Novel attack slashes power to crack textbook RSA
The digital fortress we rely on—encryption—is facing a seismic shift. For years, factoring attacks against the widely used RSA algorithm were considered a mathematical impossibility, a secure barrier impenetrable to even the most determined attackers. That perception is now being dramatically challenged by new research, suggesting that the rules of cryptographic security may need a serious rethink.
A team of researchers from UC San Diego and Inria Nancy, working together, has published findings that outline a new method capable of breaking RSA in record time. If these findings hold up under peer review, the implications for global data security could be enormous, particularly since RSA-based encryption remains the backbone for a substantial portion of online services.
This isn’t just theoretical speculation; the research details an attack that drastically reduces the number of operations required to crack an RSA key. For RSA key lengths of 1024, 2048, 3072, and 4096 bits, the theoretical effort required to crack the key drops from an exponentially large number to a much more manageable level. For example, for a 1024-bit key, the required calculation is theoretically reduced from 2^80 down to a level involving 2^65.
While the sheer number of calculations still presents a monumental challenge for an ordinary attacker, this discovery shifts the vulnerability landscape. The research highlights that the algorithms used in this attack are not fully optimized, meaning further refinements—potentially aided by artificial intelligence and GPU power—could significantly accelerate the process.
Crucially, the attack targets the “raw” or textbook variant of RSA, which is often used in specific, high-security contexts rather than commonplace activities like standard website certificates. However, this class of encryption is embedded in many critical systems, including services like Cloudflare Privacy Pass and Apple’s iCloud Private Relay.
Systems utilizing PKCS #11, hardware security modules (HSMs), and trusted platform modules (TPMs) also rely on this vulnerable variant. The danger lies in the fact that an attacker doesn’t need to brute-force the key directly; they can exploit an “oracle.”
An oracle in this context is essentially a mechanism that takes an input, runs a complex computation involving the private key, and returns an output, all without revealing the key itself. By repeatedly querying this oracle—for instance, querying a live server encrypted with the target key billions of times—an attacker can gather enough data points to perform the mathematical wizardry necessary to finally inspect encrypted traffic.
While network administrators are well aware that billions of queries against a live service would result in immediate IP bans, the potential remains. If key management systems are not closely monitored, or if hardware-based security mechanisms are less scrutinized, the potential for data harvesting through such oracle attacks increases.
The key takeaway is that even if a direct, real-time attack is currently high-bar, the finding opens the door to a future threat. Data collected through these oracle methods today could potentially be leveraged to decrypt previously saved traffic in weeks or months, especially as hardware continues to advance.
Buy on Amazon
Credit
