Everyone in the IT world is wondering how big a threat quantum computers pose to cryptography and how to deal with the problem. This series of articles tries to explain the problem in a popular way. After getting acquainted with the available technologies, it is possible to try to estimate the capabilities of quantum computers around 2035.
Based on this information, it is possible to create the following overview, which is again an estimate for the year 2035. To attack the following algorithms, a quantum computer, like a digital computer, needs a specific time, and in the case of a quantum computer, a minimum number of passes. This is followed by post-processing on a conventional digital computer. In addition to the specified time, it also requires a certain minimum number of qbits, but there is constant development in this area, so at the time of writing the article the result is slightly overestimated. For these requirements, the most significant input is probably the cooling power. Here is a modified table that takes into account the power consumption of individual qbits and gates in the range of pW, thermal radiation of the cabling of 1mW/m, and reading the state from one qbit of 100mW. The system mentioned assumes high-quality insulation and the need for cooling to a temperature of 1 mK , and uses a steel supporting structure with mechanical parameters for the weight of the device corresponding to a steel ball of a given volume. The need for internal cooling for heat dissipation is deliberately neglected, this is only a theoretical model. The necessary construction for heat dissipation would extremely change the conditions of the entire approximate calculation. The resulting energy stated is needed for the implementation of one calculation, not all repetitions to ensure an adequate level of probability. At the same time, it does not include post-processing, which, on the other hand, does not have significant power requirements.
| Algorithm | Logical qubits | Number of gates | Time (s) | Volume (m3) | Energy (J) |
| DSA 1024 | 2058 | 1.074·1010 | 1.074·103 | 0.028 | 1.667·1014 |
| DSA 2048 | 4106 | 8.590·1010 | 8.590·103 | 0.055 | 2.661·1015 |
| DH 1024 | 2058 | 1.074·1010 | 1.074·103 | 0.028 | 1.667·1014 |
| DH 2048 | 4106 | 8.590·1010 | 8.590·103 | 0.055 | 2.661·1015 |
| DH 3072 | 6154 | 2.899·1011 | 2.899·104 | 0.083 | 1.346·1016 |
| DH 4096 | 8202 | 6.872·1011 | 6.872·104 | 0.111 | 4.252·1016 |
| DH 6144 | 12298 | 2.319·1012 | 2.319·105 | 0.166 | 2.152·1017 |
| DH 8192 | 16394 | 5.498·1012 | 5.498·105 | 0.221 | 6.800·1017 |
| brainpoolP256r1 | 522 | 1.678·108 | 16.78 | 0.007 | 6.607·1011 |
| brainpoolP384r1 | 778 | 5.662·108 | 56.62 | 0.010 | 3.324·1012 |
| brainpoolP521r1 | 1052 | 1.414·109 | 141.42 | 0.014 | 1.122·1013 |
| NIST P-256 | 522 | 1.678·108 | 16.78 | 0.007 | 6.607·1011 |
| NIST P-384 | 778 | 5.662·108 | 56.62 | 0.010 | 3.324·1012 |
| NIST P-521 | 1052 | 1.414·109 | 141.42 | 0.014 | 1.122·1013 |
| Curve25519 | 520 | 1.658·108 | 16.58 | 0.007 | 6.505·1011 |
| Curve448 | 906 | 8.992·108 | 89.92 | 0.012 | 6.146·1012 |
| RSA 512 | 1029 | 6.711·108 | 67.11 | 0.014 | 5.210·1012 |
| RSA 1024 | 2053 | 5.369·109 | 536.87 | 0.028 | 8.316·1013 |
| RSA 2048 | 4101 | 4.295·1010 | 4.295·103 | 0.055 | 1.329·1015 |
| RSA 2540 | 5085 | 8.194·1010 | 8.194·103 | 0.069 | 3.143·1015 |
| RSA 3072 | 6149 | 1.450·1011 | 1.450·104 | 0.083 | 6.725·1015 |
