« first day (853 days earlier)      last day (4325 days later) » 
00:00 - 19:0019:00 - 00:00

00:00
@TildalWave So now there'll be even more monkeys taking 'selfies' on instgram
@D3C4FF hehe nice one :) can't upvote any more tho you rep #!@*$ :)))
Although running that query against both Sec.SE and SF show that, at least on the first page worth, SF has higher averages.
LOL that reads as tweet me not at starbucks ... sry
@TildalWave DAMNIT. I've passed 1337 rep!! ;__;
DAMN YOU UPVOTES!!! DAMN YOOUUUUU!!!
00:07
@D3C4FF I can always take it back now that you edited :P Just say the word :))
@ScottPack interesting
@D3C4FF xD? strike dat? Was that the word? :P
@TildalWave The word was bird. Cause bird is always the word. w3rd.
@ScottPack I'm somewhere in the middle, as expected (I've tried to answer some of the old unanswered questions... which isn't the best thing to do if rep is wanted)
@D3C4FF Yes, by The Trashmen :)) So... I trash your upvote? :P Be clear damn it!
@ScottPack Interestingly, you're above both bears in terms of average score on your answers
@TildalWave Clarity is only for clairvoyants
And sales people.
Certainly the two most clear groups of people.
00:17
@D3C4FF beware of the clairvoyant salespeople!
I thought all salespeople were clairvoyants, thats how they KNOW their product will stop APTs!
Google actually produces some funny stuff with "clairvoyant salespeople" search
00:50
3
A: Risks of giving developers admin rights to their own PCs

DKNUCKLESThe security risks associated with providing developers the ability to install their computers are numerous. Here's why I would object (speaking as a sys admin) 1) Potential violation of best security practices - One of the 8 rules of security is the rule of least privilege - only give employees...

Currently the most controversial post on the site (10 up / 7 down votes)
@TildalWave lol
01:05
Does anyone here use the contagio malware dump site?
01:16
@D3C4FF me no, but it's a good source of malware if you ever need one
I wonder, who holds the copyright to that black macaque's self-portrait?
@TildalWave quick. Tell me how long you think it'd take to do a 'quick' VA against 30K hosts?
@D3C4FF I guess it depends on what you mean by quick VA
Depends on your available resources
And the deliverables as outlined within the scope of work
And your mom.
@ScottPack yeah that too
Do you have a single box with a single core P4 and 4GB of ram with OpenVAS or do you have 40 boxes with 8 cores and 8-16GB of ram with Security Center?
Do you have login creds for the in scope hosts?
01:30
Do you use analogue modem? :)))
This is my corporate connection
user image
3
Single I7 on a laptop, 8GB RAM, Wired gb NIC, no creds, hypothetically lets say we only look at vulns on 20 ports on each host.
@ScottPack I've actually got one of those at home... :D
Oh jesus
@D3C4FF How many months do you have to do the job?
We scanned lets say 15k machines, and I might add by scan I mean PING sweep. And THAT took 4 days.
So lets say AT best 4K machines pinged per day. Lets say it takes just as long to ping a machine as it does to scan a single port on that machine. That means we’re looking at the better part of 245,756 days if we’re scanning every single port on 4k machines. 20k machines… in theory? Approximately three thousand years? I’m hoping my math is a bit off there.
Ping sweep is bad news bears. Starting with XPSP3 ping was blocked by default.
01:32
@D3C4FF what scans do you do? just banner info might be pretty fast... but complete VA... another story
Your better option there is to do a ping sweep, on each broadcast domain, and pull your live hosts from the arp table.
@ScottPack Sorry, 'ping sweep' = nmap scan with flags i can't recall off the top of my head. But it basically looks at more than just ICMP replies.
With a few Nessus scanners slightly better than yours, and credentials, I've been able to do sweeps of about that scale in a few weeks.
@ScottPack Few weeks= more than 4 weeks?
Hey gus, if someone's cracking my wirless AP password, can I know the attacker's location?
01:35
@D3C4FF 3-4
@ScottPack man -_-
Fuck this scoping shit.
@warl0ck Depends on the hardware you have available. With the right kit you can triangulate their position, but it'll require several beacons.
@warl0ck somewhere in the vicinity then... I guess you could use a heavy shield of metal or similar and then see at what direction it kicks him out, so that's your vector then... then take a baseball bat and start walking in that direction :)
Oh, and take out all the antennas but one from your router
@D3C4FF And that was with credentials using Nessus. In that setup it optimizes the scans, checks to see if the WMI ports are open, tries to connect, if so it grabs the open ports from netstat. Otherwise it starts doing portscans
@ScottPack IIRC that guy later realized it wasn't the right car? :))
01:40
@TildalWave Yup
Love that flick. Haven't seen it in a while.
@ScottPack Thanks, that's good to know actually.
@TildalWave hah funny
@D3C4FF Yeah, if Nessus can log into the system it won't bother with a network port scan. Saves a shitload of time.
@ScottPack it's nothing more than a normal cisco router
01:43
@warl0ck mrr?
@D3C4FF Is this you as a contractor doing a job or you as an employee for a company?
@ScottPack Contractor, we're trying to scope the work for a proposal
@D3C4FF Are you using any kind of automated tools other than nmap?
Nmap, nessus, Metasploit and Acunetix is the usual tool kit
McAfee foundstone sometimes as well
(foundstone is suprisingly quick and accurate)
Have you looked at Tenable's Perimeter service?
Yeah but isn't that external only? These are internal systems.
External VA/PT was completed already (1k hosts)
01:57
Ah, ok. It is external, which is a downside. The upside is that it's pretty cheap and will probably scan faster than your laptop. :)
So are you doing a VA or a PT?
@ScottPack VA
VA internal (but only a VA cause of the size)
@ScottPack I haven't used it before but its not a bad idea for future engagements.
I dunno, I think a full PT with social and physical testing using one guy on a 30k system site is totally doable.
3
@ScottPack Yeah, i'm actually a robot, don't require sleep or food and I can rapidly multiply.
@D3C4FF It's basically just a hosted install of Nessus. Fun Fact: That service has been stamped as a ASV.
That's lewd.
@ScottPack in a good or a bad way?
02:04
meh
@D3C4FF Might I suggest, during your next client meeting, you kindly request credentials for the scans? I mean, it sounds like your in a world of butthurt no matter what, but you might as well ease it in if possible.
@TildalWave Isn't it like 4 in the morning?
@ScottPack yup
Seriously.
 
