« first day (66 days earlier)      last day (5112 days later) » 

00:00
And...he's gone. WTG Avi.
@AviD The uid/gid of a process in unicees can be changed mid-run, but to the best of my knowledge the binary must implement the change itself.
@packs shorrry.... looks like chat.se can detect malicious activity and then bluescreened me....
I take that back, anything running with euid 0 (particularly the kernel) should be able to do so, but I'm not aware of any userspace mechanisms to do it.
well, actually it was stupid flash. I'm starting to agree more and more with steve jobs, about banning flash from istuff...
@AviD np, I'm still catching up on that transcript.
@packs be able to, but does it?
windows apps can do that to, but the question @Iszi had was does it do it by itself
I didnt think so...
00:06
I think I figured out that the reason I had so much trouble understanding Iszi's issues with how that question was being handled is that I've internalized those aspects of access control. I wasn't really finding a way to see it from his perspective
@packs Im curious, didja just bang on the keyboard for a while, or is that really an acronym...?
You can actually really see it in action for group membership changes.
@AviD That was a total keyboard bang.
I have no respect for overly long acronyms.
@packs aww shuckhs, you disappointed me...
was hoping i could get some MILSWANCA points....
I'll turn it into a backronym if you would like....
Anyway, back to groups and such in unicess.
@packs right, and I realized that he hadn't - and thats why it looks like a bug.
00:08
If a group membership change occurs, it only takes effect on all new login sessions.
It really is an interesting set of tradeoffs there, though... especially if you're seeing it with fresh eyes
Without question.
@packs right, as I thought - and same as windows.
I feel like the whole thing might be more prevalent in the Windows side because of kerberos.
that could be...
00:09
That's just a stab in the dark, I can't really justify it right now
hmm, just noticed a small difference (I think....)
new process, existing login.... inherits token, or gets new token for new process?
Should inherit. Not token based, remember?
brb
right, "token" whatever....
okay, so no difference...
00:33
Although....
Since there isn't a security token/ticket, it won't ever get refreshed.
So changes to any environment based access restrictions will only take affect by restarting the process
So I retract what I said earlier about prevalence.
hello again
@RoryAlsop!
@packs!
:-)
@AviD - ahh - haven't looked at fortify for ages, but yeah - trying to get back up to speed as have a few months of code review work coming up :-)
Now we just need to find a bar tended by a former pitcher.
@RoryAlsop does it have to be with fortify?
I've actually been using it a lot lately - customer requirement
to my great frustration
00:40
yeah - that bit already seems to be agreed :-)
what alternatives do you use over there?
aww
well I am a big fan of checkmarx
dont have my own license, though...
too expensive
checkmarx - will look at that one
ahhh
and, my biggest code review client insists on fortify
the awesome thing about checkmarx - it's infinitely extensible.
Fortify - they do seem to have market share, and they provided decent chocolate at RSA :-)
it exposes the internal graphs through a very scriptable platform
hehe
I gotta tell you, I've done comparisons numerous times
and EVERY time fortify comes up very short
I think the only real thing it has going for it is coverage
that, and HP is practically giving the licenses away
00:44
yeah
I will have to see if I can get a test play with Checkmarx then
but at least it's gotta be simpler than trying to review x million lines by hand :-)
@RoryAlsop Why in my day, we printed it all out on green bar and took the red pen to it!
exactly!
And I loved...seeing the looks on those kids faces when they got it back
The Akamai talk concentrated on just why that isn't possible any more with large firewall rulebases (the 50k plus rules) - you could see so many wistful faces remembering how it used to be :-)
I'll bet.
00:49
The Cisco dude did a great piece at the Qualys booth on why scanning your IP range doesn't work so well in IPv6
similar problem - sheer scale
Shit, I can show you where it doesn't work so well in v4
hahahaha
very true
The numbers on that slide scare me
I know! - who understands Undecillion anyway
I've never heard the word 'nonillion' before.
00:53
yeah, htye must have had fun putting that slide together....
@RoryAlsop one caveat you should know going in - I wouldnt be so interested in running it out of the box. Expect to invest some time in writing the queries you want...
not so much finetuning, but really writing what you want to find. Automating your brain, so to speak....
@RoryAlsop do you do a lot of codereviews?
yeah - the way timings are right now I probably won't get the leisure, we don't do a huge amount of code review
