Dark Reading is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them.Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Threat Intelligence

8/7/2019
10:00 AM
Connect Directly
LinkedIn
RSS
E-Mail vvv
50%
50%

Rethinking Website Spoofing Mitigation

Deception technology is evolving rapidly, making it easier for organizations to turn the tables on their attackers. Here's how.

Website spoofing is a common problem that has doubled in the last year, resulting in $1.3 billion in losses, according to the 2019 Thales Access Management Index (registration required). In a high-profile example of website spoofing that has left the business world more than a little rattled, hackers successfully diverted about 500,000 customers visiting the British Airways website last summer to a realistic-looking but fraudulent site, without the airline having any idea it had been spoofed.

The spoof site gathered names, addresses, login information, payment card details, and other data. After a review by the EU's Information Commissioner's Office, British Airways faces a possible record-breaking fine for violating the terms of General Data Protection Regulation (GDPR), to the tune of £183.5 million — about 1.5% of the airline's annual revenue. Because of website spoofing, not only does the airline suffer losses to its customers and its brand but the additional government-levied fines and financial pain.

This attack vector has been around for decades and continues to be popular because it's difficult to detect until it's too late. For a fairly small investment, adversaries can acquire all the tools they need on the Dark Web to set up highly convincing website-spoofing schemes, as part of a larger phishing campaign. Website spoofing works on all of the major Internet browsers and is not prevented by "secure" connections. The adversary can observe and modify all website pages and form submissions, even when the browser's "secure connection" indicator is lit. The user sees no easily discernible indication that anything is wrong. Even sites protected by two-factor authentication cannot escape the assault.

Under Europe's GDPR, multinational corporations must make every effort to provide "reasonable security" mechanisms to protect personal data. As the British Airways fine shows, "reasonable security" now extends to spoofed sites that the organization never even knew about. This incident should serve as a wake-up call that current methodologies aren't effective enough. It's time to rethink the way we approach detection and mitigation of website spoofing incidents.

Shift the Focus
To date, options for detecting when a website has been spoofed have largely relied on monitoring domain registrations and manual web searches. However, this approach is susceptible to human error and is only capable of identifying spoofed websites after the fraud occurs. A common approach for detecting website spoofing has been to search for brand images and relevant content from the copied site. But this approach leaves the attacked company with no insight into how many and which of its clients were potentially victimized and does nothing to deter the next attack.

Another familiar approach has been to train workers and consumers to be able to spot fakes. This is a good idea in theory because it educates and empowers individuals. It's true that in some cases, spoofed sites can be detected because something "looks off" compared with the original website. However, the sheer volume of spoofs being created and the increased level of sophistication behind these attacks makes it quite difficult for the user to spot them with certainty. It also places an unfair burden on the victim of the attack, making them responsible for detection instead of the owner of the original website that's been targeted.

Domain registration monitoring and training employees fail to help organizations detect spoofing early in the life cycle of an attack, before data is stolen. Further, these approaches cannot help businesses understand how long the fake site has been active, and most importantly, how many customers or employees may have been victimized. The two biggest challenges of all — who initiated the attack and deterring the adversary from trying it again — are completely left out of the mitigation process with these approaches.

Deceiving the Deceivers
A more sensible strategy is to focus on early detection and a smarter mitigation strategy called deception technology that is painful to the attacker and provides actionable data for organizations. (Disclaimer: Allure is one of a number of vendors that market deception technology products.) The idea is to flood the adversary with highly believable decoy credentials and personal information. This causes a great deal of doubt about what may have been stolen, making it hard for the fraudster to discern what is real and what is fake. The only way to know is to test all credentials at the real site, causing greater overhead costs for the adversary, and providing an opportunity to gather information about the attacker when decoy logins occur, such as endpoint IP addresses under his or her control.

Deception techniques could finally shift the advantage in favor of the legitimate businesses victimized by website spoofing. At Columbia, our Computer Science and IDS Lab has been conducting experiments to determine how certain deception techniques can bait adversaries with highly convincing but false credentials embedded with tracking mechanisms that are triggered when the attacker attempts to open or exfiltrate them. We called this the "BotSwindler," and it aims to detect crimeware such as spoofing by deceptively inducing attackers into an observable action during the exploitation of monitored information injected into the guest OS. To entice attackers with information of value, the system supports a variety of different types of bait credentials, including decoy Gmail and PayPal authentication credentials, as well as those from large financial institutions.  

