Skip to content

Month: August 2008

GoDaddy’s spam filter is broken

GoDaddy is rejecting mail with URLs that appear in the Spamhaus PBL. As this thread on the Amazon EC2 forum notes, this is creating false positives, causing nonspam mail to be rejected. Here’s what GoDaddy reportedly said about this policy:

Unfortunately, our system is set to reject mails sent from or including links listed in the SBL, PBL or XBL. Because the IP address associated to [REMOVED] is listed in the PBL, any emails containing a link to this site will be rejected. This includes plain-text emails including this information.

If this is true, it’s utterly broken.

Spamhaus explicitly warn that this is not to be done, on the PBL page:

Do not use PBL in filters that do any ‘deep parsing’ of Received headers, or for other than checking IP addresses that hand off to your mailservers.

And more explicitly in the Spamhaus PBL FAQ:

PBL should not be used for URI-based blocking! Consider the false positive potential: legitimate webservers hosted with services such as dyndns.com or ath.cx! Or consider that ISPs and other networks are encouraged to list any IP ranges which should not send mail, and that could include web servers! Use SBL or XBL (or sbl-xbl.spamhaus.org) for URI blocking as described in our Effective Spam Filtering section. Use PBL only for SMTP (mail).

Critically, the PBL now lists all Amazon EC2 space, since Spamhaus interpret Amazon’s policy as forbidding email to be delivered via direct SMTP from there. (Note — email, not HTTP.)

With this filter in place at GoDaddy, that now means that if you mail a URL of any page on any site hosted at EC2 to a user of GoDaddy, your mail won’t get through.

Note: this is much worse than blocks of SMTP traffic from EC2. In that case, an EC2 user can relay their legit SMTP traffic via an off-EC2 host. In this case, there is no similar option in HTTP that isn’t insufferably kludgy. :(

Links for 2008-08-28

The real reason cycling is such a pain in Dublin

Cian Ginty at the Irish Times writes:

As clunky helmets, yellow reflective gear, and Lycra could be used as a stereotype for Irish cyclists, it might come as a surprise that women wearing high heels are a common sight on bicycles in Copenhagen.

The general image of cycling here is vastly different to so-called bicycle cultures where cycling is normalised and there is talk of a “slow bicycle movement”.

“Among thousands and thousands of cyclists on my daily routes, I think I see one or two reflective vests a week, if that,” says Mikael Colville-Andersen, a cycling advocate living in Copenhagen.

With Denmark, the Netherlands and Germany – where bicycle usage is high – the helmets and reflective clothing we think of as “a must” for cyclists are far from standard.

It then goes on to rehash some of the stuff that has cropped up recently on cycling blogs about cycling safety, helmets, etc.

The only problem with casualization of cycling, removing gear like helmets, is that without corresponding changes to the road and cycleways to make them safer, it will increase accidents and fatalities. I looked this up a couple of weeks back when I came across an anti-helmet site. Chasing up the figures and doing some research, it became clear that if you simply want to cycle without hurting yourself, the facts were not on their side — helmets save lives, especially when dealing with shared roadways as we have here.

Copenhagenization is a result of a better, safer road environment for cyclists, as seen in Denmark and the Netherlands, which makes safety gear not as much of a requirement. But on the other hand, Ireland’s roads are designed mainly for cars, and Dublin Council have done little to help — that makes safety gear a requirement, unfortunately :(

However, I think this is the real reason why people don’t cycle in Dublin:

Let’s take a fictional person, let’s call her Kassandra. Kassandra lives a little north of Copenhagen and rides every to work every day between 07:25 and 07:55 and back again between 15:35 and 16:05. Kassandra doesn’t mind a little light showers, but if the intensity increases to over 0.4 mm over 30 minutes (light rain), then she thinks it is too wet. Kassandra works five days a week and has weekends and holidays free. That gives her 498 trips between September 2002 and the end of August 2003.

How often does Kassandra get wet either to or from her job that year? The answer is, in fact, rarely. On those 498 trips it was only 17 times. That is only 3.5% or on average 1.5 trips a month.

3.5%. Compare that with what’s happened in Dublin this month — I’d estimate that’s meant that at least half of my rides have involved some degree of rainfall, occasioning many cries of woe.

