论文标题
通过基于请求的碎片化，DNSSEC中的量子后签名
Post-Quantum Signatures in DNSSEC via Request-Based Fragmentation
论文作者
论文摘要
域名系统安全扩展（DNSSEC）使用数字签名对DNS响应进行身份验证。 DNS主要在UDP上运行，这导致了几个限制：值得注意的是，数据包最多应长1232字节，以避免在传输过程中避免出现问题。较大的DNS响应要么需要分散成几个UDP响应，要么需要通过TCP重复该请求，在当今的DNS生态系统中，这两个响应都不是足够可靠的。尽管RSA或椭圆曲线数字签名足够小，以避免此问题，即使对于包含公钥和签名的DNSSEC数据包，考虑到更大尺寸的量子后方案时，此问题是不可避免的。 我们提出了ARRF，这是一种基于请求的应用程序层（而不是传输层）的DNS资源记录的方法，这意味着初始响应包含截断的片段，然后请求者发送其余片段的后续请求。使用基于请求的碎片避免了针对先前提出的几种（和拒绝）应用程序级DNS片段化技术发现的问题。我们在128位安全级别上使用NIST选择标准化（Falcon，Dilithium和Sphincs+）选择的三种后量词后数字签名方案时，在模拟网络中进行了方法。我们的实验表明，对于所有测试的量子后算法，我们基于请求的碎片方法与UDP的标准DNS相比，分辨率的时间大大降低，而在Falcon和dilithium的情况下，数据传输的数据较少。此外，我们基于请求的碎片设计可以相对容易地实现：我们的实现实际上是一个小型守护程序，可以坐在DNS名称服务器或解析器前面，以透明地片段/重新组装。
The Domain Name System Security Extensions (DNSSEC) provide authentication of DNS responses using digital signatures. DNS operates primarily over UDP, which leads to several constraints: notably, packets should be at most 1232 bytes long to avoid problems during transmission. Larger DNS responses either need to be fragmented into several UDP responses or the request would need to be repeated over TCP, neither of which is sufficiently reliable in today's DNS ecosystem. While RSA or elliptic curve digital signatures are sufficiently small to avoid this problem, even for DNSSEC packets containing both a public key and a signature, this problem is unavoidable when considering the larger sizes of post-quantum schemes. We propose ARRF, a method of fragmenting DNS resource records at the application layer (rather than the transport layer) that is request-based, meaning the initial response contains a truncated fragment and then the requester sends follow-up requests for the remaining fragments. Using request-based fragmentation avoids problems identified for several previously proposed (and rejected) application-level DNS fragmentation techniques. We implement our approach and evaluate its performance in a simulated network when used for the three post-quantum digital signature schemes selected by NIST for standardization (Falcon, Dilithium, and SPHINCS+) at the 128-bit security level. Our experiments show that our request-based fragmentation approach provides substantially lower resolution times compared to standard DNS over UDP with TCP fallback, for all the tested post-quantum algorithms, and with less data transmitted in the case of both Falcon and Dilithium. Furthermore, our request-based fragmentation design can be implemented relatively easily: our implementation is in fact a small daemon that can sit in front of a DNS name server or resolver to fragment/reassemble transparently.