1 hour later…
03:16
@ScottPack That sounds fun.
 
1 hour later…
04:35
@ScottPack You don't understand... These guys... They want it black box ;____;
 
4 hours later…
08:46
@StackExchange As usual, xkcd makes a very good point.
 
4 hours later…
12:32
@TildalWave All the answering of old questions just to get the 200 sesame to gold tag badges is taking its toll. It is hard to get upvotes out of cold corpses.
12:54
so, whose for mars-one?
13:16
@D3C4FF That's fine. Annoying, but fine.
@D3C4FF Personally, from a VA, I would want both views. Run your tools with creds on the inside of our network (to get a worst case realistic view) and then run your tools without creds on the outside of the network (to get an environmental world view).
Isn't a VA usually just running some scanning tools and drawing conclusions from the results? Should be easier than a pentest.
@D3C4FF But they're the customer. Just so long as the customer and contractor agree to common terms during the scoping document so that each side knows what the other is expecting. :)
@TerryChia It is. A VA is effectively a report that lists all of the potential vulnerabilities found. A pentest will do that, then attempt to exploit them.
@TerryChia Take, for example, the backported patch problem.
RedHat will backport security fixes to versions of software they actually want to support. That way a fix that was released in Apache httpd 2.4.23 may be extracted and patched into RedHat's release of 2.2.3 (what they support for el5).
A simple banner grab, which is what a non-credentialed scan will get you, might show your webserver running 2.2.3 which is all kinds of vulnerable to all kinds of things.
Your penetration team can then attempt to exploit that laundry list of vulnerabilities, and probably fail.
After all, the holes had been patched.
Similarly, a credentialed scan would log into the box and see, "Oh, this is running RedHat's version of 2.2.3 which I know has patches for all of these holes. So I'm not going to report them."
@TerryChia Make sense?
@TerryChia Oh, here. I did a whole big writeup on it.
10
A: Fresh Graduate Going into Penetration Testing

Scott PackA Little Background Penetration testing probably feels like the sexiest part of security, but it is also a very small part. It's also exceptionally broad. The term "penetration test" is often used as an umbrella to mean any one of the following: Vulnerability Assessment Security Assessment Sec...