| RSA 4096 | 8197 | 3.436·1011 | 3.436·104 | 0.110 | 2.125·1016 |
| RSA 7168 | 14341 | 1.841·1012 | 1.841·105 | 0.193 | 1.992·1017 |
| RSA 8192 | 16389 | 2.749·1012 | 2.749·105 | 0.221 | 3.399·1017 |
| RSA 13550 | 27105 | 1.244·1013 | 1.244·106 | 0.365 | 2.544·1018 |
| RSA 37100 | 74205 | 2.553·1014 | 2.553·107 | 1.000 | 1.429·1020 |
| RSA 76608 | 153221 | 2.248·1015 | 2.248·108 | 2.064 | 2.599·1021 |
The article contains only an estimate, not a specific implementation. It is not even an optimal estimate, because the development of algorithms for quantum computers will accelerate attacks. Algorithms are already significantly simpler, but their explanation is challenging and it is extremely difficult to find a suitable idea. The question is to what extent the risks arising from quantum computers are overestimated in terms of the time frame. Another issue is the effect of multiplexing, where several qubits use one gate. While this reduces the required number of gates (connections) and their power consumption, the total processing time is also increased in the corresponding proportion. An attack using quantum computers is definitely within the capabilities of the entire human civilization, and perhaps even within the capabilities of some state actors. Nevertheless, it can be somewhat problematic to attack even simpler elliptic curves for a simple reason. This reason is to ensure an adequate amount of available energy. In addition to the design difficulties, a cryptographically relevant quantum computer will also have a problem with the aforementioned power source. According to this basic description, an attack on asymmetric algorithms could require significant investment in resources. On the other hand, increasing the operating temperature by a single order of magnitude will significantly reduce the aforementioned impacts. Added to this is the question of how to solve the problem of waste heat. According to this estimate, quantum computers are still a theoretical attack. However, this attack may become feasible in the foreseeable future, where any technological breakthrough threatens current cryptography.
If, based on the above results, I were to consider power supply using blocks with 1GW reactors for simplicity, I would get the following results. For Curve25519 , brainpoolP256r1 and NIST P-256, it would be 39 blocks, for brainpoolP384r1 and NIST P-384, it would be 58 blocks, for Curve448, it would be 60. blocks and for brainpoolP521r1 and NIST P-521 even 78 blocks.I'm really curious to see how close this estimate will be to reality. Given the possibilities offered by quantum computers, I'd like to be wrong.
At the same time, the above overview shows that for small key sizes, a classical digital computer could be more suitable for decryption than a quantum computer. Examples include DSA 1024b, DH 1024, DSA 2048b and DH 2048b. For small key sizes of the RSA algorithm, this difference is not so great, quantum computers have dominated here from the beginning. For large keys, of course, the quantum computer has a clear advantage in all cases. The biggest problem here is cooling, the temperature close to absolute zero is as limiting as the speed of light. The goal of the development is therefore to shift the operating temperatures of the system to more comfortable temperature ranges. Here is an overview of what this would mean in terms of energy consumption, i.e. the energy required for one calculation.