It takes dedication — and lots of wet-weather gear — to ride a bike here…

(Of course, having said that, I look out the window and it’s immediately sunny ;)

Update: Ryan Meade corrects me in the comments:

Justin, you need to take a look at Owen Keegan’s paper to Velo-City 2005, “Weather and Cycling in Dublin : Perceptions and Reality”. The probability of getting wet is actually pretty comparable to the Copenhagen scenario detailed above – 5.5% for a 30 minute journey if you take 0.2mm per hour at the threshold for “getting wet”. On the other hand the vast majority of both cyclists and motorists think it’s more than 15%, with half thinking it’s above 30%.

Amazing how the psychological, “glass half-empty” factor influences my thinking on this. I had no idea!

How tightly linked are the top spam botnets?

I was away on holidays last week, and when I got back, I found my feed reader full of some good discussion as to whether today’s bigger spam botnets — Srizbi, Rustock, Mega-D, Cutwail/Pushdo — are sharing components, such as “landing” sites, exploits, customers, and even command and control networks. It started with this post on the FireEye Malware Intelligence Lab’s blog noting:

‘Some malware researchers have described Srizbi and Rustock as rival botnets, our data indicates that this apparent rivalry is a sibling rivalry at best. Srizbi and Rustock seem to be supported (controlled) by the same parent (bot herder).’

and in this followup:

‘We can clearly see that Srizbi, Pushdo and Rustock are using same ISP, and in many cases, IPs on the same subnet to host their Command and Control servers. It seems extremely unlikely to our research team that three previously “rival” Botnets would share nearly consecutive IP space, and be hosted in the same physical facility. Of all the data centers and IPs in the world, the fact that they are all on the same subnet is very intriguing. This fact makes the FireEye research team conclude that either the Botnets are operated by the same organization, or that the datacenter (McColo) is a shell corporation that leases out it’s IP space and bandwidth for nefarious actions.’ […]

‘IPs at a typical datacenter are leased out in a /30 or more commonly, a /29 block. However, here we can see that in a given succession of IPs, the three Botnets have C&C servers dispersed throughout. This gives us an impression that same Bot herder leased out a larger range and then distributed it amongst its different Botnets.’

Marshal say: ‘at the very least, the major botnets have common customers.’

Dark Reading cover it like so:

Rustock, which recently edged Srizbi for the top slot as the biggest spammer mostly due to a wave of fake Olympics and CNN news spam, and Srizbi, known for fake video and DVD spam, have been using the same Trojan, Trojan.Exchanger, to download their bot malware updates, researchers say. “This is the first time” we had seen this connection between the two botnets, says Fengmin Gong, chief security content officer for anti-botnet software firm FireEye. “That’s why when we saw it, it was surprising. They definitely have a relationship,” he says. “There’s not the rivalry we used to think about.” […]

Joe Stewart, director of security research for SecureWorks, says the Srizbi-Rustock connection is most likely due to a spammer using both zombie networks — not that the operators of the two botnets are actually collaborating. “What is confusing people is that you’re seeing Rustock bots sending out emails that essentially infect people with Srizbi, so they think it must be Srizbi that’s sending it, but it’s not,” he says. “Srizbi is not just one big model. It’s rented out to lots of different spammers.”

A major spammer may be trying to diversify by using the two botnets, he says. “It could be because they want to separate their malware-seeding operation from their spamming operation,” Stewart says. “Maybe their bots are getting blacklisted faster when they’re sending out URLs with fake video files because they’re easy to spot, so their spam doesn’t get through. So they send malware from this botnet, and spam from this one, to keep out of the blacklists longer.”

I agree that Joe’s scenario is very likely; the spammers aren’t always the same people who operate the botnets, and it only makes sense that some of them would spread their business among multiple nets, to minimize the risk that all of their output would be blocked if one ‘net runs into trouble (or indeed, good filtering ;). But seeing C&C servers sharing LANs also strikes me as unusual. One to watch.

Anyway, it’s good to see that the malware research blogs are now actively tracking and posting updates when the botnets change topics and format; this info is very valuable for us in anti-spam, as it allows us to map from the received spam mails back to the sending botnet, and determine which rules are good at detecting each botnet. Thanks, guys.