@ScottPack Well yeah. I actually knew that before. :P Well, most of it.
</selfpromotion>
@ScottPack +1 because you made the effort. That's rare.
13:26
I think most admin types can intuitively understand the difference between credentialed, non-credentialed VA and pentest. Really it's all about just putting the pieces together.
At least, most of the admin types, I've talked to get it pretty quickly once you point out what the connections are.
While we don't do any PenTesting a big part of my job is Vulnerability Assessment and Management. I really enjoy that stuff.
What I don't enjoy is decoding the nasl to figure out why a plugin is generating a false positive.
Recently we found out that the applied vendor patch, for a specific software product, had 5 layers of version number. That is, version 1.2.3.4.5, whereas the plugin only checked 4.
Clearly it was vulnerable because version 1.2.3.4 was vulnerable. It couldn't tell that we had 1.2.3.4.5 installed.
Fucking developers.
@ScottPack You know I gotta star that...
Hey man, whatevs.
Sometimes I feel like half the things I say in here are for the star lulz
@ScottPack haha, 'Hot or not'
@D3C4FF I couldn't resist. :)
And cheers for the views on the engagement
13:37
This is the other thing that came to mind.
@ScottPack That's hot.
@ScottPack That's hot.
@D3C4FF No worries. I mean, I feel a little disingenuous spouting off like that. I've never negotiated a pentest contract or lead a team or anything useful.
I just have impotent opinions on the matter.
@ScottPack Fair enough, its still an opinion outside of my team which was certainly worth listening to, so cheers :)
Hopefully we'll win the work and i'll be travelling around for the next month and a half :P
Yay me!
13:51
does anyone on here play ingress?
@lynks 90% of the room I believe.
Cheeri-o all
@D3C4FF Bye :)
@AntonyVennard well thats good, I just got my invite. One thing I dont understand is the tactical purpose of fields. I mean they add to the team's global score, but do they have any effect on the enemy within the field?
@lynks The Rories, Neal, myself.
The only tactical purpose I know of a field is that you cannot generate a field inside of, or crossing the boundaries, of an existing enemy field.
13:55
@lynks I got the invite but I'm lacking an android phone. :(
@ScottPack so creating a field doesn't destroy all enemy portals inside it?
@lynks That's correct.
@TerryChia You can play on your Nexus
shame :P that would make massive fields epic lulz
@ScottPack My Nexus 7 is wifi only. I don't think it's possible considering the type of game?
13:57
I think the fields are more part of the metaplot
@TerryChia Depends on your area's wifi coverage
@TerryChia turn your 3G phone into a mobile hotspot :P
and connect to that on the nexus
@lynks I have no idea, I don't play it at all!
@lynks heh. i'm actually finding an excuse not to play.
i don't need to get hooked on yet another game.
especially one that i can play walking around.
@TerryChia the fact that I cant play it from home or from work (no portals nearby) means it isn't a time-sink or distraction. it mostly just gives me a reason to venture into the city once in a while
@AntonyVennard I saw your comments on that selinux answer.
I don't really see how MAC can be done in a more secure way by default than what already exist.
To be fair, selinux is great if you bother to properly configure it. which is a mighty pain in the ass to be fair.
14:02
@Terry - I have a couple of friends who use the Nexus for its screen and GPS and tether it to their phone for the data connectivity.
Yeah, trying to use my phone is a bear. The screen is small and the GPS sucks.
@TerryChia These days selinux requires almost no custom configuration.
@TerryChia Basically, selinux is a) useless for X11, since by definition you have to let X applications connect to the X server (which runs as root and lets you do things like dump the x screen out, including all other content) and b) quite a lot of fedora runs as unconfined_t, a domain where it is allowed to do anything. That's most desktop apps. It's a different case for system services which are nicely locked down.
To be fair - that selinux answer is totally irrelevant to the question.
But to be more fair - the question kinda misses the point of the comic.
As I understand it there are several ways people have tried to overcome this: nested X (Xephyr), rewriting X11 entirely to do security properly (Wayland) and patching what exists (X Access Control Extensions a.k.a. LSM for X11, letting you create SELinux-type policies for X apps)
