Traditional security programs were predicated on protecting the typically internally hosted technology infrastructure and the data within that environment. This led to an ecosystem composed of numerous discrete tools and processes all intended to detect adversaries and prevent harm. It included a multitude of controls spanning network and infrastructure security, application security, access control, and process controls.
In a world now dominated by the cloud, where applications leverage public cloud providers or are subscribed to as SaaS apps, the challenges of security have dramatically changed. Secure Access Service Edge (SASE) security architectures, with Next-Generation Secure Web Gateways (NG-SWG) at their cores, will transform security functions to fully address the cloud. They’ll also change how CISOs think about the role of security within businesses.
I’ve been a CISO for some of the most complex organizations in the world, and I see SASE as the logical evolutionary next step for our security programs. That’s a big part of why I joined Netskope, whose progress in the SASE framework is described by Gartner as being farther along than any other vendor. But we’ll need to approach it the right way to succeed. By shedding light on where the industry needs to get to, my hope is that the interim steps will become clearer.
Out with Old Perceptions
Security teams must jettison some old thinking and entrenched beliefs. Let’s start with the following:
Discard the data center mentality: As the percentage of company data stored in the cloud and SaaS apps has risen, the data center has become just one destination among many that people need to access. It’s time to realize that your world is defined by where your data lives. The data center is no longer a “center.”
Forget about perimeters: Data center-focused security was about protecting perimeters and regulating traffic. Cloud security is about understanding users and connections, applying appropriate levels of control, and monitoring and regulating traffic in real-time. You cannot retrofit data center security, with its rigid boundaries and technology stack, to accomplishing this task in the cloud.
Stop thinking of the “corporate network” as a well-defined place: The pandemic accelerated the distribution of the workforce. Digital transformation continues to swell the expanse of third-party suppliers and services with which a business interacts. The security landscape is vast and shifting. Work has to happen wherever an employee or a third party happens to be. Establishing the data center as a single control point to which you routed all network traffic worked fine when all data was within your environment. Now that data is everywhere, a security network and associated security services must follow it. This requires a new network architecture that reflects that dispersal.
Embrace SASE
SASE currently offers security teams the most complete architectural design for addressing this new world. It’s also been the subject of so much hype that things have become confused. So, from one CISO to another, here’s how it works.
SASE creates a security cloud: In SASE, network access points are distributed throughout the globe. Security services are also distributed and operate at those access points. This security cloud becomes the new control point for monitoring traffic and protecting everyone. Every user, whether they are an employee or a third party, is effectively a branch office of one.
SASE employs an integrated collection of real-time services: Instead of a loose conglomeration of separate hardware appliances and services with each element serving an independent role in network security, SASE is a single set of integrated services that work together to do the work of CASB, DLP, SWG, ATP, and other important functions. In fact, those narrow acronyms will fade away as the focus shifts to “big picture” descriptions of the work required: protecting data, users, and applications, and doing all of those using a unified platform, single-pass inspection, a single console, and a single policy engine.
SASE understands what users are doing in real-time: Data center-focused security grants access, but traffic flows are left effectively unmonitored as users work. The work of SASE does not stop at access. SASE makes it possible to understand, for example, if a user is sharing sensitive information inside different instances of the same SaaS applications such as Microsoft OneDrive or Google Docs. Furthermore, properly-implemented SASE can inspect and apply policy based on user actions—what the user is doing with the data. This can only be achieved by a system that understands the user, the application, and the data, creating context that can then be applied to policy implementation.
Here is a practical example. Let’s say an employee logs into the corporate G-Suite using their personal laptop and then accesses Docs to do their work. We want to allow that. But we want to detect if, while using the same device, that user also logs into their personal Gmail account. We want to detect that behavior and permit the multiple logins, but protect corporate data by making it read-only and prevent it being uploaded to the personal instance.
Another example is if a business partner uses Google Drive to share a doc. We want to allow our users to download from that third-party instance of G-Suite because it is a business requirement. The old hard and fast approach that blocks everything is too coarse and inhibits the business and end-user needs. SASE lets us see the details that describe instances, actions, and behaviors to then enforce risk-based policies.
SASE enables the granular control we need for business to function in the perfect “Goldilocks zone”: not too much risk by allowing everything, but not constraining business opportunity by blocking things. SASE recognizes context, which enables us to understand what the user needs to do, why they need to do it, and then enforce policies that are optimized to meet the business’ interests.
Running a SASE World
SASE isn’t a product you buy today; it’s an architecture that you will build over time. One of the first steps is to replace Secure Web Gateways and other appliances that are limited to access control and watching web traffic patterns for threats. A Next-Generation Secure Web Gateway (NG-SWG) refactors a much wider array of network security functions into a unified system that protects business interests in the Cloud and has data protection at its core. This provides a more capable foundation for building digital security capabilities and makes the job of managing cyber risk much easier. Consider what is then achievable:
Context: In the SASE world you know much more about users, user actions, devices, network connections, applications, and data. This not only gives you rich context about what is happening but also makes it possible to evaluate risk relative to that context. This makes it possible to have granular, flexible control over sophisticated policies that are implemented automatically and vary appropriately in response to changes in context.
Policies: In the data center-centric security paradigm, security staff spent lots of time configuring systems at a low level to implement the desired security controls. In the SASE world, the security staff instead sets detailed policies. The NG-SWG system implements those policies by orchestrating the collection of services available to it. These integrated services work together to create the result specified by the policy–the control objective. Security teams can then spend more of their time thinking about the posture required to protect users, data, applications, instead of focusing on the low-level configuration.
Automation: The availability of rich context and integrated services opens up new possibilities for automating the detection of threats and of unusual patterns of activity, as well as the variety of responses that can be taken when something happens. The services are all on the same platform and use the same context, so automation is much easier to achieve than when data and context are trapped and processed by discrete systems acting independently, as has been the case when we focused entirely on data centers.