I don't know who coined the term "data breach", but the vision of data leaking from a company like water from a burst dam seems particularly appropriate this year.
Some notable recent data breaches have included:
- Ashley Madison / Avid Life Media. If you run a service to assist cheating spouses, confidentiality has to be the name of the game.
It's not clear whether any customer data was stolen, but
enough data about the internal operations was leaked to suggest that customer data was insecure and to cause the company serious problems. The motivation appears to have been blackmail of Avid Life Media, with threats to release customer data unless the company shuts down various websites. It's been suggested (without any evidence) that this may have been an inside job, but personally I would be looking for a betrayed spouse.
- Office of Personnel Management. In two
"separate but related incidents" (if that's indeed possible), personnel data on 4.2 million current and former federal government employees was stolen, as well as background investigation records of about 20 million current, former, and prospective government employees and contractors. It has been suggested that this may have been the work of Chinese military intelligence.
- LastPass. The maker of widely used password manager, had the email addresses and lost password
questions of its customers stolen. Although user passwords were not apparently stolen, the combination of email address and possible password reset question answers might be useful in penetrating other systems.
- mSpy. A mobile spyware maker, its database of email addresses and passwords was stolen, among "hundreds of gigabytes of data". Discussion on darknet forums suggest that the stolen login credentials may be used to hijack accounts where naive users have re-used
- Trump Hotels, Hershey Park, Sally Beauthy, Harbtouch plus many others. Credit and debit card data stolen from point of sale transactions. The data was / is being used to make fraudulent online purchases, or to create fake credit cards for offline fraud.
- Hacking Team. The spyware company's source code, customer list, emails, and other information were made available on the internet. The information revealed contracts with some
questionable regimes, deals to buy software vulnerabilities, usernames, passwords, etc. It allowed software vendors to plug the security holes being exploited by its products and anti-virus companies to build detectors for its spyware.
- Gamma International. The spyware company's source code, and some 40GBytes of internal documents were published on the internet. This enabled detection of its FinFisher spyware, as well as identification of some of its less savory
- ICANN. This non-profit company which is ultimately responsible for the domain name system (that's the mapping of names such as www.example.com onto IP addresses) and is effectively the domain registrar's registrar. It lost a set of email addresses and hashed passwords for domain registrars.
- Saudi Aramco. Although the attack is quite old, some details are only now being published. The politically motivated attackers succeeded in wiping
the disk drives of 35,000 computers.
When examining the risks of deliberate (as opposed to accidental) incidents, we need to look at the motivations of potential perpetrators. These can give us both an idea of the type of attacks to expect and the amount of effort a potential attacker may go to in order to extract data. It also gives us an idea of what the attacker might do with the data and the harm it may cause. If we examine the incidents above, the motivations seem to fall into a
number of categories:
Attackers with this objective are mostly interested in directly monetizing the data they can steal. A primary objective here is frequently the theft of large numbers of credit card details. These will typically be resold for cash on the black market: others will use them to make fraudulent purchases.
Also of value to this type of attacker is the collection of the personal information needed to impersonate someone. Once again there is a black market
for the data: the attacker will typically resell the data to others who will use it to fraudulently apply for loans and credit.
On a different scale, there may also be targeted attacks against individuals in a company with access to banking and payment systems: the company's bank accounts can be drained with fraudulent transfers, employees added to a payroll, or suppliers paid for non-existent deliveries. Any fraud which could be committed by an insider may also be possible by an outsider
with stolen credentials.
Means to an End
Even if the only IT in our company is a website devoted to the exchange of knitting patterns, we may be at significant risk of attack, although the victims are likely to be our customers or clients. Password re-use is common by naive system users. This means that Jane's knitting site password may also be her banking password, her PayPal password, and her Google play store password. While our company may not lose directly if this data is
stolen, it can quickly lose reputation and customers. There are simple encryption techniques which can be used to minimize the value of this sort of data if it is stolen: even if all you own is a list of the login credentials of knitting enthusiasts, it is important that they be applied.
Stolen credentials can also be a target. Some of the recent attacks on corporate point of sales systems were enabled by first attacking contractors and vendors who had legitimate access to the target
network. Once inside a company's network, attackers can launch attacks as privileged insiders rather than outsiders.
Another class of means to an end attack, is the "watering hole" attack. If you want to steal credentials from bank managers, why not attack a company with a large number of bank managers as customers? Attack or control places where your targets hang out, and attack them from a direction that they may assume is safe.
Whistleblowing and Hacktivisim
company is doing things that some people find morally repugnant, then it may be the target of whistleblowing (from inside) or hacktivism (from outside). The objective here is to expose the corporate behavior to the wider world. Targets here may include email messages, customer records, meeting minutes, etc. Whereas a single document may be dismissed as a forgery or a lapse of judgement, a mass release of data showing repeated patterns of questionable behavior and attempts at concealment is
generally enough to convince the public at large of the genuineness of the data.
