions at levels 3 and 5. No mention of FN-DSA but that doesn’t even have a draft version from NIST yet.

The BSI representative also emphasized that TR-02102 should be read as a whitelist. Those algorithms are definitely ok, others can be ok if sufficiently justified.

Finally, not from BSI in particular but from the conference in general there were reliable reports of ISO working on standardizing FrodoKEM and Classic McEliece in addition to the NIST standards. That’s happening behind what I’d personally call closed doors but it’s the org where most governments maintain their crypto standards. I expect that BSI will start referencing ISO documents for FrodoKEM and Classic McEliece once they become available (at the cost of the usual ISO document fees). N.B. I have no first hand information on what’s happening at ISO. I can only report what I hear from others at conferences.

Best,
Simon

On 14 Nov 2023, at 13:37, Wrenna Robson <wren....@gmail.com> wrote:



Wrenna Robson

unread,
Nov 14, 2023, 8:20:06 AM11/14/23
to John Mattsson, pqc-forum
Which one of the two of those are paywalled?

Anyway, thanks for the clarification. Clearly in some sense it would be good if everything were singing from the same hymn sheet (though obviously I think CMcE is still under consideration by NIST in Round 4?)

Best,

Wrenna

Wrenna Robson

unread,
Nov 14, 2023, 8:22:46 AM11/14/23
to Simon Hoerder, pqc-forum
I have also reliably received these reports. The process by which ISO standardisation works seems rather roundabout though - to say the least.

Best,

Wrenna

John Mattsson

unread,
Nov 14, 2023, 8:36:25 AM11/14/23
to Wrenna Robson, pqc-forum

Hi Wrenna,

All current ISO standards are paywalled and as stated several times on this forum, ISO is planning to standardize Classic McEliece and FrodoKEM. Unless ISO realizes that paywalled security standards are a security risk I assume future ISO standards of Classic McEliece and FrodoKEM will also be paywalled.

Yes, NIST is still considering standardizing Classic McEliece and an internet draft (draft-josefsson-mceliece-00) was recently published.

FrodoKEM and McEliece specifications are available from the NIST PQC project website but deploying non-standardized cryptography is typically not a good idea.

Wrenna Robson

unread,
Nov 14, 2023, 8:40:25 AM11/14/23
to John Mattsson, pqc-forum
If we end up in a world where people end up standardising multiple different versions of the same cryptoscheme I am going to have to have a little scream into a pillow. But thank you very much for the clarification - I get what you mean now. My hope is that ISO just follow along with whatever is happening in public.

Best,

Wrenna

Kris Kwiatkowski

unread,
Nov 14, 2023, 8:52:05 AM11/14/23
to pqc-...@list.nist.gov

IMHO, such fragmentation may also have negative impact on overall adoption of PQ schemes.

Regarding FrodoKEM, indeed standardization in ISO is done behind closed doors, but
it is worth to note that the frodokem.org has a new release of the algorithm spec [1].
The page says it is a "FrodoKEM Preliminary Standardization Proposal (submitted to ISO)".
It gets some idea what's going on behind those "closed doors".
I've looked at it some time ago and it seems very similar to FrodoKEM from NIST PQC Round3,
but it does have some additional parametrisation.

[1] https://frodokem.org/files/FrodoKEM-standard_proposal-20230314.pdf

Best,
---

Kris Kwiatkowski
Staff Cryptography Architect
PQShield, LTD

Ira McDonald

unread,
Nov 14, 2023, 9:26:34 AM11/14/23
to Wrenna Robson, Ira McDonald, pqc-...@list.nist.gov
Hi Wrenna,

And here's the Classic McEliece proposal that was submitted to the ISO 18033 update 
project this spring:

BTW - Relative to the question "Hybrids - when and how long?" - on the Global Platform
Cybersecurity Vehicle Forum this morning (Hamburg, Germany), Dr Julian Brough (BSI)
answered that BSI will continue to recommend hybrid Classical/PQC constructions for
the long term.

Cheers,
- Ira

Ira McDonald (Musician / Software Architect)
Chair - SAE Trust Anchors and Authentication TF
Co-Chair - TCG Trusted Mobility Solutions WG
Co-Chair - TCG Metadata Access Protocol SG
Chair - Linux Foundation Open Printing WG
Secretary - IEEE-ISTO Printer Working Group
Co-Chair - IEEE-ISTO PWG Internet Printing Protocol WG
IETF Designated Expert - IPP & Printer MIB
Blue Roof Music / High North Inc
http://sites.google.com/site/blueroofmusic
http://sites.google.com/site/highnorthinc
mailto: bluero...@gmail.com
(permanent) PO Box 221  Grand Marais, MI 49839  906-494-2434


Anjan Roy

unread,
Nov 14, 2023, 11:33:02 AM11/14/23
to Kris Kwiatkowski, pqc-...@list.nist.gov
Dear all,

When I was implementing FrodoKEM as a header-only C++20 library I found that an older specification version (2021-06-04) describes algorithms in more detailed way (particularly helpful for implementers like me) compared to latest ISO document (2023-03-14). ISO submission document also specifies new instantiations for using FrodoKEM in ephemeral settings, as already mentioned by Kris. These are the two major changes I noticed.

