The transition to post-quantum cryptography (PQC) is often framed as a mathematical race to replace vulnerable algorithms like RSA and ECC before a cryptographically relevant quantum computer arrives. However, as Vasileios Mavroeidis and colleagues point out in The Impact of Quantum Computing on Present Cryptography, the primary hurdle for real-world adoption isn't just the complexity of the math—it is the physical size of the data.
Post-quantum primitives are "fat." While a classical Elliptic Curve (ECC) signature might take up 64 bytes, a quantum-resistant counterpart can be orders of magnitude larger, ranging from several kilobytes to megabytes. This "size tax" creates a massive communication bottleneck that threatens to break the very protocols—TLS, DNSSEC, and blockchains—that secure the modern internet.
The Packet Fragmentation Barrier
In standard networking, the Maximum Transmission Unit (MTU) is typically 1,500 bytes. If a cryptographic handshake exceeds this limit, the data must be fragmented into multiple packets. As Hyeongmin Choe and the team behind HAETAE explain, fragmentation is a performance killer: it increases latency, raises the risk of packet loss, and can even cause some network middleboxes to drop the connection entirely.
The goal for many researchers is to fit a digital signature within a single 1.5 KB packet. While NIST-standardized algorithms like Dilithium are efficient, their signatures are often just large enough to trigger this fragmentation. This has led to a push for "shorter" lattice-based signatures that use more precise mathematical pivots to compress data closer to its entropy limits, even if it means a slight trade-off in signing speed.
The Storage Tax and Blockchain Bloat
The size problem extends beyond live communication to long-term storage. In decentralized ledgers, every transaction must be signed. Transitioning to PQC signatures like SPHINCS+ or Dilithium would lead to what Maxime Buser and others describe as "blockchain bloat" in A Survey on Exotic Signatures for Post-quantum Blockchain.
If every transaction signature grows from 64 bytes to 2,500 bytes, the storage requirements for a full node would skyrocket, potentially centralizing the network by pricing out smaller participants. This has forced researchers like Dragos Ioan Ilie to propose configurable defenses for Bitcoin that balance the speed of the transition against the risk of network congestion, as detailed in Committing to Quantum Resistance, Better.
Resource-Constrained Edge Devices
The bottleneck is most acute at the "edge" of the internet. Internet of Things (IoT) devices and embedded systems often operate with severe constraints:
- Memory: Many microcontrollers have as little as 64KB of RAM, making it difficult to even buffer a large post-quantum public key.
- Energy: Transmitting more data requires more radio power. Abid Khan notes in On the Role of Hash-Based Signatures in Quantum-Safe Internet of Things that packet fragmentation in low-power wide-area networks (LPWANs) leads to significantly higher energy consumption, shortening the battery life of remote sensors.
Ben Dong evaluates this in Evaluating Post-Quantum Cryptography on Embedded Systems, finding that while lattice-based schemes like Kyber are viable for IoT, others like SPHINCS+ impose a heavy performance penalty due to their massive signature sizes.
Strategies for Slimming Down
To make PQC practical, researchers are attacking the size problem from several angles:
1. Information-Theoretic Compression
By treating the encryption process as a noisy communication channel, researchers like Antonia Wachter-Zeh have shown that the ciphertext and key sizes of standards like Kyber can be reduced by 25% to 39% without sacrificing security. As described in Reducing Ciphertext and Key Sizes for MLWE-Based Cryptosystems, this is achieved by applying tighter bounds to the noise added during the "Learning with Errors" process.
2. Batching and Amortization
In scenarios where a sender needs to talk to many recipients, "batching" can reduce the overhead per user. Hongxiao Wang introduced mmCipher, which uses a technique called Extended Reproducible PKE to achieve a 45x reduction in communication overhead for group communications compared to standard post-quantum methods.
3. Optimized Primitives
Some researchers are moving away from lattices entirely to find smaller footprints. Seongkwang Kim and Jincheol Ha developed AIM, a symmetric primitive designed specifically to make "MPC-in-the-Head" signatures smaller. By reducing the multiplicative complexity of the underlying one-way function, they can generate zero-knowledge proofs that are significantly more compact.
4. Protocol-Level Integration
Finally, high-performance implementations like OpenSSLNTRU by Daniel J. Bernstein and others demonstrate that with optimized AVX2 software and batch key generation, post-quantum handshakes can actually outperform traditional X25519 in terms of raw throughput, provided the network environment has sufficient bandwidth to handle the larger keys.