Yeah, X11 really needs to be replaced soon.
14:06
hmmm, im trying to come up with an API definition, is it massively antipattern to ask API users to pass JSON in a GET variable?
@TerryChia Actually, all mention of OS security/application isolation is irrelevant. That's to protect applications from doing things they shouldn't. Browsing to facebook and typing "I am an idiot" into the status field is not an application misbehaving, it's just a user misbehaving.
@AntonyVennard Or the user totally behaving. I'm not sure which is worse.
@TerryChia Well... in any case the only solution to the problem at hand (allowing me to leave my laptop unlocked with everything signed in) would probably make UAC look like a dream.
@AntonyVennard A laptop that shocks the user every time something stupid is being done would be a great idea.
Training chimpanzees like @ThomasPornin will say.
@lynks I think maybe a big enough object might exceed the maximum url request size?
I don't think I'd do it - I'd use POST/PUT depending on the action it represented.
14:13
@AntonyVennard yeah you're probably right, ill go with something nice and POSTy
@lynks In REST terms, one of them means create and the other means update. I can never remember which way around it is. But remember your http status codes, you can do things like 201 created, 204 deleted and so on to have proper error reporting.
As well as the 403 no!, 401 who are you?, 405 what? You make no sense error codes.
Are you people still going on about MAC?
@ScottPack Naw, some REST-y programmy shit.
That sounds worse.
@ScottPack No doubt.
14:18
i hate web
just in general
but i have to define an API that makes some sense. and it has to take reasonably complex objects as input : /
I am personally much, much happier inside a debugger, but web isn't too bad every once in a while.
@ScottPack That's a +1 ;)
only one upvote left for the day :( damnit I sometimes feel like SE is rationing my daily reading allowance
Great Success!
@TildalWave You trying to get that gold badge?
14:34
Don't forget to upboat questions in sufficient quantities!
@TerryChia Taking advantage of slow business days and going through the questions that tickle my fancy ;)
@ScottPack I always do that, It's a reflex thing by now. If I see a good answer I upvote the question to promote the answer (well, hopefully)
@ScottPack This makes me really nervous
@TildalWave Not all questions with great answers deserve upvotes.
No, although as a rule I think any question I answer deserves an upvote. It is very rare that I would answer a question that I don't think is somehow deserving.
For me to be interested enough to answer than it either has to be something I find interesting, or something I'm happy somebody asked.
@ScottPack That's cus you aren't a repwhore like most of us.
14:39
@ScottPack How else are you going to catch the cheap rep-trains, though?
@Iszi I've only ever jumped on one reptrain, thank you very much.
@TerryChia I'm not sure I agree with that. I follow different rules for voting on questions and answers. Answers get my upvote if they were thorough, thoughtful, or in any other way helpful to me or the OP. Questions get an upvote, if they enabled that to happen.
@TerryChia That's probably true. :) I tend to put a lot of effort into my answers, even the short ones tend to take a long time. I like to test and validate what I'm posting.
@ThomasPornin Agreed on that, tho I personally don't mind it that much. I can't afford to take Sec.SE (or any other SE sites so far) as a rep thing, I'd be utterly dissapointed most of the days then. Plus, I'm subscribed to the newsletter, and there's this "can you answer these questions" section that I sometimes can't ignore ;)
@TildalWave You just need to put in slightly more effort to repwhore.
It has gotten easier since @Polynomial got his new job.
14:48
Yep. What we need now is to find @ThomasPornin a nice comfy new job with lots of responsibilities and voila!
Of course, the downside is that we'd lose all those awesome answers, so I'd actually rather not...
Ah see, StackExchange needs to offer @ThomasPornin a nice comfy new job writing answers.
4
@ScottPack that "that's hot" anim. I'm glad it's pushed way up now, made me feel really uncomfortable. With such peeps I'm always anticipating their cranium to implode any minute now. Just makes me... uneasy.
@TildalWave River had the same concern with Book because of his hair.
-1
Q: Is Using MD5 Sufficient Reason to Reject This Payment Processor?

