Turning the Tide on Cyber Attacks with a New Perimeter to Protect Enterprises

Today's business ecosystems span across multiple organizations, security perimeters and trust zones, increasing the number of attack vectors. With the migration of enterprise applications to the cloud, attack surfaces have expanded significantly. Compounding the problem, today's wave of attacks are more sophisticated and diversified than ever. The combination of enterprise complexity and advanced attacks has broken the traditional enterprise perimeter, creating the perfect storm for threats and information loss, where a new security paradigm is needed.

The security landscape is constantly evolving in today's extended enterprise, but security solutions have not kept pace to protect the ever-changing enterprise environment. Today, a remote, mobile, and contract workforce, combined with a complex partner ecosystem, make it difficult for IT to control access to the high value, high risk enterprise applications. Providing full network access via VPN to untrusted external organizations or users opens up many security risks, given the increased sophistication of modern day attacks.

The push to the cloud only adds to this risk, as enterprises migrate applications to the cloud with the promise of business agility and cost reduction. But this leads to increased cybersecurity risk due to the lack of secure connectivity to the applications residing in the public cloud. Since these applications are no longer hosted behind the traditional enterprise LAN perimeter, they become difficult if not impossible to secure against attacks such as server exploitation, denial of service, and credential theft.

Enterprises need an innovative approach to defend the new enterprise perimeter. One emerging method is the Software Defined Perimeter (SDP), an initiative of the Cloud Security Alliance (CSA), backed by more than 200 organizations globally and now being included in Gartner's Hype Cycle for Virtualization. The SDP introduces a more versatile and secure way of extending the secure enterprise perimeter that includes lower trust zones such as cloud, contractors, and partners. It uses a need-to-know architecture derived from work at the Defense Information Systems Agency (DISA), where the client device's security posture and identity must both be verified before it is granted connectivity. This zero-trust approach starts with no visibility and connectivity until the device and user have both been authenticated using multifactor authentication (MFA). This inherently secure architecture enables anyone to securely access protected applications over untrusted networks using any device, even unmanaged devices, which have become so popular.

The SDP architecture separates the control plane and the data plane, providing protection by isolating the point of authentication and the point of application access. In addition, the SDP allows for real-time access permissions within the system for granting and revoking access.

Source: Vidder, Inc.

SDP Client
SDP uses client software (SDP Client) on the endpoint in order to create a secure, multi-protocol tunnel. The client runs over the application layer, providing a versatile solution capable of spanning across any arbitrary network and trust zone. The endpoint security posture is further attested by the client software before accessing the SDP-protected application(s), including probing for local information to determine if the device is company-owned or a bring-your-own-device (BYOD). This information is then used to enforce policies around which applications the device is granted access.

SDP Controller
The first server-side component in the SDP architecture is the SDP Controller, which is used to authenticate devices, enforce authorization on users, and manage policies related to access and attestation. Upon executing the SDP Client, it will first attempt to contact a SDP Controller in order to perform device-level authentication. This step is referred to as transparent Multi-factor Authentication (tMFA). Once the device is authenticated and identified, the user will be prompted to login to the system with username and password. This user authentication is implemented using Security Assertion MarkUp Language (SAML), an industry standard implemented by Identity Access Management (IAM) teams. After successful user authentication, the SDP Controller will obtain a list of authorized groups that are ultimately mapped to protected applications. Finally, a set of SDP Gateways is propagated to the SDP Client in real-time in order for it to initiate access to the SDP-protected application. The crucial step in mapping undertaken by the SDP Controller is the sharing of this authenticated client's credentials with the SDP Gateway (described below).

SDP Gateway
Another primary component in the SDP architecture is the SDP Gateway. This component provides secure access to a User of a SDP-protected application. Importantly, SDP Gateway uses a dynamically changing firewall that starts with a default "DENY ALL" state. The firewall is only opened for SDP Clients after they have successfully authenticated to a SDP Controller. Once opened, the SDP first connects using a secure, mutually trusted TLS connection, which will then provide access to the protected application.