A secondary objective may be to punish the company: by publishing trade secrets and revealing customer identities it may be possible to destroy or severely curtail the company's business operations.
While the motives of whistleblowers and hacktivists are altruistic, a blackmailer may wish to steal the same classes of data for purely financial reasons. Instead of simply leaking the data
to the public, the blackmailer will try and extort money, goods, or services from a company by not publishing the data he or she holds.
Spies, whether corporate or government, are in the business of obtaining information and access to information.
Information on current negotiations, proposals, customer lists, and customer agreements may be of interest to a competitor. Design, performance,and production information may of interest to a foreign
Also of interest is any information that enables the spies to identify people with access to useful information, as well as (obviously) any information that can be used to persuade an individual to assist them in their work.
If an employee, prospective employee, or customer becomes sufficiently slighted by a company or its staff, then their thoughts might turn to revenge. The objectives here are to cause disruption or destruction, without any concern as to
whether the actions make money for the perpetrator or are in the public interest.
Once again, any embarrassing or important collection of information may be the target for exposure, as can collections of private customer data which would damage the business if revealed. The wholesale disruption or corruption of data is also a possibility.
If your company experiences a data breach, what can it expect?
Once class of effects is purely financial. Disclosing
credit card information may result in fines from financial institutions. And, although not strictly in the same class as a data breach, any internal banking credentials stolen may lead to theft of money from corporate bank accounts.
A second effect may be disclosure of trade secrets. Some companies (but far fewer than most people think) depend upon the confidentiality of trade secrets. Disclosure may render these secrets worthless, or at a minimum enable competitors to replicate costly
research. The two spyware companies listed above are cases in point.
However, the main effect of a data breach is likely to be a significant loss of trust in the company.
Will people still wish to become our customers still if there is a risk that the relationship might be disclosed? If that is the case, disclosure of a customer list may be enough to put you almost completely out of business. Will customers still wish to deal with us after they have been subjected to credit card
fraud or had their bank account hijacked due to a data breach at our company? Will clients and vendors still trust us if they can see the email discussions we had about them? Will employees still wish to work for us if their personal details and emails are published for all to see? Or what if they see what managers and other employees are saying about them?
Reducing the Risks
The methods of reducing the risks are all well-known (if not always practiced) in IT circles. They
- Principle of least access / minimum privilege.
Only allow people or systems the minimum access required to do their job. Development staff do not need access to production systems, the staff in sales do not need access to the personnel database, the public web server does not need access to the personnel database, etc. It's sometimes inconvenient and tricky to maintain, but this is one of the first lines of defense.
- Use firewalls and
physically distinct networks.
Assuming that an internal system is "owned" by an attacker, limit its usefulness by restricting what other systems could be attacked at a physical / hardware level.
- Split responsibilities
It's a basic fraud prevention principle in accounting, and it extends here too. Don't put everything under one person's control. If a person can give themselves permission to access everything, there is little to stop them (or somebody
impersonating them) doing everything.
- Use logging and intrusion detection systems
When a member of staff plugs in that "found USB drive" to see what is on it a nd gets their computer "owned", it's important to be able to detect the attempts to find and "own" other computers as soon as possible, before too much damage has been done. It's also important to be able to figure out what happened after the event, otherwise there's little to stop the same thing happening
- Don't store un-salted or simply encrypted passwords
Assume that an attacker has been able to steal your database of simply encrypted user passwords due to website vulnerability. (A surprisingly common scenario). An attacker can do two things: they can determine if two users have the same password by simple inspection; they can try guessing user passwords for all users at the same time. And if you are using a simple encryption algorithm rather than a properly
designed password hashing algorithm, they can search for the right password very fast. Indeed, they may even have a list of the encrypted value of the million most common passwords they can use to look up the passwords. Password salting throws a random per-user value into the mix, which makes it impossible to simply compare encrypted passwords or to search for all user passwords simultaneously. A password hashing algorithm is designed to run slowly enough that searching through a long list of
possible passwords becomes infeasible.
- Enforce strong passwords or pass phrases and discourage password re-use.
If you read the announcements of data breaches by hackers, you will know that a constant source of amusement is poorly chosen passwords by system administrators. These are passwords that would have been impossible to guess if properly chosen, but were trivial. Often they were also re-used across multiple systems, making the attacker's job even easier.
"Clever" passwords are often easy to guess. (Here is how to come up with passwords that even the NSA couldn't guess.). And if you have more than a few passwords to remember, use a good password
manager such as STRIP instead of re-using passwords. (No relationship, just a fan.)
… never do anything as a company that you would be ashamed of if it was revealed. That's not just a case of protection from hackers, but also one of protection from legal discovery and eternal