My C++ implementation of FrodoKEM is conformant to the latest ISO submission document. In case you're interested, please find the repository @ https://github.com/itzmeanjan/frodokem. I ensure conformance to the specification using Known Answer Tests, which I generate by following (reproducible) steps described @ https://gist.github.com/itzmeanjan/38d506a69073bdeb0933245401f42186.

Cheers,
Anjan Roy

Constantinos Valakas

unread,
Nov 15, 2023, 5:25:24 AM11/15/23
to Anjan Roy, Kris Kwiatkowski, pqc-...@list.nist.gov

Dear John & all,

 

Per your feedback:

 

>All current ISO standards are paywalled and as stated several times on this forum, ISO is planning to standardize Classic McEliece and FrodoKEM. Unless ISO realizes that paywalled >security standards are a security risk I assume future ISO standards of Classic McEliece and FrodoKEM will also be paywalled.

 

Do you know by any chance if ISO has already registered any draft versions on McEliece and FrodoKEM? And if there are the corresponding website links?

 

For example, for QKD the corresponding link (this is an official standard though): https://www.iso.org/standard/77097.html

 

Thank you in advance,

Costas.

 

Stanchion Payments | Senior Solutions Consultant

 

From: pqc-...@list.nist.gov <pqc-...@list.nist.gov> On Behalf Of Anjan Roy
Sent: Tuesday, November 14, 2023 4:32 PM
To: Kris Kwiatkowski <kr...@amongbytes.com>
Cc: pqc-...@list.nist.gov
Subject: Re: [pqc-forum] BSI Status of quantum computer development - Entwicklungsstand Quantencomputer

 

You don't often get email from anjan...@gmail.com. Learn why this is important

CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.

 

Notice: This communication contains information that may be privileged, confidential and/or exempt from disclosure under applicable law. If you are not the intended recipient, you are hereby notified that any disclosure, copying, distribution, or use of the information contained herein (including any reliance thereon) is strictly prohibited. If you received this transmission in error, please immediately contact the sender and delete this communication.

John Mattsson

unread,
Nov 15, 2023, 7:07:35 AM11/15/23
to Constantinos Valakas, Anjan Roy, Kris Kwiatkowski, pqc-...@list.nist.gov

Dear Costas,

As discussed in this forum there are ISO draft versions of McEliece and FrodoKEM. I am quite sure there is no public information. ISO is a very secret organisation out of touch with the modern world. People engaging in ISO are even forbidden to discuss anything in public. I think it would be beneficial for global cybersecurity and democracy if ISO stopped producing paywalled security standards.

 

According to the FrodoKEM and Classic McEliece teams the following documents were submitted to ISO:

 

https://classic.mceliece.org/iso-mceliece-20230419.pdf

https://frodokem.org/files/FrodoKEM-standard_proposal-20230314.pdf

 

>For example, for QKD the corresponding link

 

QKD is completely useless for any non-military use cases. Not only is QKD very expensive and inflexible, practical implementations are not secure at all. Trusted relays are the complete opposite of current best practice for how to build modern infrastructure. If the availability and side-channel problems are solved in the future it could maybe make sense to use point-to-point QKD as a defense-in-depth measure in militarty applications where cost is not relevant. QKD can theoretically be used together with a pre-shared key to get information theoretical security. Alternatively, you could use QKD, a pre-shared key, and a symmetric cipher to create a system that do not rely on asymmetric cryptography but still provides PFS. But even in military use cases, I think a hybrid system using KEMs based on different mathematical problems (ECC + ML-KEM + Classic McEliece + CSIDH) would be a better, cheaper, and practically more secure choice.

 

(My current understanding is that the availability and side-channel issues are fundamentally tied to QKD and will not be solved).

 

Cheers,

John Preuß Mattsson

 

Constantinos Valakas

unread,
Nov 15, 2023, 7:25:18 AM11/15/23
to John Mattsson, pqc-...@list.nist.gov

Thank you very much, John for the feedback provided, also in QKD.

 

I totally agree that the way ISO works does not help much with all this secrecy. I hope that this will change in the future.

 

Regarding QKD, although off-topic, as it stands now is not a practical solution for commercial use, indeed.

 

Best regards,

Costas.

Blumenthal, Uri - 0553 - MITLL

unread,
Nov 15, 2023, 8:15:31 AM11/15/23
to Constantinos Valakas, John Mattsson, pqc-...@list.nist.gov
Somewhat off-topic: I think currently OKD is useless especially for the military, exactly because of its  practical implementation security concerns. 

Regards,
Uri

On Nov 15, 2023, at 07:26, Constantinos Valakas <Constantin...@stanchionpayments.com> wrote:


Thank you very much, John for the feedback provided, also in QKD. I totally agree that the way ISO works does not help much with all this secrecy. I hope that this will change in the future. Regarding QKD, although off-topic, as it stands now
ZjQcmQRYFpfptBannerStart
This Message Is From an External Sender
This message came from outside the Laboratory.
 