Whether the motivation is to spread fake news in pursuit of influence, steal customer login credentials or credit card numbers for financial gain, or break into cloud shares and networks to exfiltrate intellectual property, website spoofing has devastating effects on company reputation, consumer trust, and corporate revenues. It's time to take a more modern approach to solving this pervasive security problem. Simply detecting IP anomalies isn't enough. Deception technology is evolving rapidly and is well-positioned to detect website spoofing schemes sooner, giving organizations the ability to turn the tables on adversaries.

Related Content:

 

Black Hat USA returns to Las Vegas with hands-on technical Trainings, cutting-edge Briefings, Arsenal open-source tool demonstrations, top-tier security solutions, and service providers in the Business Hall. Click for information on the conference and to register.

Dr. Salvatore Stolfo is the founder and CTO of Allure Security. As a professor of artificial intelligence at Columbia University since 1979, Dr. Stolfo has spent a career figuring out how people think and how to make computers and systems think like people. Dr. Stolfo has ... View Full Bio
Comment  | 
Print  | 
More Insights
Comments
Newest First  |  Oldest First  |  Threaded View
tdsan
50%
50%
tdsan,
User Rank: Ninja
8/17/2019 | 7:54:52 AM
Re: A new way of looking at a problem,

Also, a token could be issued by the company to validate their authenticity, this would be stored on the customer's browser or computer to ensure they are working with the right organization. This ensures the data is sent to the right company, even if it was not, the domain providers could provide a pop-up to the user stating that this could be a nefarious or ill-advised action, do you want to proceed or not.

One of the other options I mentioned below (from the quote in our earlier discussion) would be to add a token to the site, this would allow the browser to determine if the site was valid by using a SHA256 hash, site descriptor, and purpose all built into this number. This could be used to ensure the site is not a compromised site or if it is a site they have visited before, the browser would determine that much in the same way we use certificates. 

For me, I go to a few sites on a regular for personal and business purposes. There are others but this would help to address the security issue, not all but at least some aspects). Also, we need to start migrating the DNS environment to specifically use DNSSEC and move away from IPv4. This would be much harder for hackers to penetrate defenses because we would secure DNS traffic, reduce MITM attacks, create truly secure connections using IPSec VPN AES256 connections (all built into the protocol - IPv6). We can start identifying where the attack derived from (1-to-1 connections using IPv6) and the token would help to validate the site with the help of ML (I wanted to reiterate the point listed below because ML was only one of the points brought up in the beginning phases of the discussion).



Tokenization is the future of business and personal transactions. Blockchain is looking into that as well.

T
SSTOLFO000
50%
50%
SSTOLFO000,
User Rank: Guru
8/17/2019 | 6:39:54 AM
Re: A new way of looking at a problem,
T

Very wise observation. (I've been a proponent of using ML in security since the inception of my lab at Columbia University, sponsored by DARPA, since 1996.) There are indeed companies that access and analyze DNS and domain registry data to infer suspicious sites using ML and NLP techniques, to head off problems before they start. ML indeed is broadly applicable to many of the problems in cybersecurity. This will help immeasurably, but not completely. The latest work in research is now focusing on Adversarial ML, techniques employed by adversaries to comfound and confuse ML systems. This is accomplished either through "training attacks", where training data consumed by an ML system is poisoned, or "testing attacks", where test data is suitably altered to avoid accurate classification of the ML system. This adversarial attack on ML first appeared by clever spammers who altered their content to avoid detection by the spam filters in email traffic. The same can be applied to domain and DNS data, leaving the problem still unsolved. Attackers are clever, and we have to be dligently one step ahead. My view is a combination of techniques raises the bar high enough to make it very hard for adversaries to succeed in scale. ML based approaches definitely have a role. 
tdsan
50%
50%
tdsan,
User Rank: Ninja
8/14/2019 | 10:21:31 AM
Re: A new way of looking at a problem,
You made some interesting and valid assertions, I think there was a section that you may have overlooked. It stated we should be able to add "ML" or "Machine Learning" to the solution; basically look into creating an intelligent system where the system is constantly looking for illegal or corrupt sites (allow the machine to sift through false positives); this will help carriers to block sites if they are considered nefarious, this will help to improve the entire DNS ecosystem (this process helps to identify key relationships and ascertain if the site is used for valid business or for criminal purposes. It can also be used to help improve performance if use correctly to identify breaks in communication.

