When Your SASE Provider Shares a Data Center With Everyone Else

Hyperscaler SASE PoPs share the cloud's failure domains and routing. See how dedicated PoPs change resilience, performance and sovereignty, and what to ask your vendor.

Matthew Brooks
Technical Marketing Manager
  • Read Time: 5 min read
  • Published: November 19, 2026
  • Modified: November 19, 2026
  • 5 min read read
  • November 19, 2026
  • November 19, 2026

Summary

Hyperscaler SASE PoPs inherit the cloud provider's failure domains and routing decisions. This post explains the shared-fate, SLA and sovereignty risks of hyperscaler-hosted points of presence, how dedicated SASE PoPs with their own ASN change traffic engineering, and the questions to put to any SASE vendor before an outage answers them for you.

  • A SASE service built on a public hyperscaler shares that cloud's failure domain, so a zone or region outage can become a connectivity and security outage for the customer.
  • Uptime SLAs typically cover the vendor's service boundary rather than end-to-end application performance across cloud and ISP networks, which is what users actually experience.
  • Vendors running PoPs inside a hyperscaler generally cannot set their own peering, choose upstreams, or steer around congestion, because routing follows the cloud provider's design and commercial priorities.
  • Dedicated PoPs operated under the vendor's own ASN allow direct peering, multiple Tier 1 transit providers, and performance-based route selection which tends to mean lower and more predictable latency for voice, video and interactive SaaS.
  • Sovereignty requires knowing who operates the infrastructure and which law governs them, not only which region the data sits in.

Public cloud outages are inevitable. A utility power failure, a backup system that doesn’t start, a faulty configuration push or a network fault can take a cloud zone or region offline for hours. When that happens, the impact reaches well beyond the cloud provider’s own services. It extends to every third-party vendor that builds on that infrastructure, whether for compute, storage or networking.

That “blast radius” includes SASE vendors that host their points of presence (PoPs) on public hyperscalers. For organizations that route their users’ traffic and security inspection through those PoPs, a cloud outage can quickly become a connectivity and security outage. This raises important architectural questions every SASE buyer should investigate ideally during an RFP rather than during an incident.

The Risks of Hyperscaler-Hosted SASE PoPs

Running security processing on public clouds can have real advantages for SASE vendors, like elastic capacity and fast geographic expansion. It also creates delivery risks that may be a disadvantage for customers:

Shared Fate

A zone or region failure at the cloud provider can affect the SASE service built on it, however well the SASE vendor designs around it. Multi-zone redundancy helps, but much of that redundancy is designed and operated by the cloud provider rather than by the vendor you contracted with.

SLAs That Stop at the Vendor’s Boundary

An uptime SLA covers the vendor’s service. It does not necessarily cover application performance across cloud-provider and ISP networks, which is what users actually experience.

Sovereignty Means Control (Not Just Location)

Choosing a region addresses data residency. Organizations bound by regulations like NIS2 or GDPR, or by national security requirements, may also need to know who operates the infrastructure, which jurisdiction governs that operator, and whether authentication, logging and inspection all stay in-country. These are requirements of a Sovereign SASE service.

Did you know? Redundancy inside a hyperscaler region is largely designed and operated by the hyperscaler, not by the SASE vendor running on top of it. A vendor’s failover plan can therefore depend on controls it does not own, which is worth confirming in writing before you need it.

Optimal Routing: Whoever Owns the ASN Owns the Path

How traffic travels to and from the PoP, toward the SaaS application, cloud workload or branch it is headed for, depends heavily on how the SASE vendor connects to the rest of the internet.

SASE vendors that host PoPs on public hyperscalers typically rely on the hyperscaler’s network and transit to carry traffic after inspection. Hyperscaler backbones are large and well connected, but routing on them is engineered around the cloud provider’s own network design and commercial priorities, not around any one SASE customer’s applications. The SASE vendor generally cannot set up its own peering relationships from those PoPs, choose which upstream carries a given flow, or steer around a congested exchange point. As a result, traffic may take a longer or less consistent path to a destination than a direct route would, which adds latency and jitter that users experience as slow applications and choppy calls.

Vendors that run their own dedicated PoPs, with their own autonomous system number (ASN), can make those decisions themselves. They can peer directly with major SaaS and cloud providers, buy transit from multiple Tier 1 carriers, and select routes based on measured performance rather than accepting a single provider’s default path. Fewer network hops and direct handoffs to the destination network tend to mean lower and more predictable latency, especially for real-time traffic such as voice, video and interactive SaaS.