Secure IT/Business Outsourcing & Ecosystem Collaboration
Unlike VPNs, the SDP securely connects partner organizations to authorized applications without putting them on the internal network. This is done using a secure tunnel terminated on the SDP Gateway, effectively removing network visibility. The SDP Gateway is invisible to attackers on the Internet because its default condition is a "DENY ALL" state in the firewall. The firewall is opened dynamically as users are identified and proven to be trustworthy, allowing the SDP to mitigate the leading attack vectors that are exploited on internet-facing systems. The SDP provides very granular access to only the protected applications that are authorized by the enterprise. This can include a variety of protocols and application types - including browser based and fat client applications such as SAP. By keeping the partner's untrusted device from the network, the attack surface is reduced and the risk to the organization is lowered.

Secure Cloud Migration
The pace of cloud migration is increasing as enterprises continue to move their applications to the public cloud with the goal of increasing agility and reducing cost. However, CIOs know that moving important data and applications to the cloud has a risk that must be considered seriously, especially when it comes to providing users secure access to these cloud applications. The SDP isolates the applications from attacks by requiring multi-factor authentication (MFA) before connectivity is granted.

Defeating Top Attacks Cited by Verizon Breach Report
The following are examples of common threats and attacks today, and how the SDP's innovative approach can help.

Credential Theft is one of the most common attack vectors prevalent in today's enterprise environments. The theft can occur from a variety of channels - including key logging, phishing or brute force attacks. The SDP's transparent MFA functionality mitigates stolen credentials by requiring that all connections to a protected application come from a SDP Gateway. In order to gain access to this Gateway a user must pass the transparent MFA process that acts as a second factor of authentication.

Denial of Service is an attack that is used to limit the availability of a network, application, or server. Often performed by activists or competitors, it can ruin the reputation of an organization and cause a loss of services for customers. The SDP can defeat a large proportion of these attacks by limiting connectivity to only authorized users. For example, a common denial of service attack is called the "SSL Renegotation DoS" which continually re-establishes SSL connections to a server until it runs out of resources to service future requests. Malicious adversaries generally perform this attack against internet-accessible websites. This attack vector is eliminated when the application is protected by the SDP, which will not allow a connection to the application until it is proven trusted.

Server Exploitation can come in many different varieties, including SQL Injection, Buffer Overflows, and Cross-site Scripting. Attackers perform these attacks remotely and without legitimate credentials. For example, one of the most common vulnerabilities is SQL Injection, which is often performed against login and search forms. These forms can generally be accessed and attacked without having a username and password. Attacks such as these are a very serious problem since many applications have been moved to be internet-accessible, ultimately allowing attackers to hack into systems and gain access to an internal network and corporate data. The SDP protects these applications by requiring an attacker to have valid security artifacts and a valid device before gaining any connectivity to the application. This limits the attack surface of the application to authorized users only, which greatly reduces the likelihood of an attack.

Connection Hijacking, also known as man-in-the-middle, is an attack vector that is used to steal or modify data in flight. It's not uncommon to hear about connections to a wireless hotspot in public coffee shops that resulted in an eventual compromise of an organization. This is caused by insufficient in-flight encryption systems. Most applications use the system's "Trusted Certificate Store" or "Keychain" to enumerate a trusted list of Certificate Authorities. This store contains approximately 300 unique Certificate Authorities that could be potentially compromised and result in a connection hijacking attack. The SDP uses a methodology referred to as "Certificate Pinning" that trusts only a small subset of Certificate Authorities. This greatly reduces the likelihood that the Certificate Authority can be compromised.

What You Need to Know
The SDP has advantages over today's current approaches because it implements mutual trust in the authentication process. This requires both the client and server to present security artifacts during the connection and authentication process. This mutual trust provides an extremely high level of confidentiality and integrity on remote connections being initiated from untrusted locations.

While the waves of attacks continue to pound enterprises in this perfect storm of vulnerability due to changing enterprise environments and increasingly advanced threats, approaches such as the SDP that start with zero connectivity and provide access to authorized resources only after multiple authentications are met can help organizations turn the tide on today's continued threats.