Podcasts

Today on Risky Business 2 we speak with Lane Bess, the CEO of Palo Alto Networks. Founded by firewall pioneer Nir Zuk, Palo Alto makes what it calls a next generation firewall.

We don't normally talk to suits like Lane on Risky Business, but hey, that's what this second podcast feed is all about. We thought it would be interesting to get his take on movements in the security market given everything that's happening in world markets.

This week's episode is sponsored by Microsoft and hosted, as always, by Vigabyte virtual hosting.

We're shifting focus a little bit in this week's feature and taking a look at DECT hacking. DECT is the Digital Enhanced Cordless Telecommunications standard, and as you'll hear, it's not always implemented correctly. That can be a lot of fun for the evil guys out there.

Blair Strang will be joining us to talk about that.

Also on this week's show we'll catch up with the host of the PaulDotCom security podcast, Paul Asadoorian. He's popping by to do this week's news segment, and boy, what a week for news it's been.

Microsoft's Internet Explorer product manager, James Pratt, pops by to discuss the new security-related features in the browser in this week's sponsor interview.

If you'd like to comment on anything you've heard on Risky Business, or suggest something you'd like to hear on the show, you can call Sydney 02 8569 1835 or USA +1 877 688 8417 (Toll free).

Normally at these sorts of events protocol dictates that I have a sales department chaperone present at all times to make sure I use the correct fork for the shrimp cocktail, etc, and this was no exception.

My technical colleague and I riffed away, deftly interspersing witty-yet-topical infosec anecdotes with sales patter and doomsaying while we charmed the gathered CIOs with our analysis of the threat insiders posed to their organisations.

Now, you and I know that any sort of insider access is game-fuckin-over, but for the purposes of making the presentation more sales-friendly than a singe powerpoint slide saying, "you're all fucked, plz give us some money while you're still in business," we humoured them.

As I drew to a close, I looked around the audience, fruit platters on the table, a few shunned greasy pastries (they did have bacon, at least) and stewed coffee. I went for my concluding slide -- the last bit of useful information to be shared with the room before the sales drones would activate and attack.

When my sales-chaperone guy saw it he started twitching up the back -- it was off topic and he knows how I roll.

"People sometimes ask me 'Adam, if you were in a room with two dozen CIOs and you could tell them one thing, what would it be?'," I began.

They don't, by the way, but hey -- I get to use any sort of shabby segue I like when I'm clucking on my particular nest. So here's what I'd say.

"Security is hard. It's hard to buy and it's even harder to know if you've bought it. But you have to care, so you hire experts in this arcane field, just as you would any other technical niche. And if the expert says 'your stuff is broken,' then you know where you are. But if they say 'your stuff is great,' then you've got a problem. \t

"Is it really great, or are they awful? Did your expert have a bad day? Is he covering for the fact they just lost half their tech team to a competitor? Did they give you a junior guy, or a box ticker, or even worse, are they out to sell you kit? You don't know, because quality is opaque to someone who isn't an expert here.

"If you take one thing away from today, its that this stuff is hard, and the quality of my work is opaque to you. The only rational choice is not to trust me. So don't hire us. Hire Deloitte, or IBM or whoever you want. But next quarter, pick someone different. Rotate your audit providers. Use one now, another for the audit next quarter, maybe even two different parallel providers on a critical project. Pit us against each other and make us compete. Then at least you have relative quality metrics, which is more than you have at the moment."

Its true, you know -- I'd much rather be going into a pen test or an audit knowing that some high-priced big-5er has already been through with Nessus and Impact, picking off all the dumb shit that just wastes everyone's time to write up. There's no joy in savaging the poor fish in their nesting barrels. (Of course this assumes that the big-5er actually did spot all the low-hangers, which is, uhhh, Not My Experience.)

Yes, your domain controller is still vulnerable to MS06_040. No, you've never patched, your passwords are crap and you have 100,000 clear-text credit cards in /tmp on your RHEL3 box. I'd much rather write up a report about some point of entry that forced me to write a python script to exploit it -- at least then I get to use the Courier font non-ironically.

That's actually the best bit of advice I could ever give a CIO. Please, for the love of God, don't just pick one security supplier. Don't let them cut and paste you the same report every quarter. Get someone else in. Compare the reports, the findings, the quality of the write-ups and mitigation advice and ./sploit.py scripts attached.

Please. Please? I mean, we worked real hard writing them up for you. I know you've only got a 40-minute project meeting, and best to just glance at the summary-table and cross out everything rated less than 'ohmigod'. But please. Get someone else. Don't make me write the same report twice. Let me write a report that I know is going to make some big-5 infosec team look like the boxtickers they are. Please. Let me at them. You know I'm going to find their report on your \\\\fileserver anyway after I MS06_040 your win2k domain controller and have to resist the urge to open it up for the epic lulz that will be within.