(image credit: cobalt123, used under CC license)

Links for 2008-08-21

My Omnivore’s 100

Here’s my results for The Omnivore’s Hundred, a silly foodie “purity test”. Bold are items I’ve eaten; crossed-out items are ones I wouldn’t eat again. I score 70 out of 100, and clearly need to eat less Asian and more European cuisine ;)

  1. Venison
  2. Nettle tea
  3. Huevos rancheros
  4. Steak tartare
  5. Crocodile
  6. Black pudding
  7. Cheese fondue
  8. Carp
  9. Borscht
  10. Baba ghanoush
  11. Calamari
  12. Pho
  13. PB&J sandwich
  14. Aloo gobi
  15. Hot dog from a street cart
  16. Epoisses
  17. Black truffle
  18. Fruit wine made from something other than grapes
  19. Steamed pork buns
  20. Pistachio ice cream
  21. Heirloom tomatoes
  22. Fresh wild berries
  23. Foie gras
  24. Rice and beans
  25. Brawn, or head cheese
  26. Raw Scotch Bonnet pepper
  27. Dulce de leche
  28. Oysters
  29. Baklava
  30. Bagna cauda
  31. Wasabi peas
  32. Clam chowder in a sourdough bowl
  33. Salted lassi
  34. Sauerkraut
  35. Root beer float
  36. Cognac with a fat cigar
  37. Clotted cream tea
  38. Vodka jelly
  39. Gumbo
  40. Oxtail
  41. Curried goat
  42. Whole insects
  43. Phaal
  44. Goat’s milk
  45. Malt whisky from a bottle worth $120 or more
  46. Fugu
  47. Chicken tikka masala
  48. Eel
  49. Krispy Kreme original glazed doughnut
  50. Sea urchin
  51. Prickly pear
  52. Umeboshi
  53. Abalone
  54. Paneer
  55. McDonald’s Big Mac Meal
  56. Spaetzle
  57. Dirty gin martini
  58. Beer above 8% ABV
  59. Poutine
  60. Carob chips
  61. S’mores
  62. Sweetbreads
  63. Kaolin
  64. Currywurst
  65. Durian
  66. Frog’s Legs
  67. Beignets, churros, elephant ears or funnel cake
  68. Haggis
  69. Fried plantain
  70. Chitterlings or andouillette
  71. Gazpacho
  72. Caviar and blini
  73. Louche absinthe
  74. Gjetost or brunost
  75. Roadkill
  76. Baijiu
  77. Hostess Fruit Pie
  78. Snail
  79. Lapsang souchong
  80. Bellini
  81. Tom yum
  82. Eggs Benedict
  83. Pocky
  84. Tasting menu at a three-Michelin-star restaurant
  85. Kobe beef
  86. Hare
  87. Goulash
  88. Flowers
  89. Horse
  90. Criollo chocolate
  91. Spam
  92. Soft shell crab
  93. Rose harissa
  94. Catfish
  95. Mole poblano
  96. Bagel and lox
  97. Lobster Thermidor
  98. Polenta
  99. Jamaican Blue Mountain coffee
  100. Snake

(thanks to this generator and mordaxus at Emergent Chaos for the link.)

Links for 2008-08-20

Links for 2008-08-19

Links for 2008-08-18

Irish gang skims 20,000 bank cards through retail POS terminals

Wow, this is pretty massive. The Irish Payment Services Organisation has again released details of a credit-card breach, this time on retail Point-of-Sale card terminals. Quoting the Irish Examiner story:

Una Dillon, head of card services at the Irish Payment Services Organisation, said the criminals went into the shops pretending to be doing maintenance work on behalf of the banks.

“We have discovered only in the last 48 hours that a number of retailers have been affected by a point-of-sale compromise,” she said. “We are in the lucky position that it was discovered quickly and the gardaí are working on it.

“Gardaí have uncovered a lot of the devices and CCTV footage. We have a list of all the card numbers that have been used. They have either been blocked or restrictions put on those cards.

“With the devices recovered it may just be that the cards were only saved and the criminals did not have a chance to get hold of the card numbers.”

“There will be an emergency meeting today with the gardaí, the terminal vendors and the banks to try and close down on this,” she said, adding the gardaí were in pursuit of the gang.

