Phishing still on the decline, despite Whois privacy

The number of detected phishing attacks almost halved last year, despite the fact that new Whois privacy rules have made it cheaper for attackers to hide their identities.
There were 138,328 attacks in the fourth quarter of 2018, according to the Anti-Phishing Working Group, down from 151,014 in Q3, 233,040 in Q2, and 263,538 in Q1.
That’s a huge decline from the start of the year, which does not seem to have been slowed up by the introduction in May of the General Data Protection Regulation and ICANN’s Temp Spec, which together force the redaction of most personal data from public Whois records.
The findings could be used by privacy advocates to demonstrate that Whois redaction has not lead to an increase in cybercrime, as their opponents had predicted.
But the data may be slightly misleading.
APWG notes that it can only count the attacks it can find, and that phishers are becoming increasingly sophisticated in how they attempt to avoid detection. The group said in a press release:

There is growing concern that the decline may be due to under-detection. The detection and documentation of some phishing URLs has been complicated by phishers obfuscating phishing URLs with techniques such as Web-spider deflection schemes – and by employing multiple redirects in spam-based phishing campaigns, which take users (and automated detectors) from an email lure through multiple URLs on multiple domains before depositing the potential victim at the actual phishing site.

It also speculates that criminals once involved in phishing may have moved on to “more specialized and lucrative forms of e-crime”.
The Q4 report (pdf) also breaks down phishing attacks by TLD, though comparisons here are difficult because APWG doesn’t always release this data.
The group found .com to still have the most phishing domains — 2,098 of the 4,485 unique domains used in attacks, or about 47%. According to Verisign’s own data, .com only has 40% market share of total registered domains.
But new, 2012-round gTLDs had phishing levels below their market share — 4.95% of phishing on a 6.83% share. This is actually up compared to the 3% recorded by APWG in Q3 2017, the most recent available data I could find.
Only two of the top 20 most-abused TLDs were new gTLDs — .xyz and .online, which had just 70 attack domains between them. That’s good news for .xyz, which in its early days saw 10 times as much phishing abuse.
After .com, the most-abused TLD was .pw, the ccTLD for Palau run by Radix as an unrestricted pseudo-gTLD. It had 374 attack domains in Q4, APWG said.
Other ccTLDs with relatively high numbers included several African zones run as freebies by Freenom, as well as the United Kingdom’s .uk and Brazil’s .br.
Phishing is only one form of cybercrime, of course, and ICANN’s own data shows that when you take into account spam, new gTLDs are actually hugely over-represented.
According to ICANN’s inaugural Domain Abuse Activity Reporting report (pdf), which covers January, over half of cybercrime domains are in the new gTLDs.
That’s almost entirely due to spam. One in 10 of the threats ICANN analyzed were spam, as identified by the likes of SpamHaus and SURBL. DAAR does not include ccTLD data.
The takeaway here appears to be that spammers love new gTLDs, but phishers are far less keen.
ICANN did not break down which gTLDs were the biggest offenders, but it did say that 52% of threats found in new gTLDs were found in just 10 new gTLDs.
This reluctance to name and shame the worst offenders prompted one APWG director, former ICANN senior security technologist Dave Piscitello, to harshly criticize his former employer in a personal blog post last month.

Could crypto solve the Whois crisis?

Could there be a cryptographic solution to some of the problems caused by GDPR’s impact on public Whois databases? Security experts think so.
The Anti-Phishing Working Group has proposed that hashing personal information and publishing it could help security researchers carry on using Whois to finger abusive domain names.
In a letter to ICANN, APWG recently said that such a system would allow registries and registrars to keep their customers’ data private, but would still enable researchers to identify names registered in bulk by spammers and the like.
“Redacting all registration records which were formerly publicly available has unintended and undesirable consequences to the very citizens and residents that electronic privacy legislation intends to protect,” the letter (pdf) says.
Under the proposed system, each registry or registrar would generate a private key for itself. For each Whois field containing private data, the data would be added to the key and hashed using a standard algorithm such as SHA-512.
For items such as physical addresses, all the address-related fields would be concatenated, with the key, before hashing the combined value.
The resulting hash — a long string of gibberish characters — would then be published in the public Whois instead of the [REDACTED] notice mandated by current ICANN policy.
Security researchers would then be able to identify domains belonging to the same purported registrant by searching for domains containing the same hash values.
It’s not a perfect solution. Because each registry or registrar would have their own key, the same registrant would have different hash values in different TLDs, so it would not be possible to search across TLDs.
But that may not be a huge problem, given that bad guys tend to bulk-register names in TLDs that have special offers on.
The hashing system may also be beneficial to interest groups such as trademark owners and law enforcement, which also look for registration patterns when tracking down abuse registrants.
The proposal would create implementation headaches for registries and registrars — which would actually have to build the crypto into their systems — and compliance challenges for ICANN.
The paper notes that ICANN would have to monitor its contracted parties — not all of which may necessarily be unfriendly to spammers — to make sure they’re hashing the data correctly.