David Alan HjelleI'm evaluating a credit card processor[1], and I noticed they are using MD5 as part of a salted hash algorithm to protect a secret key. Since I know MD5 is generally considered broken, this feels like a poor solution. Is that enough of a criteria to reject them as our processor? Is it possible to...

This looks odd to me.
15:03
it might be time to upgrade my work machine when I have to turn off font antialiasing in openoffice to get it to scroll smoothly.
From what he describes, the hash is used to verify the integrity of the contents he sends over. But the hash is sent in the same form as the contents?
@lynks heh. what crappy piece of junk are you using?
@TerryChia a venerable old machine of my own. AMD64 3200+ with 1gig of RAM :)
@ScottPack speaking of .nasls and false positives have you come across one recently with IPSEC VPNs and agressive mode?
@lynks Maybe pump in a few more gigs of RAM? That should help things. Unless it's DDR2 ram, which is crazy expensive now.
@RoryMcCune I haven't, but I don't see many of those.
15:07
@TerryChia its original, DDR ram, I have 3 gig sat beside me on the desk, but for some reason the machine will not boot with anything other than 1 stick :(
all sticks pass memtest on their own
@ScottPack we're seeing it quite a bit as a finding in nessus and the other tool we use for the scanning ike-scan doesn't show it...
as soon as i add a second, memtest is sad
@lynks woah. upgrade time!
@RoryMcCune Have you filed a report yet?
I'm betting even a tablet has almost as much computing power. ;)
15:08
@TerryChia yeah its old, im only using it at the office because the company cant afford to buy me my own machine.
@lynks I think I said it a while back. You need a new job.
@ScottPack not yet, I'd need to nail down a test case well enough to confirm. Unfortunately only come across it on short-term externals so I don't have access to the target system long enough to play about with it...
@TerryChia yep ive been saying it for months :P i have some applications in the works!
@RoryMcCune Ah, that sucks. In theory the logs, the .nessus output and the kb should be enough to diagnose. Testing a fix wouldn't be possible, though.
@ScottPack yarr downside to the life of a consultant.. when I get some time I'll set up an IPSEC VPN and see if I can replicate the problem...
15:10
Good luck with that.
Truth.
15:31
@TerryChia I'm actually as confused as you are about this whole request signing thing.
Have you found something about that?
@Adnan Yeah I got it figured out. I was just confused.
@TerryChia Rain your knowledge all over me then
@ScottPack What's the average price of a 15cm sandwich at subway?
@Adnan Like the guy said in the answer's comments. It's to prevent MITM attacks from tampering with a single form value.
@Adnan No clue. I haven't been to to a Subway in a few years. You want me to go check the website and get back to you?
Hmm actually... why don't they just use SSL.
Guess I have not thought it through yet.
15:34
@TerryChia Actually it doesn't make any sense. If the MiTM can see the request and potentially tamper with it, the he has all the information to compute a new signature, am I wrong?
@ScottPack Oh, sorry. Forgot about the website, I'll check it.
@Adnan Presumably the signature is recomputed at the server side. This is assuming that the attacker does not know the "secret key"
Ahh yes. It makes sense now.
@ScottPack I can't get to the U.S. website, it's automatically redirecting to the Finnish one. Would you please check it for me?
It's to prevent attacks from forging new payments in the name of a particular merchant.
@ScottPack It doesn't have any prices
@TerryChia What? How does that make sense? If the signature is computed server-side, what is it being compared against?
15:38
@TerryChia We can give them all GUIDs!
I personally look forward to naming my child {2312BEE3-2B98-41A3-854C-2EBF7CE8550C}.
@Adnan The signature being sent over.
If you noticed, the secret key isn't in the forms.
It only did not make sense to me because I had thought the secret key was being sent over the form as well.
@TerryChia Then how is client signing the request?
If it doesn't have the secret key, which is used for signing
@Adnan If you look carefully at his scheme, the client concats the secret key together with the form values and send it over as the "tamper proof seal"
The client has the key, the client just isn't sending over the key to the server.
@Adnan Let's say $4-$5 USD
And yeah, like The Bear's answer states, just use HMAC.
15:45
@TerryChia I get it now
@Ladadadada ping
@Iain Opened up your email to reply to it and got distracted learning about Flattr.
@TerryChia sack?
15:47
/me enjoys being flattered
@Iszi of flour?
@Ladadadada ;)
@TerryChia for Algernon?
@Iszi CCCCCCCOMBO BREAKER!
Wow
Thanks @Iain.
15:49
@ScottPack always a pleasure
I think I need a drink after all that.
@TerryChia That's why I'm still Ladadadada everywhere. In 14 years since I invented this username I've only once run into a place where it was already registered and I think that was me anyway and I'd just forgotten about it.
@TerryChia Mwahaha, I shall be back to annihilate you all! :P
I was going to make an old joke, but we're the same age. :(
@Polynomial Go away.
15:56
@Ladadadada You didn't try the password reset function? That's the first thing I do when I find an Iszi already registered. I wonder how the other Iszis in the world have reacted to that?
@Iszi if i receive an unsolicited password-reset email, i go into full DEFCON 1 combat ninja mode of 1337ness and panic.
Weirdly, I sometimes get password reset emails in my ladadadada@free-email-provider accounts. Recently, a Facebook one got so insistent that I had to reset the guy's password, log in and delete his account.
@TerryChia :o :(
@Polynomial :D
I can't really write an answer to this, but this is like expecting your wife to have sex with you after she discovers you've met with a divorce attorney. — Mike Brown 12 mins ago
Just brilliant.
16:19
True tho
Doh ignore me I'm having one of those 1=1 moments
must be that gif @ScottPack posted working its magic :)
You're welcome.
I'm still shocked after yesterdays business meeting... two guys from same client came to my place to discuss the project's status then became involved with themselves what it's actually expected of me... that was productive, two hours wasted... if I could demand they return the coffee I made them, I would
16:38
So I bought this.
It's a bit bigger than the VGA dongle I already had. A bit bigger than expected.
17:06
We had something like this one here, right?
9
Q: Forcing users to use a particular pattern for passwords

