As I’ve mentioned here once or twice, the Minneapolis-Saint Paul chapter of the Open Web Application Security Project put on its first conference on Tuesday. By most accounts, it was a success and we’re likely to have another. I believe there were 150 attendees or thereabouts, which I think is pretty dang good for a few weeks of basic word-of-mouth advertising.
The University was gracious enough to donate the use of the theater in the student center, but we needed somewhere for lunch and were at least lucky enough to be able to stay in the building instead of walking across campus or send people out on their own. So logistics around the space were a little weird. I spent an hour or so of the morning just directing people to the registration desk on the second floor, then back down to for the talks in the theater in the basement (or is it 3rd and 1st floors? whatever). In the end, although it might have seemed mighty strange in the morning, I don’t think anyone minded much.
While I was playing usher, I missed Kuai Hinojosa’s introduction and the first part of Jeff Williams‘s presentation, but I did make it down for most of Jeff’s talk about ESAPI, the OWASP Enterprise Security API. The popular frameworks don’t do nearly enough to guide developers toward building secure software, which is where ESAPI steps in as a set of APIs for building secure web applicationbs, with both an extensible interface and a reference implementation. Right now, the development of the main project is happening in Java, which I know was disappointing to many in the audience who don’t work with Java. But there is an active .NET ESAPI, as well as a less active PHP port, to which contributors are welcome and encouraged. If you are writing Java web apps, you should look at ESAPI now. It’s good stuff. Talks are underway to see about getting some of this in the next servlet spec, which would be fabulous.
Arshan Dabirsiaghi then gave an entertaining and engaging talk about the OWASP Intrinsic Security Working Group, which is a new project trying to get at the heart of the problems in web application security, largely having to do with browser security. Which is a mess. They’ll have their say about HTML 5, too, and will provide input hoping to steer the spec away from security disasters.
I was surprised at lunch to have an actually tasty vegetarian option, a portabello mushroom sandwich on ciabatta. It was probably the bread that sold me. Ciabatta’s the new hotness in bread, after all.
Anil Kumar Revuru from Microsoft spoke about a few things they have going on in the Connected Information Security Group. I had to step out during the first part, so missed much of what he had to say about their framework, but I did catch some of the tools he demoed. Pretty cool stuff, and the Anti-XSS library is a must-have if you do .NET web apps.
I’m torn about the threat modeling tool. On the one hand, it is clearly good work that I’m sure can prove beneficial once a team has worked with it for a while. I believe very strongly that threat modeling is a Good Thing. On the other hand, the threat modeling tool seems extremely heavy-handed, a lot of fiddling with an external application, and I can’t imagine working with a development team that would tolerate it. If I tried to introduce threat modeling with the use of that tool, I’d never get it off the ground. That said, a new version is due out next month, and Adam Shostack is involved. His paper on threat modeling experiences at Microsoft (PDF) was enlightening. So maybe I can hold out hope again. I’ll at least watch the demo available on an SDL progam page.
Brian Chess talked about static analysis and some of the interesting work in that space. I greatly enjoyed this talk, although I can’t remember much of it. :-) He talked about how static analysis tools have evolved, and what you can and cannot do with them. I will say this: a compelling metaphor goes a long way. His saying that “writing secure software is like making safe-to-eat burritos” caught some Twitter-love.
If you liked hearing about ESAPI and CISF, you might also have enjoyed hearing Elliot Glazer speak about the security framework at the Depository Trust and Clearing Corporation, which last year moved $1.86 quadrillion in transactions. Ahem. I hope to get a copy of his slides and post them to the OWASP site, since they were very text-heavy and hard to read but seemed worth reading. The framework seems well thought-out and practical, a bit process-heavy for some but not nearly as bad as you might think it would be, as everything he identified in the process always serves a clear purpose. And has saved them more than once.
Corey Benninger of the Intrepidus Group treated us to real-world phishing examples and trends, and walked us through the discovery of a simple but effective session hijacking attack against a brokerage that cost real money. These are always fun and frightening to see. A great way to round out the regular talks.
We were then treated to an appearance by Richard Stallman. Given fifteen minutes, he explained how free software is an ethical concern. Free-as-in-freedom, of course, not free-as-in-beer. I’ve seen him do this a couple times, and I have to say that he does it well. His talk might not have had an obvious bearing on security, but I’m glad he was there. We don’t talk about ethics enough. I don’t, anyway.
All the talks will be posted online. Stallman’s will of course not be available in Flash, because that is not a free format. For him, expect Ogg Theora.
All in all, I think this was a good conference, I can’t believe we pulled it off and charged only $25 per person! That was no doubt a key factor in getting people to attend. I mean, for twenty-five bucks it’s almost easier to pay out of pocket than to try to convince your boss to pay. ‘Course, for $25 it’s hard to imagine a boss not paying. And to hear these great speakers at that price… marvelous. We were fortunate this time in that the speakers found their own way here instead of having OWASP pay their way. Hopefully we can keep the cost down in the future.
There are things that we can do better in the future. I already mentioned the weird physical logistics, for example. I’d like to see even more along the lines of practical guidance on how to build security into web applications. That’s a core strength of OWASP conferences that I think we should play up as much as possible: they’re for builders more than for breakers. If you look at the agenda, the talks were largely focused on building, but they sometimes got a bit abstract. The only other real concern I had is the perception that OWASP is focused on the enterprise to the exclusion of, well, non-enterprise. That’s not my impression of the organization, but with a focus on Java and .NET in the talks, and with almost no Macs in the audience, it’s an easy impression to give. We have a thriving tech community in the Twin Cities, not all of it so enterprisey, and it would be good for everyone to engage them.
A tip of the hat to Lorna Alamri, who did most of the leg-work for the logistics of actually getting the conference going, and to Kuai Hinojosa, who has done a tremendous job this year growing the chapter and getting the word out about OWASP and web application security. Both had great ideas for the conference, and it really came together. Bang-up job.