ICANN joins anti-phishing group board

ICANN’s VP of security has joined the board of directors of the Anti-Phishing Working Group.
Dave Piscitello is one of three new APWG board members, arriving as the group expands its board from two people to five.
APWG said the expansion “is recognition of the growing complexity and scale of Internet crime today and the challenges in responding to this global threat.”
In a press release, it noted that targeted phishing attacks are said to be the root cause of the data thefts that may or may not have influenced the US presidential election last year.
The other two new directors are Brad Wardman of PayPal and Pat Cain of The Cooper Cain Group, a security consulting firm (a different bloke to the similarly named Pat Kane of Verisign).
APWG is an independent, public-private coalition that collects and publishes data about phishing attack trends and advice for how to defend against them.
Part of this work entails tracking how many domain names are involved in phishing, and in which TLDs.
The APWG board also includes chair David Jevans of Proofpoint and secretary-general Peter Cassidy.

New gTLD phishing still tiny, but .xyz sees most of it

New gTLDs are not yet being widely used to carry out phishing runs, but most such attacks are concentrated in .xyz.
That’s one of the conclusions of the Anti-Phishing Working Group, which today published its report for the second half of 2014.
Phishing was basically flat in the second half of the year, with 123,972 recorded attacks.
The number of domains used to phish was 95,321, up 8.4% from the first half of the year.
However, the number of domains that were registered maliciously in order to phish (as opposed to compromised domains) was up sharply — by 20% to 27,253 names.
In the period, 272 TLDs were used, but almost 54% of the attacks used .com domains. In terms of maliciously registered domains, .com fared worse, with over 62% share.
According to APWG, 75% of maliciously registered domains were in .com, .tk, .pw, .cf and .net.
Both .tk and .cf are Freenom-administered free ccTLDs (for Tokelau and the Central African Republic) while low-cost .pw — “plagued” by Chinese phishers — is run by Radix for Palau.
New gTLDs accounted for just 335 of the maliciously registered domains — 1.2% of the total.
That’s about half of what you’d expect given new gTLDs’ share of the overall domain name industry.
Twenty-four new gTLDs had malicious registrations, but .xyz saw most of them. APWG said:

Almost two-thirds of the phishing in the new gTLDs — 288 domains — was concentrated in the .XYZ registry. (Of the 335 maliciously registered domains, 274 were in .XYZ.) This is the first example of malicious registrations clustering in one new gTLD, and we are seeing more examples in early 2015.

XYZ.com aggressively promoted cheap or free .xyz names during the period, but APWG said that only four .xyz phishing names were registered via freebie partner Network Solutions.
In fact, APWG found that most of its phishing names were registered via Xin Net and used to attack Chinese brands.
But, normalizing the numbers to take account of different market shares, .xyz shapes up poorly when compared to .com and other TLDs, in terms of maliciously registered domains. APWG said:

XYZ had a phishing-per-10,000-domains score of 3.6, which was just slightly above the average of 3.4 for all TLDs, and lower than .COM’s score of 4.7. Since most phishing domains in .XYZ were fraudulently registered and most in .COM compromised, .XYZ had a significantly higher incidence of malicious domain registrations per 10,000 coming in at 3.4 versus 1.4 for .COM.

APWG said that it expects the amount of phishing to increase in new gTLDs as registries, finding themselves in a crowded marketplace, compete aggressively on price.
It also noted that the amount of non-phishing abuse in new gTLDs is “much higher” than the phishing numbers would suggest:

Tens of thousands of domains in the new gTLDs are being consumed by spammers, and are being blocklisted by providers such as Spamhaus and SURBL. So while relatively few new gTLD domains have been used for phishing, the total number of them being used maliciously is much higher.