| Algoritmus | 1μK (J) | 1mK (J) | 0,1K (J) | 1K (J) | 10K (J) | 100K (J) |
| DSA 1024 | 1.667·1017 | 1.667·1014 | 1.667·1012 | 1.667·1011 | 1.667·1010 | 1.667·109 |
| DSA 2048 | 2.661·1018 | 2.661·1015 | 2.661·1013 | 2.661·1012 | 2.661·1011 | 2.661·1010 |
| DH 1024 | 1.667·1017 | 1.667·1014 | 1.667·1012 | 1.667·1011 | 1.667·1010 | 1.667·109 |
| DH 2048 | 2.661·1018 | 2.661·1015 | 2.661·1013 | 2.661·1012 | 2.661·1011 | 2.661·1010 |
| DH 3072 | 1.346·1019 | 1.346·1016 | 1.346·1014 | 1.346·1013 | 1.346·1012 | 1.346·1011 |
| DH 4096 | 4.253·1019 | 4.253·1016 | 4.253·1014 | 4.253·1013 | 4.253·1012 | 4.253·1011 |
| DH 6144 | 2.152·1020 | 2.152·1017 | 2.152·1015 | 2.152·1014 | 2.152·1013 | 2.152·1012 |
| DH 8192 | 6.800·1020 | 6.800·1017 | 6.800·1015 | 6.800·1014 | 6.800·1013 | 6.800·1012 |
| brainpoolP256r1 | 6.607·1014 | 6.607·1011 | 6.607·109 | 6.607·108 | 6.607·107 | 6.607·106 |
| brainpoolP384r1 | 3.324·1015 | 3.324·1012 | 3.324·1010 | 3.324·109 | 3.324·108 | 3.324·107 |
| brainpoolP521r1 | 1.122·1016 | 1.122·1013 | 1.122·1011 | 1.122·1010 | 1.122·109 | 1.122·107 |
| NIST P-256 | 6.607·1014 | 6.607·1011 | 6.607·109 | 6.607·108 | 6.607·107 | 6.607·106 |
| NIST P-384 | 3.324·1015 | 3.324·1012 | 3.324·1010 | 3.324·109 | 3.324·108 | 3.324·107 |
| NIST P-521 | 1.122·1016 | 1.122·1013 | 1.122·1011 | 1.122·1010 | 1.122·109 | 1.122·107 |
| Curve25519 | 6.505·1014 | 6.505·1011 | 6.505·109 | 6.505·108 | 6.505·107 | 6.505·106 |
| Curve448 | 6.146·1015 | 6.146·1012 | 6.146·1010 | 6.146·109 | 6.146·108 | 6.146·107 |
| RSA 512 | 5.210·1015 | 5.210·1012 | 5.210·1010 | 5.210·109 | 5.210·108 | 5.210·107 |
| RSA 1024 | 8.316·1016 | 8.316·1013 | 8.316·1011 | 8.316·1010 | 8.316·109 | 8.316·108 |
| RSA 2048 | 1.329·1018 | 1.329·1015 | 1.329·1013 | 1.329·1012 | 1.329·1011 | 1.329·109 |
| RSA 2540 | 3.143·1018 | 3.143·1015 | 3.143·1013 | 3.143·1012 | 3.143·1011 | 3.143·1010 |
| RSA 3072 | 6.725·1018 | 6.725·1015 | 6.725·1013 | 6.725·1012 | 6.725·1011 | 6.725·1010 |
| RSA 4096 | 2.125·1019 | 2.125·1016 | 2.125·1014 | 2.125·1013 | 2.125·1012 | 2.125·1011 |
| RSA 7168 | 1.992·1020 | 1.992·1017 | 1.992·1015 | 1.992·1014 | 1.992·1013 | 1.992·1012 |
| RSA 8192 | 3.399·1020 | 3.399·1017 | 3.399·1015 | 3.399·1014 | 3.399·1013 | 3.399·1012 |
| RSA 13550 | 2.544·1021 | 2.544·1018 | 2.544·1016 | 2.544·1015 | 2.544·1014 | 2.544·1013 |
| RSA 37100 | 1.429·1023 | 1.429·1020 | 1.429·1018 | 1.429·1017 | 1.429·1016 | 1.429·1014 |
| RSA 76608 | 2.599·1024 | 2.599·1021 | 2.599·1019 | 2.599·1018 | 2.599·1017 | 2.599·1016 |
In this series, I tried to explain the problems of attacks on current asymmetric cryptography and to explain in a somewhat popular way the problems of attacks using quantum computers and the Shorr algorithm. Although the part dedicated to quantum computers tries to explain the mentioned issue, it is mainly a popular form and includes some distortion of information, for which I apologize. This area is extremely complex and its explanation is difficult. In my opinion, however, it is necessary to draw attention to the publicly overlooked problem of attack time depending on the need to achieve a certain accuracy and, in the case of cryogenic devices, the need to effectively cool such devices.
In the future, this series will be loosely followed by a critical assessment of the successes of attacks on asymmetric cryptography using both classical computers (history) and quantum computers. The reason is an effort to draw attention to what I believe to be misleading experiments and their possible inappropriate interpretation. I would like to compute power needed for each quantum algorithms also.
