Google Ups Bug Bounty To $20,000

Trailrunner7 writes, quoting Threatpost: "Search giant Google said it is quintupling the top bounty it will pay for information on security holes in its products to $20,000. Google said it was updating its rewards and rules for the bounty program, which is celebrating its first anniversary. In addition to a top prize of $20,000 for vulnerabilities that allow code to be executed on product systems, Google said it would pay $10,000 for SQL injection and equivalent vulnerabilities in its services and for certain vulnerabilities that leak information or allow attackers to bypass authentication or authorization features."

Comments

A failure of conventional hack-ism ?

[Taco Cowboy]

I am sure Google is employing many many very able programmers, but if Google has to pay bounty to hackers up to $20,000 to find bugs, does that mean the programmers who are sitting in Google's offices around the world have phailed?

Re:A failure of conventional hack-ism ?

[mark-t]

It probably means that they realize that they've come to a point in the project where crowdsourcing QA is more cost-effective than using internal QA. This isn't because their internal QA is incompetent, it's because they are only just so many.

Re:A failure of conventional hack-ism ?

[Anonymous Coward]

the inference to be drawn is that finding a security hole would take more than 20k of programmer time, so probably the holes remaining are _hard_ to find. Seems more like a success than a failure to me.

Re:A failure of conventional hack-ism ?

[Bucky24]

I can see why you might think that, but I strongly suspect that Google has already put their own programmers to work finding bugs. This is their attempt to "crowdsource" the bug-finding. The more eyes on the code, the more bugs that can be found. Also they realize that not all the brilliant minds work for them, and some might decide to exploit a bug for monetary gain rather then turn it in. The bounty is to give those people a bit more of a reason to turn the bug in.

Re:A failure of conventional hack-ism ?

[FSWKU]

I am sure Google is employing many many very able programmers, but if Google has to pay bounty to hackers up to $20,000 to find bugs, does that mean the programmers who are sitting in Google's offices around the world have phailed?

Not necessarily. It just means that while they're confident in their code, they believe that it's always a good idea to have things vetted in the real world. The reasoning behind this is that the developers are often so close to the code that they can't possibly see EVERY conceivable bug or vulnerability. Inviting others to poke your products with a stick on a constant basis is a good thing. It lets Google get some good press, and also a MUCH more thorough real-world trial than they could do in house.

In a way, it's somewhat remniscent of the developers who worked on the flight software for the Space Shuttle computers. Teams would actually compete to see who could find more bugs in the other team's code. This lead to some of the most robust and bug-free software ever written.

Re:A failure of conventional hack-ism ?

[jhoegl]

Nope, it means they are offering proper market value for bugs found in their systems and are confident enough to offer such high bounties for them.

If, however, this were Microsoft or Apple, they would not offer such high amounts as bounties as they would soon go bankrupt from the financial burden of paying out these bounties.

So, not only is Google saying "we are confident and proud of our product" they are also saying "we know there are bugs and even though we are confident in our products we are willing to pay out for people finding them".

Re:A failure of conventional hack-ism ?

[lostchicken]

I'd love to see a more vibrant market for this. The cost paid per bug (perhaps normalized by product revenue) would be a really useful measure of software reliability.

Re:A failure of conventional hack-ism ?

[Monkier]

Which is a much better position than "Let's pretend there's no bugs, and hush up anyone who says there is". Nice one, Google...

Re:A failure of conventional hack-ism ?

[Anonymous Coward]

What they're offering is still well below the $100,000 that a digital arms dealer like Vupen charges for a year's subscription plan for exploits it discovers. And according to the Forbes article I linked to, some vulnerabilities individually cost several times more than that. It's so fucked up that NATO counties pay these security firms like Vupen, HB Gary Federal, etc. for exploits in the products of legitimate software companies for their use in cyberwarfare, espionage, and other nefarious shit. They'd rather leave everyone vulnerable, not even using the info they purchase to shore up their own government's systems lest the vulnerability become public and they lose the value of their purchase. If I were Google I'd save the bounty money and give it to their lawyers to create a tsunami of FOIA requests with every government they can to get the info about whatever exploits they have. Start a PR campaign letting the public know that their own government have knowledge that could help software companies make their products more secure for the computing public at large. Maybe if some influential people in the security field and tech firms complain loudly enough, something will change. I doubt it, but what hell else is there to do?

Re:A failure of conventional hack-ism ?

[AbRASiON]

This 20k figure has very little to do with programmer time.

It has a LOT more to do with googles size and customer base. When you look at just how much data google has, just how many customers (paying as well, with Google Apps) 20k is absoloute pocket change to possibly convince someone malicious to instead fix the problem than exploit it.

Re:A failure of conventional hack-ism ?

[lucm]

If, however, this were Microsoft or Apple, they would not offer such high amounts as bounties as they would soon go bankrupt from the financial burden of paying out these bounties.

Microsoft has a more cost-effective way to deal with bugs: the MVP, aka as "unpaid Level 1 support staff unleashed on forums and blogs that accept to do Microsoft's work in exchange for a title, a pin and a secret handshake instead of a salary". And when MVP cannot solve a problem or find a serious bug, they simply push them on Microsoft Connect and wait for a Service Pack or a hotfix, like the Common People.

Apple has a similar program (maybe even more brilliant) but only at the marketing level and they provide no formal title; the unpaid staff is informally called "fanbois" by those who know better, and they are convinced that promoting Apple products is a reward by itself so they don't need a pin or a logo.

Re:A failure of conventional hack-ism ?

[MobileTatsu-NJG]

It's more likely that a bug would do more than $20,000 worth if damage.

Obligatory Dilbert

[ace37]

Bug bounty: http://dilbert.com/strips/comic/1995-11-13/

Granted it's external rather than internal pay for a bug, but at $20k a piece, it wouldn't take a sleazy employee like ratbert long to figure out...

Re:Obligatory Dilbert

[icebraining]

Yes, I'm sure a Google employee will risk their $110k+benefits job and being unemployable for life in any major tech company to gain $20k.

Re:mmm

[multiben]

No one's forcing anyone to do it. Some people love spending their time trying to punch holes in security. This way they can do it legally and get rewarded for it. Seems like a pretty sweet deal to me.

Three reasons

[gstrickler]

1. Bugs are getting harder to find, especially ones that can be exploited
2. Criminals are paying good money for quality exploits.
3. It's cheaper than hiring more people to do it.

Re:Carry on Google..

[icebraining]

Exactly; it's "Don't be evil", not "Do no evil". And if you think about it, the difference is huge.

Having been an MVP...

[Interfacer]

I can tell you that it depends on which product group you are active in.
some teams like the C++ product group have (at least when I was an MVP) a very good relationship with their MVPs. this included getting developers to look at weird bugs, getting lots of interesting information, technical previews, etc. From my experience, the low level groups (SDK, DDK, C++) had a very active private community going with their MVPs.

For people interested in the product they were working with (C++ and SDK for me) being an MVP can be a rewarding experience, because you get a lot of technical inside information, people to talk with, an MSDN subscription, some free incidents with Microsoft support, etc.

Other (often the larger) product groups had virtually no real relationship with their MVPs, and some groups just treated them like unpaid 1st level support. Or in some cases they just plain ignored them. What it means to be an MVP and what you get out of it really depends entirely on your category and interest group. It can range from very good to very crappy.