The number of maliciously registered domains containing a variation on the targeted brand was more or less flat, up from 6.6% to 6.8%.
APWG found that 84% of all phishing attacks target Chinese brands and Chinese internet users.
The APWG report can be downloaded here.
UPDATE: XYZ.com CEO Daniel Negari responded to the report by pointing out that phishing attacks using .xyz have a much shorter duration compared to other TLDs, including .com.
According to the APWG report, the average uptime of an attack using .xyz is just shy of 12 hours, compared to almost 28 hours in .com. The median uptime was a little over six hours in .xyz, compared to 10 hours in .com.
Negari said that this was due to the registry’s “aggressive detection and takedowns”. He said XYZ has three full-time employees devoted to handling abuse.

Phishers prefer free ccTLDs to new gTLDs

Domains in free and cheap ccTLDs are much more likely to host phishing attacks than new gTLDs.
That’s one of the conclusions of the latest report of the Anti-Phishing Working Group, which found that Freenom’s re-purposed African ccTLDs were particularly risky.
The first-half 2014 report found 22,679 “maliciously registered” domains used in phishing attacks. That’s flat on the second half of 2013 and almost double the first half of 2013.
Only roughly a quarter of the domains used in phishing had been registered for the purpose. The rest were pointing to compromised web servers.
On new gTLDs, the APWG said:

As of this writing, the new gTLD program has not resulted in a bonanza of phishing. A few phishers experimented with new gTLD domain names, perhaps to see if anyone noticed. But most of the new gTLD domains that were used for phishing were actually on compromised web sites.

The new gTLDs .agency, .center, .club, .email and .tips were the only ones to see any maliciously registered phishing domains in the half — each had one — according to the report.
The APWG speculates quite reasonably that the relatively high price of most new gTLD domains has kept phishers away but warns that this could change as competition pushes prices down.
While .com hosts 54% of all phishing domains, small ccTLDs that give away domains for free or cheap are disproportionately likely to have such domains in their zones, the report reveals.
The Freenom-operated ccTLDs .cf (Central African Republic), .ml (Mali) and .ga (Gabon) top the table of most-polluted TLDs, alongside PW Registry’s .pw (Palau).
Freenom, which also runs .tk, offers free domains, while PW Registry has a very low registry fee.
APWG measures the risk of phishing by TLD by counting phishing domains per 10,000 registered names, where the median score is 4.7 and .com’s score is 4.1.
.cf tops the charts with 320.8, followed by .ml with 118.9, .pw with 122, .ga with 42,9 and .th (Thailand) with 27.5. These number include compromised as well as phisher-registered domains.
Read the APWG report here.

Phishing domains double in 2013

The number of domain names registered for phishing attacks doubled in the first half of the year, according to the latest data from the Anti-Phishing Working Group.
The APWG identified 53,685 phishing domains, of which 12,173 are believed to have been registered by phishers. The remainder belonged to compromised web servers.
This 12,173 number — up from 5,835 in the year-ago period — is the important one for the domain name industry, as it is there that registries and registrars have the ability to make a difference.
“The increase is due to a sudden uptick in domain registrations by Chinese phishers,” the APWG said in its Domain Name Use and Trends 1H2013 report (pdf). Chinese targets accounted for 8,240 (68%) of the registered domains.
This works out to about 66 maliciously registered domains per day on average, or less than half a percent of the total number of domains registered across all TLDs daily.
According to the APWG, the number of phishing domains that actually contain a brand or a variation of a brand is smaller still, at 1,244. That’s flat on the second half of 2012.
It works out to about seven new trademark-infringing phishing domain names per day that a brand owner somewhere in the world (though probably China) has to deal with.
APWG reiterated what it has said in previous reports:

most maliciously registered domain names offered nothing to confuse a potential victim. Placing brand names or variations thereof in the domain name itself is not a favored tactic, since brand owners are proactively scanning Internet zone files for their brand names. As we have observed in the past, the domain name itself usually does not matter to phishers, and a domain name of any meaning, or no meaning at all, in any TLD, will usually do. Instead, phishers often place brand names in subdomains or subdirectories.