1. Introductory Provisions
1.1. These General Terms and Conditions are, unless otherwise agreed in writing in the contract, an integral part of all contracts relating to training organised or provided by the trainer, Jan Dušátko, IČ 434 797 66, DIČ 7208253041, with location Pod Harfou 938/58, Praha 9 (next as a „lector“).2. Creation of a contract by signing up for a course
2.1. Application means unilateral action of the client addressed to the trainer through a data box with identification euxesuf, e-mailu with address register@cryptosession.cz or register@cryptosession.info, internet pages cryptosession.cz, cryptosession.info or contact phone +420 602 427 840.3. Termination of the contract by cancellation of the application
3.1. The application may be cancelled by the ordering party via e-mail or via a data mailbox.4. Price and payment terms
4.1. By sending the application, the ordering party accepts the contract price (hereinafter referred to as the participation fee) indicated for the course.5. Training conditions
5.1. The trainer is obliged to inform the client 14 days in advance of the location and time of the training, including the start and end dates of the daily programme.6. Complaints
6.1. If the participant is grossly dissatisfied with the course, the trainer is informed of this information.7. Copyright of the provided materials
7.1. The training materials provided by the trainer in the course of the training meet the characteristics of a copyrighted work in accordance with Czech Act No 121/2000 Coll.8. Liability
8.1. The trainer does not assume responsibility for any shortcomings in the services of any third party that he uses in the training.9. Validity of the Terms
9.1 These General Terms and Conditions shall be valid and effective from 1 October 2024.Consent to the collection and processing of personal data
According to Regulation (EU) No 2016/679 of the European Parliament and of the Council on the protection of individuals with regard to the processing of personal data and on the free movement of such data and repealing Directive 95/46/EC (General Data Protection Regulation, hereinafter referred to as "the Regulation"), the processor xxx (hereinafter referred to as "the Controller") processes personal data. Individual personal data that are part of the processing during specific activities at this web presentation and in the course of trade are also broken down.Information about the records of access to the web presentation
This website does not collect any cookies. The site does not use any analytical scripts of third parties (social networks, cloud providers). For these reasons, an option is also offered for displaying the map in the form of a link, where the primary source is OpenStreet and alternatives then the frequently used Maps of Seznam, a.s., or Google Maps of Google LLC Inc. The use of any of these sources is entirely at the discretion of the users of this site. The administrator is not responsible for the collection of data carried out by these companies, does not provide them with data about users and does not cooperate on the collection of data.Information about contacting the operator of the site
The form for contacting the operator of the site (administrator) contains the following personal data: name, surname, e-mail. These data are intended only for this communication, corresponding to the address of the user and are kept for the time necessary to fulfil the purpose, up to a maximum of one year, unless the user determines otherwise.Information about the order form
In case of an interest in the order form, the form contains more data, i.e. name, surname, e-mail and contact details for the organisation. These data are intended only for this communication, corresponding to the address of the user and are kept for one year, unless the user determines otherwise. In the event that a business relationship is concluded on the basis of this order, only the information required by Czech law on the basis of business relations (company name and address, bank account number, type of course and its price) will continue to be kept by the administrator.Information about the course completion document
Within the course, a course completion document is issued by the processor. This document contains the following data: student's name and surname, the name and date of the course completion and the employer's name. The information is subsequently used for the creation of a linear hash tree (non-modifiable record). This database contains only information about the provided names and company names, which may or may not correspond to reality and is maintained by the processor for possible re-issuance or verification of the document's issuance.Rights of the personal data subject
The customer or visitor of this website has the possibility to request information about the processing of personal data, the right to request access to personal data, or the right to request the correction or deletion of any data held about him. In the case of deletion, this requirement cannot be fulfilled only if it is not data strictly necessary in the course of business. The customer or visitor of this website also has the right to obtain explanations regarding the processing of his personal data if he finds out or believes that the processing is carried out in violation of the protection of his private and personal life or in violation of applicable legislation, and the right to request removal of the resulting situation and to ensure the correction.