Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

Enabling security teams requires a cybersecurity sea change 

The traditional approach to some core capabilities of network-based cybersecurity is simply not suited for the way computing has evolved. That leaves well-intentioned security teams disadvantaged trying to protect their organizations.

Man typing

[Images: putilov_denis/AdobeStock]

Profile photo of Martin Roesch

In the chaos of 2020, networks quickly evolved to atomized composites of multi-cloud, hybrid-cloud, and on-premise technologies, utilized by onsite, mobile, and remote workforces. The pandemic left no choice if businesses wanted to keep operating as best as they could. But the resulting implications for network security were massive, leaving security teams scrambling to protect an elastic and ephemeral network environment.

Realistically, this was an inevitable situation. Like the proverbial frog in the slowly boiling pot of water, we are saddled with a security paradigm that has been long in the making but dramatically de-positioned by this recent catalyst. The traditional approach to some core capabilities of network-based cybersecurity is simply not suited for the way computing has evolved. That leaves well-intentioned security teams disadvantaged trying to protect their organizations.

COMING TO TERMS WITH A NEW ABSTRACTION

Truly enabling effective security for the atomized network requires a doctrinal shift in approach: a recognition that the atomized network is a thing unto itself. Just as germ theory evoked a breakthrough medical abstraction with the discovery of disease-transmitting micro-organisms, the cyber community needs a comparable “light bulb” moment to recognize the foundational shift.

Unfortunately, prevailing security approaches rely extensively on a patchwork of solutions, rather than taking a unified approach. During the past two decades, there has been widespread adoption of first physical then cloud-based security approaches from leading vendors; the two do not interoperate well. Cloud-native vendors generally don’t address on-premise environments. Chief information security officers and security operations center teams use their chosen tools; chief technology officers frequently adopt multiple cloud platforms and have their own security teams, who also choose tools they think best fit their needs. Each tool and environment approaches problems differently, usually with its own configuration, threat description language and unique ways of defining problems.

The result is big gaps across the network environment. Attackers live in those gaps.

CONTRADICTIONS IN CURRENT SECURITY INFRASTRUCTURE

In this disjointed environment, there are many inherent technical challenges to overcome. Most current systems (e.g., security information and event management, security orchestration and response, and cloud detection and response) deal with security events, system logs, and other network metadata, aggregating, storing, and querying data on some frequent, periodic basis to look for anomalies. The time gap, even if only a few hours, is too slow when systems are being attacked. Defending the atomized network must be done in real time, detecting and responding to threats as they emerge.

Enterprises also rely on appliance-based deep packet inspection to identify malware on the network. But DPI doesn’t work on encrypted traffic, which is now extensively used. Appliance-based approaches are ineffective in the distributed, diverse atomized network environment, where there is no central network to monitor.

Another strategy, defense-in-depth security, is actually defense in adjacent scope. Each given security technology has its own finite set of capabilities and domain of responsibility. For example, a firewall filters traffic according to a set of defined constraints. If the constraints are met, the traffic is allowed through. If they’re not, the traffic is blocked. While a helpful first line of defense, the firewall is not looking for attacks.

The next link in the chain is commonly intrusion prevention systems, examining the network traffic to detect and block attacks happening within the network protocols. An IPS operates on the principle that if something can be seen, it can be prevented. Most SaaS applications are delivered over the web, but commonly used encryption techniques make it impossible to see network attacks inside network traffic. What’s more, IPSs don’t block incoming malware files embedded in protocols—it’s out of their scope, computationally expensive to inspect, and hard to do.

Even next-generation firewalls, combining firewalling with intrusion prevention, still don’t examine files containing malware. That then shifts the burden of defending against file-based attacks to endpoint detection and response-based anti-malware systems. These systems must determine if a file that is attempting to execute is satisfactory or suspect and stop it if necessary.

While most security infrastructure has been built on these approaches, their varying scopes of responsibility leave gaps attackers can exploit.

THE ZERO TRUST BLIND SPOT

Many across the cyber industry and the user community are now turning to zero trust as a sweeping approach to protect diverse environments. Encrypting disk, memory, and network traffic while leveraging identity-brokered access control, the principle of zero trust constrains an attacker’s ability to move laterally inside the network after an initial breach. It’s a useful and logical method—but with one notable shortcoming.

Encrypting all network traffic means incumbent detection and prevention technologies relying on inspecting traffic in the clear for suspicious characteristics are blinded. That renders attack detection on IPSs and next-generation firewalls ineffective; they don’t flag what they can’t see. Most organizations, however, don’t implement zero trust in a way that’s comprehensive enough to fully protect their networks. That leaves a situation in which widely adopted security capabilities are blinded, for the promise of a strategy that is not broadly deployed, with high reliance on endpoint detection and response systems to defeat attacks no longer caught on the network. When the inevitable network compromise happens, analyzing network metadata can provide a way to identify anomalies and indicators of compromise—if known devices and their correct organizational context are pre-characterized to allow it.

SHIFTING OUR COLLECTIVE MINDSET

Heavily encrypted, atomized networks mandate a new approach in which security architecture is unified, frictionless, and comprehensive, wherever and however technology is consumed.

That approach, rather than the piecemeal architectures so prevalent today, is the future that will enable the security teams of tomorrow. Getting there requires our collective industry to understand and accept a new abstraction that reflects what is already reality—that the atomized network is everywhere.


Martin Roesch, CEO, Netography.