Only 2% of phishing attacks use cybersquatted domain names

The number of cybersquatted domain names being used for phishing is falling sharply and currently stands at just 2% of attacks, according to the Anti-Phishing Working Group.
The APWG’s first-half 2012 report (pdf) identified 64,204 phishing domains in total.
Of those, the group believes that only 7,712 (12%) were actually registered by the phishers themselves. The rest belonged to innocent third parties and had been compromised.
That’s a steep drop from 12,895 domains in the second half of 2011 and 14,650 in the first half of 2011.
Of the 7,712 phisher-owned domains, about 66% were being use to phish Chinese targets, according to the APWG.
The group’s research found only 1,350 that contained a brand name or a misspelling of a brand name.
That’s down from 2,232 domains in the second-half of 2011, representing just 2% of all phishing domains and 17% of phisher-owned domains.
The report states:

Most maliciously registered domain strings offered nothing to confuse a potential victim. Placing brand names or variations thereof in the domain name itself is not a favored tactic, since brand owners are proactively scanning Internet zone files for such names.
As we have observed in the past, the domain name itself usually does not matter to phishers, and a domain name of any meaning, or no meaning at all, in any TLD, will usually do.
Instead, phishers almost always place brand names in subdomains or subdirectories. This puts the misleading string somewhere in the URL, where potential victims may see it and be fooled. Internet users are rarely knowledgeable enough to be able to pick out the “base” or true domain name being used in a URL.

Taken as a percentage of attacks, brand-jacking is clearly a pretty low-occurrence offence, according to the APWG’s numbers.
In absolute numbers, it works out to about 7.5 domain names per day that are being use to phish and contain a variation of the brand name being targeted.
Unsurprisingly, the APWG found that Freedom Registry’s .tk — which offers free registration — is the TLD being abused most often to register domains for phishing attacks.
More than half of the phisher-owned domains were in .tk, according to the report.

Will new gTLDs really increase phishing?

The US Federal Trade Commission has come out swinging against ICANN’s new generic top-level domains program, saying it will increase online fraud and should be scaled back.
In an open letter to ICANN’s top brass yesterday, the FTC’s four commissioners claimed that “the dramatic introduction of new gTLDs poses significant risks to consumers”.
Saying that more gTLDs will make it easier for scammers to acquire domain names confusingly similar to existing brands, the commissioners said the program should be rolled out as a limited pilot.
The FTC commissioners wrote (pdf):

A rapid, exponential expansion of gTLDs has the potential to magnify both the abuse of the domain name system and the corresponding challenges we encounter in tracking down Internet fraudsters. In particular, the proliferation of existing scams, such as phishing, is likely to become a serious challenge given the infinite opportunities that scam artists will now have at their fingertips. Fraudsters will be able to register misspellings of businesses, including financial institutions, in each of the new gTLDs, create copycat websites, and obtain sensitive consumer data with relative ease before shutting down the site and launching a new one.

The letter demands better Whois accuracy enforcement, better ICANN compliance programs, and a cap on approved new gTLDs in the first round perhaps as low as a couple dozen.
The FTC’s claims that new gTLDs will increase phishing may not be supported by reality, however.
The latest data (pdf) from the Anti-Phishing Working Group shows that in the first half of the year only 18% of domain names used in phishing attacks were registered by the attacker.
That was down from 28% in the second half of 2010. Phishers are much more likely to compromise a domain belonging to somebody else – by hacking a web server, for example.
Of the 14,650 maliciously registered domains 10,444 (70%) were used to phish Chinese targets, “overwhelmingly” the e-commerce site Taobao.com, the APWG found.
Furthermore, only 2% of these domains – just 1,816 over six months – were judged to have been registered due to their confusing similarity with the brands they target.
The APWG said (emphasis in the original):

These are the lowest numbers we have observed in the last past four years, and show that using domain names containing brand strings has fallen further out of favor among phishers.
…
the domain name itself usually does not matter to phishers, and a domain name of any meaning, or no meaning at all, in any TLD, will usually do. Instead, phishers almost always place brand names in subdomains or subdirectories

The APWG found only one gTLD that ICANN has introduced – .info, with 4.5% – in its top ten phishing TLDs. The .com space accounts for 48.9% of all phishing domains.
Will the increase in the number of gTLDs reverse these trends? The FTC seems to think so, but the claims in its letter appear to be based largely on guesswork and fear rather than data.
I suspect that the FTC’s letter is more concerned with ICANN’s ongoing bilateral talks with registrars over law enforcement-demanded amendments to the Registrar Accreditation Agreement.
These talks are completely separate and distinct from the new gTLDs program policies, but in the last few weeks we’ve seen them being repeatedly conflated by US lawmakers, and now the FTC.
This may be ignorance, but it could just as well be an attempt to apply political pressure on ICANN to make sure the RAA talks produce the results law enforcement agencies want to see.
ICANN does not want to be forced into an embarrassing retreat on its hard-fought gTLD expansion. By producing a strong RAA, it could deflect some of the concerns about the program.