Network access-control solutions enjoyed growing enterprise adoption 10 years ago, thanks to the rise of wireless local-area networks and the proliferation of internet worms. A decade later, security demands have escalated well beyond the vision of even the most advanced NAC vendors. The result for many enterprises deploying NAC to protect growing, complex networks has been frustration, escalating costs and declining protection.
NAC is a combination of user authentication, endpoint security assessment and access control. The first NAC solutions hit the market in 2005, at about the time when employees unplugged their company-issued laptops from the network, took them to places outside the office and brought them back to the corporate network infected with malware. NAC was designed to check the health status of the reconnecting laptop, as well as the identity and access rights of the user, before allowing full network access. If there was a problem, access could be denied or minimized until remediation could take place.
That notion seems so quaint in today’s complex computing environment. Now, the endpoints aren’t strictly corporate-issued laptops but also a variety of personally owned notebooks, smartphones and tablets, all manner of internet-connected “things” and even virtualized servers. Along with regular employees, the people connecting to enterprise networks are contractors, service providers, partners, vendors and guests. Given such complexity, NAC has never quite lived up to the reputation it earned early on. It hasn’t been an outright failure, but for many enterprises, NAC has not met the expectations or the needs.
Scale fatigue — too big not to fail
Network access control is a complex undertaking. Because it incorporates the three elements of assessment, authentication and access, there are many working pieces that must be integrated, such as an authentication service (e.g., Active Directory, LDAP, token servers, etc.); a mobile device management or enterprise mobility management solution; endpoint security; and perhaps even a security incident and event management system.
Aside from the interoperability struggles of the different technologies, an implementation requires people with expertise in network management, endpoint management, mobility, authentications services and network security to collaborate. When one technology or group shifts, it creates ripples down the line — and there have been plenty of ripples in the past decade. Consider, for example, the addition of unmanaged smartphones and IoT devices to the mix.
The sheer number of people and devices that need to connect has jumped by an order of magnitude, making policy management a huge challenge. For many enterprise networks, enforcing granular access within the network can require millions of access-control lists. And given that NAC solutions are priced by the number of connecting devices, the initial cost as well as ongoing operations and support can run into the millions.
It’s very telling that the March 2016 Gartner Market Guide for Network Access Control said that customers’ priorities for selecting a NAC solution should include “low levels of complexity of administration” and “ease of integration.”
Limited visibility and misplaced trust
The notion of trust has also changed significantly. Ten years ago it was enough to simply posture-check a device before granting access to a network. Today, with advanced threats and pervasive malware, connected (and posture-checked) devices are one of the most common routes of network compromise. Even if a device is deemed “clean” when it connects, it only takes one phishing message delivered via email to circumvent many security mechanisms and implant malware on the device, from which it can easily spread to the network.
Granted, NAC is not designed to continuously monitor the state of the endpoint and break the network connection if a compromise is found. However, a post-connection compromise shows that the trust established by NAC is a fleeting thing, at best. A more durable trust connection is needed before allowing access to enterprise applications and data.
Cloudy days mean limited protection
According to the RightScale 2017 State of the Cloud Report, enterprises now run 41 percent of their workloads in public clouds. With the rise of software-defined wide-area networks, many enterprises now enable their users to go straight to cloud-based applications without first logging into the WAN. With this type of architecture, users would never go through the NAC solution, rendering it meaningless. Even if user traffic still passes through the enterprise network before hitting the public cloud, NAC offers no access control beyond internal servers. Moreover, existing NAC solutions don’t provide admission control to enforce connection to a virtualized infrastructure.
Repeal NAC and replace it with software-defined perimeter
Jon Oltsik, Enterprise Strategy Group’s senior principal analyst, said the time for NAC has come and gone, and the essential replacement for it is software-defined perimeter technology. SDP is being used in a new generation of access-control solutions that are no longer bound to various types of hardware devices or access-control lists. SDP does have some of the same elements as NAC, as it combines device authentication, identity-based access and dynamically provisioned connectivity. However, the approach to using these elements is quite different from NAC.