17.1.17. In all cases a comprehensive risk assessment should be undertaken before configurations are selected. Some general principles to be considered are: Cryptographic strength is determined by a combination of the encryption algorithm being used, the encryption protocol and the key length. Longer keys generally provide increased encryption strength over shorter keys when using the same encryption algorithm; Asymmetric cryptographic algorithms are slower than symmetric cryptographic algorithms at an equivalent cryptographic strength; Asymmetric cryptographic algorithms are recommended for the exchange of symmetric cryptographic keys when they are needed to be shared across communication channels; Encrypted data cannot usually be compressed, but compressed data can be encrypted. Data should be compressed before encryption; Encryption keys have the same requirements for handling and storage as the unencrypted data they are being used to protect; Any risk assessment should include consideration of key management – refer tosection 17.9 Key Management. 17.1.18. It is important to note that the NZISM prescribes approved algorithms and protocols and users must select combinations from these lists. Quantum Computing and Encryption 17.1.19. Developments in quantum computing have highlighted threats to classical cryptography whereby a quantum computing, can undermine all of the widely used public key algorithms used for key establishment and digital signatures. While this may be not an immediate issue, quantum developments are likely to undermine the effectiveness of encryption being used today to protect confidentiality of information. 17.1.20. A further implication is that historical and archived data protected by encryption may be at risk. 17.1.21. It is generally accepted that symmetric encryption, with sufficiently long keys, will remain quantum resistant in the short term but that quantum resistant replacements for digital signature and key establishment algorithms will be required in the near future. 17.1.22. In the longer term, quantum resistant algorithms are expected to be developed, standardised and approved for use. Until such time, however, agencies should be positioning themselves to be ready to migrate to a “post-quantum encryption” environment. 17.1.23. As it is now recognised that agencies will need to undertake future migration activities related to post-quantum encryption, it is no longer specifically advised to invest in migration from RSA to ECC-based algorithms if that has not already taken place. Emphasis should instead be placed on ensuring minimum key lengths specified in the NZISM are adhered to. Transitioning Cryptographic Algorithms and Protocols 17.1.24. It is important to use algorithms that adequately protect sensitive information. It is also important to recognise that all cryptographic algorithms and protocols have a finite life. Challenges are posed by new cryptanalysis techniques and methods, the increasing power of classical computing technology, and the continuing work on the development of quantum computers. In addition, there is an active field of work that continuously seeks to compromise algorithms and protocols currently in use. 17.1.25. Planning for changes in the use of cryptography because of algorithm breaks, the availability of more powerful computing techniques or new technologies is an important consideration for agencies. Awareness of retirement or deprecation of algorithms and associated protocols is essential. RSA 17.1.26. RSA was announced in 1976 and is now over 45 years old. Several flaws and attacks have been identified since creation, each of which required specific mitigations, careful implementation and management. Unfortunately there is ample evidence that implementers continue to have difficulties in securely implementing, using and managing RSA. 17.1.27. To counter identified threats from shorter RSA key lengths, longer key lengths have been specified in the NZISM since 2010. Minimum key lengths have been subsequently increased over time. 17.1.28. For a number of years there had been several indicators that RSA was likely to be deprecated by the cryptographic community and standards bodies. For example, TLS 1.3 has deprecated the use of RSA for key exchange in favour of elliptic curve cryptography, but RSA is still supported for digital signatures in the current standard. Previous guidance from NIST was also indicative of the impending deprecation of RSA. However, subsequent guidance no longer recommends moving from RSA to elliptic curve if that has not already been done. 17.1.29. Therefore, while RSA is not fully deprecated in the NZISM, it is approved ONLY for a limited set of uses as described inSection 17.2 – Approved Cryptographic Algorithms. Product specific cryptographic requirements 17.1.30. This section provides requirements for the use of cryptography to protect classified information. Requirements, in addition to those in this Manual, can exist in consumer guides for products once they have completed an approved evaluation. Vendor specifications supplement this manual and where conflict in controls occurs the product specific requirements take precedence. Any policy or compliance conflicts are to be incorporated into the risk assessment. Exceptions for using cryptographic products 17.1.31. Where Agencies implement a product that uses an Approved Cryptographic Algorithm or Approved Cryptographic Protocol to provide protection of unclassified data at rest or in transit, that product does not require a separate, approved evaluation. Correct implementation of the cryptographic protocol is fundamental to the proper operation of the Approved Cryptographic Algorithm or Approved Cryptographic Protocol and is part of the checking conducted during system certification. Federal Information Processing Standard 140 17.1.32. FIPS 140 is a United States standard for the evaluation and validation of both hardware and software cryptographic modules. 17.1.33. FIPS 140 is in its third iteration and is formally referred to as FIPS 140-3. This section refers to the standard as FIPS 140 but this should be considered to encompass FIPS 140-1, FIPS 140-2 and FIPS 140-3. 17.1.34. FIPS 140 is not a substitute for an approved evaluation of a product with cryptographic functionality. FIPS 140 is concerned solely with the cryptographic functionality of a module and does not consider any other security functionality. 17.1.35. Cryptographic evaluations of products will normally be conducted by an approved agency. Where a product’s cryptographic functionality has been validated under FIPS 140, the GCSB can, at its discretion, and in consultation with the vendor, reduce the scope of a cryptographic evaluation. New Zealand National Policy for High Assurance Cryptographic Equipment and Key Management 232 Version_3.5__January-2022

Select target paragraph3

Connect to a paragraph
Connect to an entity
Disable highlights
Add to table of contents