Insufficient authentication of maintainance staff is being blamed:

“The criminals have been going into shops claiming to be engineers working on the terminals. Staff are used to their bank officials coming to update terminals so unfortunately they have been able to do that.

Bank of Ireland estimated 3,100 of its debit and credit cards were affected and Ms Dillon said the other eight card providers could have similar numbers.

Bank of Ireland said, as a temporary measure, it had reduced the daily withdrawal limit on all its debit cards for ATM transactions outside Ireland to just €100 to protect customers from fraud.

They haven’t released the names of the affected shops yet; 20,000 cards, though, sounds like it’s been going on for while, on a large scale. Yikes.

The SiliconRepublic story claims that the gang ‘plugged in wireless devices that pushed the data to the internet and allowed the card numbers to be used overseas.’ However, in the past, these ‘wireless devices’ didn’t use the internet; instead they use parts from mobile phones, which relayed PINs, card numbers and CVV security codes via SMS text messages to Romania. That model seems more likely here, I would guess, due to the reliability of phone networks.

Update: last night’s RTE Six-One news bulletin (viewable as streaming RealVideo or transcoded 5MB AVI file), made it clear that the hardware used phone components and SMS. It had some pretty good pics of what appears to be a sample subverted POS terminal:

VISA have been warning about attacks on petrol-pump-based POS terminals since 2006, e.g. this story, but they’re more easy to attack since there’s few or no staff present by the pumps when the POS terminal subversion takes place. This has resulted in most petrol stations in Ireland disabling POS credit-card payment systems, requiring customers to pay at the counter; we lose convenience, but at least we’re probably not being skimmed. But these in-store POS terminals seem to be increasingly under attack; there are reports from Livigno, Italy, Rhode Island, and Canada.

The tamper-proofing of POS terminal hardware is unreliable; it’d be nice to see them made harder to tamper with. I would guess the gang used secondhand, hacked POS terminals, which supposedly should be tamper-evident (ie. easy to spot modifications).

Better yet, if Chip-and-PIN cards used end-to-end crypto between a crypto smartcard and the bank’s central systems, POS hacks would be impossible. But there’s no sign of that happening.

Most importantly, IPSO has promised that ‘banks will refund any customers whose details have been used to make fraudulent transactions.’ That’s key. It’s interesting to note that IPSO have been hammering home this point repeatedly in their stories — they’re worried about customer confidence, I’d guess.

Links for 2008-08-14

Links for 2008-08-13

Links for 2008-08-12

Links for 2008-08-11

Links for 2008-08-10

On Threading

Luis Villa, in a post to FoRK:

I have found that [mail] threading is overrated, in part because I’ve realized that any conversation so baroque as to actually require threading probably isn’t worth following.

Even though I wrote a threader for MH, I have to admit by now that he has a point ;)

AppEngine — only useful for toys

Noted on Twitter:

simonw: So apparently http://www.news.com.au/ used json-time for their Beijing countdown widget and blew my App Engine quota! They’ve stopped now.

uh, great. That’s useful.

Google — how are we supposed to host useful services with those limits?

Links for 2008-08-07

Links for 2008-08-05

Why San Francisco’s network admin went rogue an “eyewitness account” with allegations about the SF network admin in question — no documentation, passwords kept to himself instead of shared with his team, the entire network maintained by 1 person, never took holidays, bad tempered and stubborn — sounds like a recipe for classic BOFH disaster

Yehrin Tong Illustration cool, hyper-detailed hand-drawn tiling patterns

working around installation bug in File::Scan::ClamAV running the test suite results in “ERROR: Can’t open/parse the config file clamav.conf”; looks like File::Scan::ClamAV is now unmaintained :(

ALIFE Conference to reveal bio-inspired spam detection ‘this bio-inspired spam detection algorithm, based on the cross-regulation modeal of T-cell dynamics, is equally as competitive [sic] as state-of-the-art spam binary classifiers and provides a deeper understanding of the behaviour of T-cell cross-regulation systems.’

Aaaand we’re back

Due to unfortunate scheduling, taint.org had some downtime there as it moved to a new server. It’s back now though.

Some of the other services running on taint.org, jmason.org, yerp.org, twit.ie, planet.spam.abuse.net and so on, are still intermittently offline as I bring them back up in their new home… so have patience, they’ll be back soon. ;)