The Cloud Security Alliance (CSA), a group that promotes security best practices for cloud computing, is making progress on its plans to develop a software-defined perimeter (SDP) framework for protecting Internet-connected systems against a range of security threats.
The non-profit group has teamed up with Waverley Labs to work on an open-source SDP framework that it says will give organizations a very different approach to protecting on-premise and cloud-based applications than traditional perimeter tools offer.
SDPs are based on protocols that were originally developed for the Department of Defense and the National Security Agency, says the CSA. They are designed to enable authorized devices and users to connect securely with enterprise applications and systems over the public Internet, while denying access to everything else.
SDPs incorporate the idea of dynamically provisioned perimeters for mitigating network security threats. Typical security approaches involve the use of fixed perimeter technologies to control access to network connected assets. In an SDP, all network-connected systems are completely hidden from public view and are made visible only to devices that have been properly validated.
The goal is to make the application infrastructure completely undetectable with no DNS information or IP addresses being publicly available. Only devices that properly authenticate themselves to an SDP controller are provided visibility to the network, and only to systems and applications to which the device owner is authorized.
An SDP allows enterprises to overlay what essentially amounts to a "black," closed private cloud on top of the pubic Internet, says Jim Reavis, CEO of the Cloud Security Alliance. It allows organizations to mitigate common security threats like distributed denial-of-service attacks more effectively than traditional perimeter tools, he says.
Large-scale SDP deployments, like at Coca-Cola and at some federal agencies, signal the growing interest in the SDP model in securing network connected assets, Reavis says. But most have involved the use of proprietary technologies.
The initiative with Waverley Labs is to develop an open-source framework that will allow organizations to take advantage of readily available commercial tools to deploy software SDPs, he says.
According to Reavis, an SDP consists of five distinct layers of security controls. It starts with a single packet authorization system where the first packet sent to the SDP controller by a device seeking access to the network, and has to cryptographically verify the device’s identity.
An SDP also uses Mutual Transport Layer Security to enable two-way authentication between the client and server, and a separate device validation step to ensure the device is running trusted software and is being used by the rightful owner of the device.
Dynamic firewalls are another crucial component of SDPs. They are designed to automatically deny all incoming requests except those made by previously approved systems. Instead of generalized rules, dynamic firewalls apply access control rules based on the access rights assigned to the device.
Once all the device validation and user authentication tasks are completed and access has been properly provisioned, an application-binding layer creates an encrypted TLS tunnel to protect access to the application and to ensure the applications only communicate through the tunnel.
The open source SDP framework initiative will give enterprises an alternative to conventional perimeter tools, says Juanita Koilpillai, CEO of Waverley Labs. Under the effort, Waverley and CSA will work on delivering the design for the various security layers in a phased manner, starting with the single packet authorization layer, she says.