Permalink
Fetching contributors…
Cannot retrieve contributors at this time
108 lines (87 sloc) 4.95 KB

Full Disclosure Versus Responsible Disclosure

Very special and crucial note: This article discusses my personal feelings regarding vulnerability disclosure methods. This article is not written on behalf of HardenedBSD. This article does NOT reflect the opinions, beliefs, or actions of The HardenedBSD Project.

Meltdown and Spectre, along with the other CPU microarchitecture speculation vulnerabilities, will be the prime use case in this article.

It's fine if you don't agree with this article. If everyone held my same opinions and beliefs, the world would be a very boring place. Diverse thought makes experiencing and living in this world exciting.

My goal with this article is to start a much-needed discussion. I hope we in the infosec industry can learn from recent events and move forward together.

My perspective may change with time. I'll keep this article updated with regards to my perspective. I try to be malleable and understanding. If I've completely missed the mark here, that's fine. This is only a reflection of some of the recent events regarding large architectural vulnerabilities that impact an extremely broad audience. If you feel I'm out in left field, please reach out to me and we can discuss the matter.

Definitions

We live in a time when vendors prefer security researchers disclose newly-found vulnerabilities responsibly. We don't have a firm definition of what "responsible disclosure" is, though. Usually, researchers will follow this pattern:

  1. Researcher finds exploitable vulnerability
  2. Researcher notifies vendor, giving a projected embargo date
  3. Vendor resolves issues prior to expiration of embargo
  4. Researcher and vendor release a joint announcement

This workflow typically works when the vendor produces easy-to-update products that don't have many (or any) downstream dependent consumers. Firefox is an easy example. All Mozilla has to do is fix the bug and cut a new release. The package maintainers for the various operating systems will perform a version bump and the problem is largely solved.

The definition of "Responsible Disclosure" becomes murky when a researcher finds a vulnerability in a hard-to-update product used in a massive scale, like Intel CPUs. Sure, Intel should be notified, but certain vulnerabilities have much larger impact than Intel's product line by itself. Meltdown and Spectre provide an excellent case study.

Thus, when we define "Responsible Disclosure," we also need to define "Coordinated Disclosure." Vendors must responsibly decide which downstream projects get notified and in which order.

The Coordinated Disclosure Problem

A lot of business today is about relationships. This is especially true in information security. We value the opinions of people we know and trust and distrust everyone else.

When a vendor receives a security notification, the vendor must evaluate who could be impacted and of those impacted, who to notify. Coordinated disclosures usually come with either a Non-Disclosure Agreement (NDA) or an embargo until a specified date. The vendor can choose to collaborate with zero or more impacted entities of the vendor's own choosing.

When Intel received notice from Google regarding Meltdown and Spectre, Intel chose to notify only a few larger entities (eg, Amazon, Google, and Linux). Those entities got a six-month advanced notice and had at least some semblance of a fix the day the story broke in January 2018. FreeBSD received notice the week around Christmas 2017 only after certain researchers in-the-know pushed Intel hard to notify FreeBSD. Smaller hosting companies and all of the other BSDs were left out.

Thus, Intel only coordinated the disclosure with a few larger players. Favoritism is at play here. Such favoritism has a direct impact on the economy of the less fortunate. Why would I go with a smaller hosting provider when Amazon receives and can take action on advanced vulnerability notifications? If security is of any concern, I should only use Amazon's and Google's services. Screw the smaller players, right?

Coordinated disclosure, as currently performed, should be called unethical selective disclosure. Pay the right people your agreed extortion, and you're now on the favorites list. When did it become acceptable to pay someone to tell you you're vulnerable because of their mistakes?

The Case For Full Disclosure

Since so-called responsible disclosure inevitably leads to greed and extortion, full disclosure is the responsible way forward. All downstream vendors become equals.

Full disclosure lends itself to a better collaborative effort. Now that all entities are on equal footing, they can work together to provide their communities the best quality fix. No one single vendor has an omniscient developer capable of writing "the perfect fix." If vendors work side-by-side, they can anticipate future problems along with providing innovative fixes and mitigations for current issues.