SalmanMany websites, specially government ones, enforce users to use a password which conforms to particular criteria. For example: use between 8-13 characters with at-least one integer and one capital letter and sometimes a special character. I have found such password patterns are rather hard to foll...

@ScottPack coolio does it work well? I've got a VGA one at the moment, but that could come in handy if I run into a projector that does HDMI...
ya know it takes a special company to not just have default credentials for their web interfaces but to actually put them on the login page, just in case you couldn't remember admin/admin ?!?!
17:23
if I ever actually program my self out of a Job I would put it on my resume and double the amount I charge. -- Chad Mar 28 at 13:00
@RoryMcCune that is deeply magical
@RoryMcCune given that metasploit scanners try a password equal to the username first, and that 'admin' is right at the top of my usernames-wordlist, that login would require precisely 1 attempt.
@lynks happen to have seen it on two consecutive tests. On one of the two occaisions the product actually automatically logged me in. Looked in Javascript and they had hardcoded an admin/admin login to the page
the mind boggles
Anyone experiences with giving very short presentations?
@RoryMcCune I haven't plugged it in yet. I wanted a DVI output, I only had a VGA before. I liked how dirt cheap this thing was and that it had the extra outputs was just a bonus.
@LucasKauffman Define "very short"
17:47
@LucasKauffman flashing is illegal in my country
5 minutes
@LucasKauffman that's pretty short. I'd recommend having one point and stick to covering that, don't make it too wide ranging and avoid any fluff round the edges.
Some of the cons are doing that now. I think they usually call them Lightening Talks.
@ScottPack yea I think I better look up how to do one of those and how to prepare the slides
18:04
@LucasKauffman if you want something a bit different from PowerPoint you could try Prezi
@LucasKauffman In Crypto conferences, this is a typical duration for Rump Session talks. Rump Session is for presenting new results which have not yet turned into a full article.
Also used by Biham to demonstrate the extent of his grasp of the word "humour".
@RoryMcCune And give everybody vertigo?
@RoryMcCune I suggested prezi but one my colleagues doesnt like it at all :P
@ScottPack rly? I like it :) lends itself to non-linear talks..
I've never seen it used reasonably.
18:07
@ScottPack one of the presentations I did that went down the best was a Prezi one. prezi.com/oqsovr2gzie-/it-security-prepare-to-fail
I'll take a gander at it later. I'm trying to do data analysis.
@ScottPack surely that's easy. the data says that we need more money for security :)
3
A: Why do some websites force users to come up with difficult passwords?

Tom LeekWhen a site implements something, it is always to serve their interests; however, this sometimes coincides with serving the user's interests as well, e.g. as in "protecting the users against themselves", which is the case here. The point of the restrictions is to (try to) make the password harde...

