Back 
Co-authored by Mark Day and Jeff Brainard
In recent blogs, we’ve explored the role of Security Service Edge (SSE) technologies as part of a SASE architecture, and the key differences between SSE and SASE. But so far, we’ve focused more on overall functionality than on its realization and what SSE means from a cloud design and implementation perspective. In this post, we shift gears to put a spotlight on networking and infrastructure as it relates to security clouds. Let’s start by stepping back and looking at legacy enterprise networking.
What’s Changed with the Cloud?
Until fairly recently, most enterprises hosted their data and applications in enterprise-managed data centers. Accordingly, the role of enterprise networking and infrastructure leaders was primarily to support secure connectivity to the data center. The security services inspecting traffic and enforcing policies had certain features in common: They were protecting traffic flowing to and from the data center, and they were packaged as appliances running in the data center.
Recent years have seen the data and applications of interest migrate out of the data center and into the cloud. Two examples are Microsoft Office 365 for office productivity tasks and Salesforce for sales automation and customer relationship management. Accordingly, it’s no longer realistic to treat the data center as the only concern for connectivity, as well as security. Instead, the primary focus becomes the user and the user’s varied requirements. The emphasis is now on an integrated platform of connectivity with embedded security services. From the user’s perspective, this architecture simplifies safe and secure access to web, cloud, SaaS, and private applications.
From the perspective of enterprise IT, SSE is appealing because it replaces a jumble of different connectivity and security technologies. Perhaps more importantly, the traditional approach is no longer workable: A collection of physical or virtual appliances daisy-chained in a data center simply won’t work after the apps, users, and data have moved away. SSE can become how every user reaches every service. It combines the best aspects of an enterprise WAN and the internet, with high-performance secure connectivity to a broad collection of destinations.
SSE is also a path for traffic to flow between users and application services; a set of inspection and policy-enforcement mechanisms to avoid security problems like malware or data leakage; and a means of isolating an organization’s public services from direct internet connections, making them more difficult to find and attack, as well as mitigating the risk from lateral movement.
What are the Core Security Cloud Requirements for SSE?
Underpinning SSE is some flavor of a security cloud. In simplest terms, the key requirements of a security cloud are a distributed implementation with relatively intense computations performed at a relatively large number of locations, and management and visibility tools sufficient to build an elastic, but high-performance service.
Broadly speaking, a security cloud must:
- Support connections from enterprise users all over the world, on diverse devices. Part of that diversity includes both IT-managed devices and IT-unmanaged devices (the latter sometimes called “bring your own device,” or “BYOD”).
- Control access to a variety of cloud services, including diverse SaaS services as well as enterprise app workloads that may be deployed in a multi-cloud arrangement.
- Perform inspection and policy enforcement on enterprise traffic traveling between users and cloud services.
- Maintain proper localization and proximity information for each user/service combination.
- Be fast. More specifically, a security cloud needs to be fast enough so that user experience doesn’t motivate users to bypass or reject the security controls provided by the security cloud.
The key problems for good performance are distance (speed of light constraints) and crossing multiple networks (peering and market constraints). Both of these problems are well-known to network and cloud architects. A centralized implementation of a global security cloud cannot achieve high performance. Instead, the security services must be globally distributed to deliver on SSE advantages: both for performance, and also for availability and resilience.
It would be lovely if we could simply build a global security cloud in a public compute cloud and call it SSE, but unfortunately, the universe doesn’t work that way — at least, not if we want to get the best results. Instead, a modern, global security cloud is a different kind of cloud. It effectively needs to combine the intense computing capabilities of a public compute cloud with the global reach of a Content Delivery Network (CDN).
Realizing the promise of SSE – and ultimately the security and networking convergence of SASE – requires solving these performance problems. Accordingly, it’s crucial for customers making investment decisions to understand a vendor’s architectural choices in this area. Netskope NewEdge is a unique security private cloud leveraging these insights–fast, cloud-native, and data-centric–built from the ground up for SSE and SASE, backed by industry-leading Service Level Agreements.
For a deeper dive into the networking and infrastructure considerations in this rapidly evolving era of SSE and SASE, download your complimentary copy of “The Network Is the Security: Perspective On the Design & Implementation of Security Clouds” by Netskope Chief Scientist and Distinguished Network Advisor Mark Day. Or visit the Netskope NewEdge micro-site at: www.netskope.com/newedge.
Read the blog