Do it for the good of your shareholders. You owe it to them to get a second opinion on something as important as your security. Currently you may not know who is doing the quality work, but it wont take you long to find out -- all you have to do is shop around. You can't tell if you're getting quality, so make us all work to show you.

Hell, maybe we should give discounts to customers who provide us with their previous provider's reports after we've written ours. The lulz would be so worth it. I'll suggest it to salesguy. Well, I would if he wasn't too busy talking scoping with the douche bag who called me a wiseass.

Metlstorm is a New Zealand-based freelance security consultant. He's created several tools including Hai2IVR, Winlockpwn and SSH_Jack. He's also an organiser of the annual Kiwicon security conference in Wellington, New Zealand.

Announced with little fanfare last week by iptables developer Patrick McHardy, the launch of the nftables alpha has barely been mentioned by the press.

That's somewhat surprising, considering the new software will represent the biggest change to Linux firewalling since the introduction of iptables in 2001.

Gordon 'Fyodor' Lyon, the creator of the nmap security scanning tool, says he's excited by the alpha release.

"I'm... looking forward to its general release in the mainstream Linux kernel," he told Risky.Biz. "The previous transitions from ipfwadm to ipchains and then to netfilter (iptables) each brought a new, more powerful firewall interfaces to the user. I expect nftables to do the same."

Administrators who learn the nftables syntax will find it much more expressive and easier to read, Lyon added.

Melbourne-based CSO Adam Pointon says he's surprised the announcement hasn't made more of a splash.

"It's the next generation Linux firewall," he says. "It's a significant milestone and people should pay attention to it."

However, it's not great news for everyone. Iptables and netfilter will be phased out as nftables becomes the norm, Pointon says, which could create some extra work for security appliance manufacturers.

"Iptables is used heavily by lots of UTM products, like routers, DSL modems and the like," he says. "Support will end for that code and everyone will move to nftables. So all the Linux boxes out there using it... will eventually have to re-write all their stuff or wind up using old, unsupported code."

The new firewall has native IPv6 support and userland queuing. "Snort and anything at that layer will be better integrated," Pointon says, adding that nftables will be faster, process rules more efficiently and allow administrators more control at the userland level.

The code base is also significantly smaller. "That can only be a good thing for its security," Pointon says. "It will take Linux firewalling to the next level."

While the alpha release is available now, nftables will go through an extensive beta testing phase before finding itself included in the Linux Kernel.

The group claims the botnet has been targeting DroneBL's servers in a denial-of-service campaign for several weeks and is the first of its kind. It uses brute-force password cracking attempts to hijack any Linux mipsel routing device that uses insecure or common username and password pairs.

"This is the first known botnet based on exploiting consumer network devices, such as home routers and cable/dsl modems," the DroneBL team wrote in a post on their Website. "Action must be taken immediately to stop this worm before it grows much larger."

DroneBL claims many devices are vulnerable to the botnet, which is spreading automatically using compromised hosts to propagate. The size of the botnet, dubbed "psyb0t," is currently unknown.

It's been easy to see why, historically, most Mac users haven't felt the same level of security-related anxiety as Windows users. Until now, no one has really bothered targeting them.

When commentators like this one dared suggest, in 2003, that Apple's OS X software was susceptible to the same sorts of vulnerablities that have plagued other operating systems, the reader reaction was so severe it was worrying.

Indeed, one of the comments posted on the piece by a particularly passionate reader suggested ZDNet's Sydney bureau would make an excellent destination for a truck laden with explosives.

Keep in mind that in 2003 there were few vulnerabilities being disclosed in OS X, leading most consumers to genuinely regard it as more secure. But from there, a trickle of bugs began to be disclosed. By 2008, OS X was giving Windows a real run for its money in terms of the number of bugs being disclosed and patched. The myth of OS X as a "secure" operating systems was destroyed among the more savvy types in the IT industry, and Apple dropped its rhetoric about its operating system's amazing invulnerability to malware.

Yet in the years since the malware never showed up. Sure, anyone with half a clue could trigger a client-side exploit in OS X, but what then? The science of writing Trojans for Windows-based operating systems is mature; staff at CERT teams and AV companies have actually found comments and evidence of revision control in modern PC malware.

Mac malware has been primitive in the extreme by comparison -- the bad guys just haven't built up their OS X chops yet.

Last year, news of simple script-based Mac malware doing the rounds surfaced. The badware would simply alter the user's DNS settings, so it was pretty simple stuff. Some may argue that's actually pretty serious -- if an attacker can control their target's DNS, a man-in-the-middle hack is trivial, thanks to browser insecurity (Hi, Safari!). Still, this early Mac malware was hardly what you'd call sophisticated.