@RoryMcCune Oh, this is for thesis work.
@ScottPack ahh
18:10
@RoryMcCune I've got alert files for plain old snortd and using my system.
@ThomasPornin I've linked your answer there.
@RoryMcCune So now I get to compare them for efficacy.
@ScottPack cool so the thesis is on a new type of IDS?
@RoryMcCune Not exactly. It's effectively a wrapper around snort. We got rather tired of having to disable potentially useful signatures for performance reasons. Sigs that seldomly fired, were exceptionally expensive to run, but were very strong indicators.
@ScottPack interesting...
18:14
@RoryMcCune So it's largely a system to dynamically run different signatures against selected addresses.
@RoryMcCune In the end it's not much more than a perl script to fiddle with a database of addresses and generate BPF files along with a special snortd conf.
@RoryMcCune That being said, it seems pretty novel and seems to work. Here's hoping.
@ScottPack so how do you select the addresses to target? Is that based on the results of other plugins?
@RoryMcCune I tried to make the input mechanism pretty generic so that how we got the addresses could be similarly generic. I'm thinking correlation rules from a SIEM and/or analyst instinct.
@RoryMcCune It's not that often one of us will be watching flow traffic and think to ourselves, "Something's...not....right..."
random aside I'm listening to Gotye at the moment and I can't help hearing the lyrics from the star wars version.
2
in case anyones not seen it.
18:25
So, yeah, while I don't expect this work to change the world, it's at least good enough for a defense and that's good enough for me.
@ScottPack well yeah it's like this isn't it? matt.might.net/articles/phd-school-in-pictures
@RoryMcCune Pretty much. :) All too often that perspective seems to be lost.
@ScottPack well when it's something you spend so much time on, it'll inevitably feel big/important/scary :)
meh
The hard part's actually been figuring out how to correlate the data for review.
I've been building out a csv where each row is keyed off the signature id, and contains a field for each source and destination where each contains a list of the respective IP addresses for that sig.
@ScottPack CSV, come on that's not webscale.. you need mongodb or some other nosql store :oP
18:37
@RoryMcCune That's just for the analysis. I'm using sqlite for the actual tracking. Bitch.
@ScottPack but but that's not webscale either!! mongodb-is-web-scale.com
2
this is rather disturbing and funny
@RoryMcCune why not use hadoop I heard it runs in the cloud!
2
@ScottPack if you want to do correlation you should try apache lucene
@LucasKauffman what's disturbing is how they are even not trying to polish out auto-key (or whatever that tech to make bad singers appear to be singing in key) anymore... it's now an expected part of every single hip-hop crap they release
18:42
@LucasKauffman I'm trying to get finished, not learn new devopsian tactics!
@Lucas, yes
@ScottPack maybe you should use a DB2 on a z/series, I heard it's quite feasible according to IBM.
Had to give a half hour presentation to an FD once. She gave me five minutes. ..
@LucasKauffman Shuddit
Not ideal, but fun.
18:43
@RoryAlsop first thing I want to avoid is putting too much info on the slides
@RoryAlsop did you speak a lot faster?
mmm what do you guys I should avoid?
@LucasKauffman If you have only five minutes you should avoid just about everything.
You need about 2 minutes for the audience to get up to speed; you can make them grasp at most one idea a minute, except "Big" ideas which take two or three.
Basically you can express one or two core ideas in your talk, little more.
@ThomasPornin I think that's the point
that's why they are only giving 5 minutes probably
so I should keep to high level description
but emphasizing all the positive things
@LucasKauffman what's your topic?
5 minutes is enough for one description, one idea and one joke. Be sure to choose the joke with great care, that's what they will remember.
@RoryMcCune well basically they had a competition where you could present an idea for a new service, we ended up onto the shortlist and that's what we need to present to the Dragon's Den. They will then decide who has the best idea and that service will be built out
18:54
@LucasKauffman interesting. so it's like a VC pitch. Are they doing Q&A after the presentation? that'd probably be the hard bit if so..
@RoryMcCune yea, we're preparing that as well by asking people to prepare difficult questions
it's 5 minutes presentation and 5-10 minutes QA
00:00 - 19:0019:00 - 00:00

« first day (853 days earlier)      last day (4325 days later) »