Inequity in Multi-Factor Authentication – choosing the right implementation for your organisation

Share on Facebook0Share on Google+0Tweet about this on TwitterShare on LinkedIn0Email this to someone

The Australian Signals Directorate (ASD) recommends the use of multi-factor authentication (MFA) within their general security control guidance known as the Essential Eight. They claim, “it is one of the most effective cyber security controls an organisation can implement,” yet, not all implementations of MFA are equally effective, so choosing which one is right for your organisation is essential. Furthermore, adversaries know about inherent weaknesses and have begun targeting organisations that use particularly weak implementations. This blog looks at some of the issues and pitfalls with modern MFA solutions and offers some guidance that supports ASD’s claim that it’s one control not to overlook.

In our earlier blog post on multi-factor authentication we explained what it is and the various options available.  You can read it here.

Multi-factor authentication is a stronger mechanism used to authenticate users to a system than the plain old username/password combination. Implementations rely on three categories of authentication services, of which one or more controls from two or more categories are used to authenticate users. These categories are:

  1. Something you know – typically a password or pin;
  2. Something you have – a unique, one-time token;
  3. Something you are – a biometric reading, such as a fingerprint or iris scan.

MFA solutions works by prompting you to supply two authentication types at the point of entry into a system. Typically, organisations opt for a combination of type 1 and type 2 authenticators, where single-use codes are sent to a registered mobile phone, combined with the username and password. In this scenario, the user enters her username and password combination and is then sent the single-use authentication code to their registered mobile phone, which is then compared against the server along with the username and password and only when all three are correct will the user be authenticated.

Understanding Multi-Factor Authentication Implementation Problems

Knowledge of the efficacy of MFA is certainly growing, with social media sites like Facebook and LinkedIn popularising the sending of a mobile token to users if they detect logins from new browsers and locations. Other online services, such as Gmail, PayPal and The PlayStation Network are also introducing two step verification services, many of which use convenient mobile apps, such as the Google Authenticator to prove the user is who they say they are.

Many commercial enterprise vendors provide MFA solutions, with strong uptake over the past few years. This indicates that users (and executives) are becoming increasingly aware of the cyber risk.  Consequently,  vendors are sometimes rushing their MFA implementations to market without proper validation and verification. One example where MFA has turned into a weaker overall security mechanism is that of the mobile phone. Many phones use a fingerprint reader to identify and authenticate legitimate users. However, the same mechanism is now allowing users access to their bank account, so there is one authentication service providing two levels of authentication to a sensitive service. Some banks have also implemented a separate verification service for all online bank transfers, one that sends a single-use pin code to the user to authorise the transaction – and guess where it goes… the very same mobile phone device. This means that anyone finding a way to spoof the fingerprint of a user has access to the bank, and if they are using the phone, they will also receive the bank’s verification code right into the palm of their hand.

There are some other disadvantages of mobile phone based authenticators, since:

  1. If the mobile network carrier is down or you are out of range, you can’t log in;
  2. Mobile phones are hackable, clonable and not infallible, so you are at risk from a compromised device.

Let’s look at another example, in this case a transportation business called We Move Stuff. It has an externally facing service that allows field-based salespeople to connect to and run client-facing applications. We Move Stuff operators can connect email using this remote service, while managers can remotely monitor and manage orders, ensure deadlines are met and deal with delays. We Move Stuff has decided to implement MFA on their remote access service, to help protect themselves from Internet-based attacks, especially now that mandatory breach notification is arriving in 2018. We Move Stuff’s technical team has posited that MFA will significantly strengthen their security posture, so it has gained executive support. However, the technical team has chosen an SMS based authentication service, with little upfront costs, especially because every employee has a company issued mobile phone. On the surface, this sounds like a good approach, but let’s just consider some of the threats and see what improvements could also be introduced.

Threats, Monitoring and Deep Defence

We Move Stuff should conduct a technical threat assessment to understand if it should be concerned. This is a crucial step in evaluating the efficacy of a security control, since it provides a factual risk calculation and assists the business in determining the breadth, depth and type of control required.

As the Essential Eight suggests, cyber security is an organisation-wide consideration and controls should be adopted that mitigate cyber risk at all layers of the business. Through their threat assessment process, We Mode Stuff has considered threats that may seek to exploit vulnerabilities in their systems through:

  • Direct targeting;
  • Indirect, collateral damage;
  • Accidental occurrence;
  • Environmental anomaly.

Threat actors are rated in terms of their capability and motivation, resulting in an understanding of how dangerous each threat actor is to their business. So, after completing the technical threat assessment, We Move Stuff finds that they are a likely target for organised criminal syndicates (highly capable and highly motivated threat actors). They determine moving goods from one state to another is a service that organised criminal entities may seek to compromise and use to their own advantage. Thus, proving an agent is who they say they are is a critical business control. In this case, any MFA solution that is as weak or weaker than standard username/password authentication is not an acceptable outcome. So, instead of tying all of the authentication capabilities to one device (the phone) they opt for an out-of-band system, using an RSA token that isn’t linked in any way to the user’s true identity, other than through the cryptographic pairing of token to user within their infrastructure.

Additional security services can also be employed to assist organisations in preventing authentication-based attacks. Protective monitoring is one such organisational security control that supports the business’s security objectives, since the introduction of a Security Information and Event Management (SIEM) tool or managed security service can detect anomalous authentication behaviour and instantly alert the incident response team. Even if We Move Stuff was targeted and a mobile phone was hacked, any additional authentication steps would be monitored and the correlated activity would result in a security investigation – which is the last thing a threat actorwants.

It is likely that We Move Stuff’s MFA solution mitigates as many as 90% of the threats targeting their business via remote attacks. When combined with a SIEM or protective monitoring service We Move Stuff’s MFA implementation has now raised their security posture to an acceptable level and made the much harder to compromise – in most cases faced with this level of control, a threat actor will look for a new target.

This is the beginning of We Move Stuff’s defence-in-depth strategy, and coupled with their firewalls, IPSs and limitations on administrator privileges, their MFA solution has, as ASD suggests, protected them from most opportunist attacks, as well as at least 85% of targeted attacks.

Multi-Factor Authentication is a critical security control

MFA helps mitigate the risks associated with cyber-attacks. Some implementations of MFA are better than others, so before you implement one, do a technical threat assessment and ensure you understand how any solution might be misused or leveraged by an attacker. Try not to implement an MFA solution that uses one device as its base for services (such as a mobile phone) since a device compromise circumvents all factors of control.

A SIEM or a security monitoring service can be an appropriate layer of additional defence and afford the security operations team with the visibility you need to respond to potential threats.

Multi-factor authentication solutions have significantly come down in price over the past few years, so really there is no excuse as to why you cannot implement one.

 

ASD Essential Eight - The Missing Ninth Step