But now we're seeing some much, much nastier stuff. Risky.Biz forwarded a recently obtained Apple malware sample to two parties -- Paul Ducklin at Sophos (disclaimer: Sophos is a sponsor) and a contact who'd prefer not to be named.

Paul had seen that sample before, and Sophos's products detected its payload. But it was what the other had to say that I found particularly interesting.

His analysis indicated the sample -- which pops up as a flash installer on, err, "video sites" -- may in fact automatically trigger upon download. How? Well, every time Safari downloads a file with a DMG (Apple disk image) extension, it will auto-mount it when the download's complete.

That's really handy, but also a security issue, especially when you remember that there have been buffer overflow vulnerabilities in the code OS X uses to mount DMG disk images. So if a user hadn't patched against the DMG overflow, all they'd have to do is click "ok" to a bogus Flash installer notification, served from the domain apple-updates.com. OS X would do the rest for you.

My contact couldn't be 100 percent sure the sample was trying to trigger the DMG bug, but even the possibility should give us pause; it would mean the badware is getting much smarter.

To be fair, Windows still does some similar, super-daft things. The Conficker malware is currently spreading left right and centre because it's basically impossible to disable autorun in Windows without resorting to a registry hack.

The payload in the Mac malware sample in question was a 'dloader,' tasked with connecting to some shady data centre in Eastern Europe and downloading more bad stuff.

This is much more sophisticated than a script that just alters some DNS settings. It's closer in sophistication to the malware we've been seeing targeting PCs for the last 10 years.

Interestingly, we haven't seen this dloader actually grabbing a payload yet. That tells me these guys haven't bothered actually writing a serious Trojan yet -- they've just sent the first stage of the attack out there to see how many bots they wind up with.

If they get enough, undoubtedly they'll actually create some "real" malware for it, and begin distributing it to pre-infected hosts.

So that's it folks. Mac malware has arrived, and what a party it's going to be. Most Mac users are convinced they're using a magical, impenetrable platform, so they don't actually use antivirus software. Apple's advertising campaigns of yesteryear actually encouraged that mentality. Combine that with Apple's expanding market share, and the average Mac user is now a very tempting target. A sitting duck, if you will.

Enjoy the next couple of malware free months, Mac users, because you're in for a rough ride in '09.

Patrick Gray is the managing editor of Risky.Biz and the host of the Risky Business security podcast.

The online environment is just like the real world, yet for some reason many consumers completely abandon their street smarts the second they fire up their browsers. When a leather-clad, toothless ruffian is walking up and down the street saying "give me $500 and I'll come back in an hour with a computer worth $1000," everyone knows not to trust him. Yet this is the same premise by which many scams, such as online auction fraud, are perpetrated.

The success of online criminals is harming consumer confidence.

In late 2008 I released the findings of the Consumer Trust and Confidence Online Survey [pdf] which was aimed at determining the level of trust and confidence of Australian Internet users within the online environment. The survey focused on e-commerce, social networking and online safety.

There were some interesting results. For example, 35 percent of respondents were more trusting of online transactions than two years ago. That sounds great until you realise 65 percent were either less trusting or had the same level of trust as two years prior.

Considering the increasing value and importance to the Australian economy the Internet plays, these statistics should ring alarm bells for anyone with a vested interest in online commerce.

Let's dig a little deeper.

The two most important factors considered by survey respondents when considering purchasing goods and services online was the reputation of the merchant and the payment method. Now we have some actionable information that tells us organisations must boost their reputation to bolster consumer confidence. Here's how:

• Be transparent -- give honest and open responses to customer questions and feedback
• Be flexible -- recognise change in systems and behaviour and implement swiftly
• Establish a reputation system -- it's a popular feature for eBay transactions
• Reflect reality - customers (and the media) are smarter than you think [They sure are.. ;) -- ed], they can sniff out a fake quickly.

Which leads into payment methods. While plenty of organisations abide by the Payment Card Industry Data Security Standards, some just don't. Media reporting of e-commerce organisations that have been compromised with the loss of customer credit card and personal information is a weekly occurrence.

But it's not just targeted hacks that are causing problems, there are far simpler forms of fraud. Consumers have proven willing to send payment for non-existent goods to unknown beneficiaries in international destinations via money transfer systems like Western Union.

Why do consumers engage in this risky behaviour? Maybe it's because online consumers are usually at home in a relaxed and comfortable environment where they can't see the normal visual cues that make us suspicious. Like the guy who's trying to sell you the Blu-ray player is covered in prison ink and has no teeth.

In a real world transaction their radar is far better attuned to detecting the potential for fraud.

The successful integration of e-commerce into the Australian economy is dependent upon the level of trust and confidence consumers have in the digital environment.