if the demand is there, we can do most things that come under the infosec category, and I used to do a fair few code reviews, but this is the first since October
Anyway - gotta go sleep before the little ones get me up at unfeasible o'clock to go swimming. Night guys.
fun fun
gnite
@RoryAlsop enjoy
Code reviews are a service that we don't really offer right now, but my boss and I feel pretty strongly that we need to.
is there a lot of custom development by you, @packs?
It's just a shame, because I'm the best we have on staff to figure out how to do something like that.
We (university IT) have absolute gobs of custom development.
01:01
@packs you dont sound enthused....
you could consider outsourcing it, instead....
I'm not a programmer :)
@packs ah, well...
then definitely consider outsourcing it :)
To be fair, I'm a few months actual work from finishing my MS in CS. So I have a lot of exposure and a fair bit of experience, but since it wasn't something I ever enjoyed then I never really got hot and heavy into it.
The biggest problem with outsourcing are funds. We're a public institution, and the state's not been looking so hot lately.
so, offshoring then? ;)
We couldn't afford your rates :)
01:04
lol
Nice try
meant those cheap workers over in india and far east, but whatevs...
On the plus side, almost all of our custom work is in java.
me is always a good option, though....
eh. plus == java, why? compared to what?
So at least they've minimized the language set.
01:05
ahh, yeah, that a plus.
I'm not putting forth that java is any better or worse a language. Just that we're looking at a single one.
yes, absolutely.
Our department didn't really exist until middle of 2006, so we've been ticking off those boxes as they've trickled to the top.
Your guys chat about Fortify has just reminded me that this one is getting there :)
NOT fortify, i hope!
I'll keep that resounding endorsement in mind.
01:09
heh
I should probably either look at the main site, or formulate this into a good Question, but really I would be encouraged to check out free and/or open source tools first.
as it says in the sidebar - "Fortify Sucks."
I think there are a couple questions around that...
Money for things like that can sometimes be found, but I can see funding code review software being very difficult to convince.
but I dont think youd have much like getting quality tooling on that.
On a free budget, you mean?
01:11
There is a project at OWASP - but it seems to be slowish. Lets see now after the latest summit...
@packs yes
So far, we've had good luck with free tools, or at least getting something going with free tools.
For example, we spent a few years with spider, and as a result got funded for a bunch of Identify Finder licenses. That's been working out very well.
I meant, specifically wrt code review tools.
there are some halfway decent free ones, if all you want is glorified grep
you could even find a couple free that do source-sink flows
but thats not really all I would look for
but, to be fair, fortify doesnt do much more than that, anyway :D
And the only review tool I have experience with is splint, which, as you said, isn't much more than a glorified grep.
Hey, sweet. OWASP has updated their website. And it looks suspiciously more usable. Hm.
01:31
dunno, looks same unusable...
prettier, though
Well, that's what happens when your website is a wiki :)
heh
ok, my turn to turn in - tomorrow's a work day for me :)
Well, tomorrow will be a work day for me as well. Different suit, different location, and a boss that I have to sleep beside every night :)
g'night
@packs ah, but you get to choose all your team :)
(I assume you're talking about your wife - otherwise you might want to get some counseling....)
Heh, yes, home chores and the wife. Although, if most tv shows are to believed, hospitals work that way as well.
Speak of the devil... Guess the kid's asleep. HangWithTheSpouseTime. ta
01:41
gnite
 
12 hours later…
14:08
Mornin
14:25
@Iszi!
@packs!
Now that I've acknowledged your existence. Time to start cleaning.
Yay weekends!
@packs Meh
Cable guy still hasn't showed. And we were promised an 0800 appointment this time, despite the general window of 0800-1000.
15:35
Finally, he's here.
 
2 hours later…
17:16
Nice, thanks guys.
I pretty well got fed up with the whole 2 hour window thing and started asking for a call when they were on their way.
Living 5 minutes (7 with traffic) from work really helps
18:09
@Iszi You get your stuff sorted?
18:21
hello can any one tell me abt security certifications
which is the best IT security certification ??
That's a pretty subjective question right there :)
Do you want policy, DR, generic, audit, technical, etc?
 
4 hours later…
22:09
@packs They came a half-hour past their window, and the job took nearly 6 hours, but it's done.
@mr_eclair Yeah, perhaps a question to be discussed in chat if at all - don't ask it on the site.

« first day (66 days earlier)      last day (5112 days later) »