Site Spoofing

The ways to use "ML" are endless, we can use it to capture information from the 9 DNS servers across the globe to look for potential actors, we can use the data from Akamai to help and address some of the issues found to even block illegal traffic going to the site (there would need to be a better relationship with the ISPs but I think this would be a start).



Something to think about.

T

 
SSTOLFO000
0%
100%
SSTOLFO000,
User Rank: Guru
8/12/2019 | 6:33:01 AM
Re: A new way of looking at a problem,
Todd

Sensible point. There are a few companies who do access and analyze DNS entries in an attempt to identify spoofing sites. But adversaries are quite clever in making it hard to accurately identify a spoofed domain without a potentially large number of false positives, and false negatives. In one recent example, an enterprise using such a service failed to identify a spoofed site, they were alerted by a customer. 

DNSSEC was proposed over a decade ago and still hasn't become pervasively deployed. It is a community project effort and it is increasingly hard to expect community spirit in the internet's global setting. In the meantime one must expect continuing growth in adversary creativity to create undetected spoofed sites and victimize users to steal their credentials. 

As it now stands, fraudsters have relatively little effort and cost to gather and sell stolen credentials. Flooding them with deceptive credentials changes all of that making their effort expensive; forcing adversaries to pay a price is a good strategy to change the balance of power they now enjoy for free. 
tdsan
50%
50%
tdsan,
User Rank: Ninja
8/8/2019 | 2:15:11 PM
A new way of looking at a problem,
The DNS companies have a huge database. What if the domain providers were able to look at the dns informaiton, implement machine learning and determine if there are companies who have similar domain names or identify if the domain came up days after their creation or just recently. They could identify if this was a problem by sending out an email. In addition, if the domain was brought up by another country, then the originating web owner would be notified of this occurrence, they in turn could notify their clients, partners and the link of any issues associated with their domain.

Also, a token could be issued by the company to validate their authenticity, this would be stored on the customer's browser or computer to ensure they are working with the right organization. This ensures the data is sent to the right company, even if it was not, the domain providers could provide a pop-up to the user stating that this could be a nefarious or ill-advised action, do you want to proceed or not.



This process gives the user an example of how to spoof a site, this needs to be looked at in more detail at a global level (9 dns organization providers), how to improve the existing process, because the existing methods are not working. I think the use of IPv6 with DNSSec could help address some of these issues, we need to fully adopt IPv6 and get off IPv4, most of the hacks occur using that protocol.

Todd

 
AI Is Everywhere, but Don't Ignore the Basics
Howie Xu, Vice President of AI and Machine Learning at Zscaler,  9/10/2019
Fed Kaspersky Ban Made Permanent by New Rules
Dark Reading Staff 9/11/2019
Register for Dark Reading Newsletters
White Papers
Video
Cartoon Contest
Current Issue
7 Threats & Disruptive Forces Changing the Face of Cybersecurity
This Dark Reading Tech Digest gives an in-depth look at the biggest emerging threats and disruptive forces that are changing the face of cybersecurity today.
Flash Poll
The State of IT Operations and Cybersecurity Operations
The State of IT Operations and Cybersecurity Operations
Your enterprise's cyber risk may depend upon the relationship between the IT team and the security team. Heres some insight on what's working and what isn't in the data center.
Twitter Feed
Dark Reading - Bug Report
Bug Report
Enterprise Vulnerabilities
From DHS/US-CERT's National Vulnerability Database
CVE-2019-14540
PUBLISHED: 2019-09-15
A Polymorphic Typing issue was discovered in FasterXML jackson-databind before 2.9.10. It is related to com.zaxxer.hikari.HikariConfig.
CVE-2019-16332
PUBLISHED: 2019-09-15
In the api-bearer-auth plugin before 20190907 for WordPress, the server parameter is not correctly filtered in the swagger-config.yaml.php file, and it is possible to inject JavaScript code, aka XSS.
CVE-2019-16333
PUBLISHED: 2019-09-15
GetSimple CMS v3.3.15 has Persistent Cross-Site Scripting (XSS) in admin/theme-edit.php.
CVE-2019-16334
PUBLISHED: 2019-09-15
In Bludit v3.9.2, there is a persistent XSS vulnerability in the Categories -> Add New Category -> Name field. NOTE: this may overlap CVE-2017-16636.
CVE-2019-16335
PUBLISHED: 2019-09-15
A Polymorphic Typing issue was discovered in FasterXML jackson-databind before 2.9.10. It is related to com.zaxxer.hikari.HikariDataSource. This is a different vulnerability than CVE-2019-14540.