Developing new kinds of commercial activities utilising the Internet hinges on assuring consumers that their use of networked services is secure and reliable, that their transactions are safe and that they will be able to verify information about transactions and transacting parties. There are too many organisations that have a commercial interest in establishing customer trust and confidence in online technologies for this not to be taken seriously.

Nigel Phair was the Team Leader of Investigations for the Australian High Tech Crime Centre from 2003 to 2007 and the author of Cybercrime: The Reality of the Threat. He is an active cyber crime analyst.

This week's podcast is brought to you by Tenable Network Security and hosted, as always, by Vigabyte virtual hosting.

It's a special day for us at Risky Business HQ -- we've launched our new Web site: http://risky.biz/

We now publish two podcasts, video and written news and opinion. There's also forums, so by all means go and sign up for an account! We'll see you in there.

On this week's show we're talking to L0pht/@stake/Veracode co-founder Chris Wysopal about the rebirth of L0phtCrack, the legendary password cracking package.

In this week's sponsor interview, Tenable Network Security analyst and Open Security Foundation dude Brian "Jericho" Martin pops in for a chat about dataloss -- are you more likely to lose data through a USB key, lost laptop or an actual attack?

Adam Pointon also pops by for a look at the week's news.

In this first post in our fresh new RB2 podcast feed, you'll hear Krusher's presentation to the second Kiwicon conference in New Zealand.

It was recorded in September 2008.

H D Moore has also done some interested work with wardialling. You can hear him discuss his work on WarVOX here.

Sure, maybe its not 1994AD any more. But let me posit this, which I culpably dub Metlstorm's Assertion:

The cost of owning a corporation is a fraction of a percent of their annual infosec spend.

Lets go with 0.1%. Can you think of any organisation you've worked for, or on, or with, or pwned that you couldn't own for the sales margin on a single Check Point device?

Let's assert the value of owning a corporation -- if you're any good at the order-fulfillment bits of crime, which I'm not -- is proportional to its market cap.

The ratio of cost-of-ownership to value-of-ownership is so low as to have an ROI to an attacker that is nearly infinite.

Stated more concisely (unusual for me, I know); the incremental cost to an attacker between not hacking you and hacking you is so close to zero we have to assume they actually do.

Which means you should proceed on the assumption that your corp is already owned.

We live in a world where our desktop machines get USB autorun worms, where a garden or variety botnet worm owns entire Ministries of Health, where insider attacks are commonplace, where biometrics doesn't work, where routers are backdoored by offshore manufacturers with various political goals, where we pay janitorial services staff minimum wage because they've only got physical access to, well, everything via their trivially clonable RFID proxcards running on building management software off a crappy old NT4 box in the basement. Ok Metl. Breathe.

You see where I'm going with this. There is no infosec industry. We're just doomsayers who take the chumps money while they've still got it, and when they don't we just scare the next lot senseless until someone pays up. We don't actually improve anything.

The infosec industry is a trinity; the boxpushers (vendors), the chumps (the users), and the doomsayers (us, the pentesters).

Boxpushers sell kit to the chumps, who've been goosed into thinking they need it. The doomsayers occasionally pity the chumps, but are generally stuck in io-wait, writing off the boxes being pushed as useless, impractically complex, and that highest criticism of all; boring.

Us doomsayers take the chump's money, then tell them in excruciating and savage detail how much they and the boxes they got pushed suck.

And they invariably do.

When we're on a typical gig we sit around, amusing ourselves intellectually by doing something we'd all probably just do for fun anyway, call it work, and then tell the chumps in serious sounding language quite how poked they are today.

There is doom. Unending grimness. Like the darkened frostbitten forests of Ukranian blackmetal album covers.

Hell, in the case of boxpushers, they actually make it worse (Hi mail antivirus gateways! Hi IDS consoles, hi shatter-prone desktop asset management and patch deployment solutions, giving up localadmin like [security researcher] Brett Moore slipped you his best Mr December smile under the digital cyber eMistletoe.)

I ask you again -- is there any corporation you've seen where the upper bound of cost to own them wasn't proportional to the janitor's hourly rate? We all know, deep in our guts, that we could own anyone. And we wouldn't be doing it with Ben Hawkes' heap technique -- that stuff's for impressing cons and talking shit in bars, not wasting on actual attacks. We'd just roll like it was 1994AD; and we'd win. Every time. You know it. And how much would it cost? To own a bank, a telco, an ISP, a critical infrastructure provider? Really, we all know the turgid, sodden, doomladen truth.

How much would it cost?

Yeah. Exactly. Fractions, my man. Fractions of a percent.

Metlstorm is a New Zealand-based freelance security consultant. He's created several tools including Hai2IVR, Winlockpwn and SSH_Jack. He's also an organiser of the annual Kiwicon security conference in Wellington, New Zealand.