dlgeek 2 years ago

This is really interesting research. But I'm a little surprised a research group was able to tap the entirety of a large university campus's internet traffic. Is that common?

  • amarshall 2 years ago

    The research is by teams at those universities whose networks were observed, not some outside group. There’s an entire section on “Ethical Considerations” in the paper that details what they did (and did not) collect. Like any research involving human subjects in some way, their methodologies were reviewed by their respective university institutional review boards (not that those are necessarily perfect).

  • londons_explore 2 years ago

    There are plenty of organisations who will let you tap Gbps network links for research purposes and a small fee.

    Network data isn't supposed to be secret, and in fact I wish more networks would publish samples of data that passes through the network so users inadvertently not using encryption can be alerted.

fulafel 2 years ago

From the abstract:

> The faulty signatures we observed allowed us to compute private RSA keys associated with a top-10 Alexa site, several browser-trusted wildcard certificates for organizations that used a popular VPN product, and a small sporadic population of other web sites and network devices.

jwilk 2 years ago

Abstract:

> It is well known in the cryptographic literature that the most common digital signature schemes used in practice can fail catastrophically in the presence of faults during computation. We use passive and active network measurements to analyze organically-occuring faults in billions of digital signatures generated by tens of millions of hosts. We find that a persistent rate of apparent hardware faults in unprotected implementations has resulted in compromised certificate RSA private keys for years. The faulty signatures we observed allowed us to compute private RSA keys associated with a top-10 Alexa site, several browser-trusted wildcard certificates for organizations that used a popular VPN product, and a small sporadic population of other web sites and network devices. These measurements illustrate the fragility of RSA PKCS#1v1.5 signature padding and provide insight on the risks faced by unprotected implementations on hardware at Internet scale.

londons_explore 2 years ago

So these researchers collected ~200 million TLS handshakes, and found a few hundred that were miscomputed, they suspect by bit errors.

However, I do not believe modern computational devices are so unreliable. If I computed 200 million TLS exchanges on my home PC over a few days, I wouldn't expect a single one to be miscomputed. Servers with ECC memory ought to be another order of magnitude more reliable.

So why do we see such high rates of miscomputation?

  • teraflop 2 years ago

    The research doesn't necessarily imply that any typical device has such a high failure probability. From the paper:

    > The three private keys revealed by the 11 faulty [RSA] signatures in our [passively observed] data were associated with three certificates that were served from four different IP addresses associated with Baidu. [...]

    > After we disclosed to Baidu, they informed us that the traffic we observed was between the clients and Baidu’s golang-based L7 load balancer BFE which offloads cryptographic operations like signature generation to a hardware accelerator. [...] Based on the temporal pattern of signature errors we observed, we hypothesize that the errors may have been due to a single failing hardware component which then passed vulnerable signatures through the unprotected software implementation.

    • WoahNoun 2 years ago

      Yea it's just a coincidence that most of the key leaks identified in a US university campus internet are from Baidu. Just a random hardware error, nothing suspicious at all...