ZjQcmQRYFpfptBannerEnd

Jarek Duda

unread,
Nov 15, 2023, 9:27:04 PM11/15/23
to pqc-forum, John Mattsson
Two decades to break the current cryptography assuming there is a stable evolution of current quantum computing technologies, without some large breakthroughs - but what if there are some surprises behind a corner?

Is there maybe a list of considered hypothetical breakthroughs which might shorten this time?

One I am aware of are two-way quantum computers (2WQC) enhancement: we influence standard electronics from both directions - pushing electrons into a chip, and simultaneously pulling from for better flow control.
If reaching such two-way control for any quantum computing technology, in theory they could attack NP problems (postBQP with postselection as physical constraints).
Below is proposed for photonic QC using laser impulse as state preparation. Using ring laser for this purpose and enclosing into a loop like below, from perspective of CPT symmetry photon trajectories are reversed - doing exactly as state preparation, but to the original final state.

2WQC.jpg

Bas Westerbaan

unread,
Nov 17, 2023, 7:58:53 AM11/17/23
to Jarek Duda, pqc-forum, John Mattsson
Indeed, development could stall or speed up.

But these two-way quantum computers are most likely not real. [1]

Best,

 Bas



--
You received this message because you are subscribed to the Google Groups "pqc-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pqc-forum+...@list.nist.gov.

Jarek Duda

unread,
Nov 17, 2023, 8:41:33 AM11/17/23
to pqc-forum, Bas Westerbaan, pqc-forum, Jarek Duda
> But these two-way quantum computers are most likely not real. [1]

As there is state preparation allowing to modify the initial state, and most physicists believe in CPT symmetry, why shouldn't it be possible to do "the same but reversed" (CPT analog) to influence the final state?
For example: pull/push, negative/positive pressure, stimulated emission absorption ... also works for EM waves/light:

In theory we could also use superfluid, for which there are considered mechanical qubits (e.g. https://www.nature.com/articles/s41534-021-00393-3 ) ... and building two-way quantum computer for such microfluidic chip, would be just connecting it to a pump.
Also e.g. (currently hypothetical) topological QC are very time-symmetric, straightforward to do two-way.
Non-intuitive does not mean impossible.

Anyway, excluding " likely not real" possibilities while designing cryptography does not seem very responsible - it might be safer to try to make it resistant also to currently less standard approaches.



On Friday, November 17, 2023 at 1:58:53 PM UTC+1 Bas Westerbaan wrote:
Indeed, development could stall or speed up.

But these two-way quantum computers are most likely not real. [1]

Best,

 Bas



On Thu, Nov 16, 2023 at 3:27 AM Jarek Duda <dud...@gmail.com> wrote:
Two decades to break the current cryptography assuming there is a stable evolution of current quantum computing technologies, without some large breakthroughs - but what if there are some surprises behind a corner?

Is there maybe a list of considered hypothetical breakthroughs which might shorten this time?

One I am aware of are two-way quantum computers (2WQC) enhancement: we influence standard electronics from both directions - pushing electrons into a chip, and simultaneously pulling from for better flow control.
If reaching such two-way control for any quantum computing technology, in theory they could attack NP problems (postBQP with postselection as physical constraints).
Below is proposed for photonic QC using laser impulse as state preparation. Using ring laser for this purpose and enclosing into a loop like below, from perspective of CPT symmetry photon trajectories are reversed - doing exactly as state preparation, but to the original final state.
 
On Tuesday, November 14, 2023 at 1:25:45 PM UTC+1 John Mattsson wrote:

Hi,

German
Bundesamt für Sicherheit in der Informationstechnik (BSI) recently published an update to the excellent report “Status of quantum computer development - Entwicklungsstand Quantencomputer”. I think this is very likely the best and most up to date overview of quantum computer and quantum algorithm development focusing on cryptography. A lot of details but not so many conclusions expect that development of any CRQC will take at least one decade and more likely two.


https://www.bsi.bund.de/SharedDocs/Downloads/DE/BSI/Publikationen/Studien/Quantencomputer/Entwicklungstand_QC_V_2_0.pdf

 

Two decades is not a lot of time and BSI encourages rapid migration to post-quantum cryptography: ”From the BSI's point of view, the question of "if" or "when" there will be quantum computers is no longer paramount. First post-quantum algorithms have been selected by NIST for standardisation and post-quantum cryptography will be used by default. Therefore, the migration to post-quantum cryptography should be pushed forward.

 

https://www.bsi.bund.de/EN/Themen/Unternehmen-und-Organisationen/Informationen-und-Empfehlungen/Quantentechnologien-und-Post-Quanten-Kryptografie/quantentechnologien-und-post-quanten-kryptografie_node.html

 

Cheers,

John Preuß Mattsson

Reply all
Reply to author
Forward
0 new messages