Not So Unbreakable

Question

Not So Unbreakable

Steve Jones - SSC Editor

SSC Guru

Points: 734418
More actions
December 7, 2011 at 11:32 pm

#264428

Comments posted to this topic are about the item Not So Unbreakable

Viewing 8 posts - 1 through 7 (of 7 total)

You must be logged in to reply to this topic. Login to reply

djackson 22568 SSChampion Points: 11733 More actions · Answer 1

I disagree with the current thinking that disclosures should be delayed. The thinking is short sighted. Those in the industry who push this, and it is probably the majority of those we hear, are looking out for the company that produced the product that has a flaw. Do we believe hackers are going to delay notifying other hackers about a flaw, to give the company a chance to correct the issue?

So what is the result of delaying?

Thoes who seek to cause harm immediately share information, resulting in widespread knowledge of flaws among those who are malicious. Thoes who seek to protect the company do not share, resulting in DBA's and other professionals not being aware of potential issues. Further, if the company affected decides the flaw isn't important, they may not even work on a fix for months, sometimes years. Sometimes, like in a well documented case involving a well known non-Microsoft media player with the initials RP, they deny it even exists even given evidence proving it does. In the meantime customers are at risk.

You appear to be suggesting a coordinated group that would investigate, identify, and report behind the scenes to companies with software flaws.

What makes us believe that group would identify risks sooner than hackers, or even all of the ones hackers find? Should they miss just one, or simply not have the power to encourage the company to fix it, there is no benefit to anyone, and much increased risk as customers are then putting more faith in the (very flawed) process. In the end, I think a fix to this process that doesn't work reliably, doesn't provide identification prior to anyone else, and that doesn't provide a reliable resolution, is worse than how it is today. The only way I see to make this work is for software vendors to release the product to this independent group, prior to public release, for sufficient time to allow all important holes to be closed. That just ain't gonna happen.

Dave

Steve Jones - SSC Editor SSC Guru Points: 734418 More actions · Answer 2

I think that some delay in reports is good. If a severe flaw is found, it's absolutely possible that x, say 10, criminals know about it. Therefore we have 10 people that can exploit the flaw.

If the researcher discloses that right away, then millions of people can potentially take advantage of the flaw. Even the average programmer at a company might have fun with some DoS or crash attack at their company if they can do it anonymously. Remember, lots of disclosures include example code or descriptions.

However if we allow, say a 3 month delay, then the vendor has 3 months to a) develop a patch, b) test a patch (I prefer they do this), and c) distribute it to allow customers to patch. There can be a delay after c) for many companies who also want to test.

I wouldn't advocate "delays until patched" as we have now, or delays for a long period of time since we have new people potentially learning of the flaw every day. I do think some reasonable delay makes sense.

chrisfradenburg SSCrazy Eights Points: 9592 More actions · Answer 3

I'm not sure that a resource issue with the development process is a good justification for delaying security patches. I do understand that other items need to be in the queue but if a companies method of prioritization is such that the number of known security issues grows while other features are getting coding then they have an issue with prioritization. If it ain't broke, don't fix it. However, if it's broke and you're finding it's more broke stop building on it until it's fixed.

Revenant SSC-Forever Points: 42467 More actions · Answer 4

djackson 22568 (12/8/2011)
. . . What makes us believe that group would identify risks sooner than hackers, or even all of the ones hackers find? . . .

Companies that really care about security have their own moles in the hackers groups and are able to judge in which direction is the hacker research going. They typically have enough resources to do their own research, with a good chance of discovering vulnerabilities first.

Don Heimiller-720851 Grasshopper Points: 16 More actions · Answer 5

I see both sides of this. If you put an announcement right away then more people could potentially exploit the Vulnerability until it is patched. On the other side with all of the legislation out there I also do not want to be responsible for vital or private data being stolen because I did not know that was a risk to my data. There needs to be some way that companies that have to be SOX/PCI/HIPPA (US) or the EU’s equivalent of those laws to know that there is a risk out there. There could also be some serious financial impact to companies because they did not protect themselves from the threat.

djackson 22568 SSChampion Points: 11733 More actions · Answer 6

cfradenburg (12/8/2011)
I'm not sure that a resource issue with the development process is a good justification for delaying security patches. I do understand that other items need to be in the queue but if a companies method of prioritization is such that the number of known security issues grows while other features are getting coding then they have an issue with prioritization. If it ain't broke, don't fix it. However, if it's broke and you're finding it's more broke stop building on it until it's fixed.

In my experience with absolutely every software company I have been involved with, either as a worker or customer, the focus is on new sales. This ends up meaning that a bug fix is not prioritized, because sales and marketing are looking for that cool new feature to market, or that part of the product that meets some government regulation, or anything like that.

Fixing a bug is not cool, nor marketable, and sales and marketing are almost never going to push for that over what they view as pushing new sales.

Whether they are correct or not is irrelevant, but it is how they think.

Dave

Revenant SSC-Forever Points: 42467 More actions · Answer 7

djackson 22568 (12/8/2011)
cfradenburg (12/8/2011)
I'm not sure that a resource issue with the development process is a good justification for delaying security patches. I do understand that other items need to be in the queue but if a companies method of prioritization is such that the number of known security issues grows while other features are getting coding then they have an issue with prioritization. If it ain't broke, don't fix it. However, if it's broke and you're finding it's more broke stop building on it until it's fixed.
In my experience with absolutely every software company I have been involved with, either as a worker or customer, the focus is on new sales. This ends up meaning that a bug fix is not prioritized, because sales and marketing are looking for that cool new feature to market, or that part of the product that meets some government regulation, or anything like that.
Fixing a bug is not cool, nor marketable, and sales and marketing are almost never going to push for that over what they view as pushing new sales.
Whether they are correct or not is irrelevant, but it is how they think.
Dave

Not so, in my experience. You have to factor in the cost of support, which goes up if the product is buggy. Sales want their features today, but top management has to see the whole picture and make a balanced decision, and it is the top management that gives a go to the new release.

That said, I do know that there are "smart" managers who release a product with known serious bugs on assumption that before the customer installs it, there will be a fix available on the Web. Sometimes it works, often it does not.