See the architecture: explore how Versa Cloud Gateways and the Versa SASE Fabric are deployed.

Questions to Ask Any SASE Vendor

  • Which PoPs run on public hyperscalers, and which on colocated or dedicated infrastructure?
  • What happens to my traffic when a single cloud zone fails, what is the documented failover behavior, and what was measured in recent incidents?
  • Does the SLA address end-to-end application performance, or only service availability?
  • For sovereign requirements, can authentication, inspection and logging all be confined to one jurisdiction, and who is the legal operator?
  • Does your status page publish root causes, and how quickly?

The Versa Networks Advantage

Versa provides customers with secure, reliable global transport using the Versa SASE Fabric: an interconnected, policy-aware, distributed system where networking, security, identity, analytics and orchestration operate as a single unified platform. It relies on Versa SASE Cloud Gateways, which run in dedicated PoPs hosted in resilient data centers such as Equinix.

Because many organizations also need multi-cloud connectivity to reach workloads hosted across the major hyperscalers, Versa additionally runs SASE Cloud Gateways inside AWS, Azure and GCP, creating on-ramps from every cloud environment. In its 2026 Magic Quadrant for SASE Platforms, Gartner calls out Versa’s PoP infrastructure strategy as a strength.

A global, dispersed network of PoPs provides greater flexibility and fault tolerance. More PoPs also tend to improve customer performance, because the connection can be chosen along the best-performing path through a Versa Cloud Gateway to the Versa SASE services.

Hosting services in dedicated PoPs is also what makes Sovereign SASE possible for jurisdictions with narrow compliance requirements. Versa launched Europe’s first fully sovereign SASE on the noris network AG sovereign cloud platform (nSC), hosted in its German data centers, giving EU enterprises subject to GDPR, NIS2, DORA or Germany’s KRITIS framework a path to full digital sovereignty.

The Takeaway

The point is not that cloud-hosted PoPs are unusable. It is that infrastructure ownership, failure domains and sovereignty are architectural properties and they are worth testing in an RFP before an outage tests them for you. To see how Versa Universal SASE fits your networking and security requirements, visit versa-networks.com or reach out to your Versa account team.

Matthew Brooks

By Matthew Brooks

Technical Marketing Manager

Matthew Brooks brings more than 15 years in cybersecurity, cloud, and enterprise networking to his technical marketing work across Versa's SASE, SSE, SD-WAN, SD-LAN, and firewall products. He previously led product marketing for Cisco's Zero Trust and identity security portfolio, and held product and marketing roles at Verizon and Citrix.

FAQs

A SASE point of presence is the location where a SASE provider terminates user and branch traffic and performs security inspection before forwarding it to its destination. Where that PoP runs, and who operates the infrastructure beneath it, determines both the failure domain and the routing options available to that traffic.

A SASE service running inside a public cloud shares that cloud's failure domain, so a zone or region outage can take the security and connectivity service offline with it. Much of the redundancy relied on is designed and operated by the cloud provider rather than the SASE vendor.

A vendor operating dedicated PoPs under its own ASN can peer directly with SaaS and cloud providers, buy transit from multiple Tier 1 carriers, and select routes on measured performance. A vendor inside a hyperscaler generally inherits that provider's routing design and commercial priorities instead.

Not necessarily. Region selection addresses data residency, but sovereignty frameworks may also ask who operates the infrastructure, which jurisdiction governs that operator, and whether authentication, inspection and logging all remain in-country.

Which PoPs are hyperscaler-hosted versus dedicated, what the documented failover behavior is when a cloud zone fails, whether the SLA covers end-to-end application performance or only service availability, and who the legal operator is for any sovereign deployment.

Subscribe to the Versa Blog

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.

Related Posts

Company Updates Feb 26, 2025

What is Sovereign SASE? Ensuring Data Sovereignty & Compliance

A foundational primer on Sovereign SASE and why data sovereignty depends on who operates the infrastructure, not only where the data is stored.

Read More
Industry Insights Apr 06, 2026

Why Data Sovereignty Fails Without Sovereign SASE

Why data-residency-only approaches fall short, the same argument this post applies to PoP location, and what sovereignty requires for data in motion.

Read More

Sovereign SASE: The Four Planes True Sovereignty Requires

Why genuine sovereignty spans four planes, data, control, management and jurisdiction, plus the questions that pressure-test any vendor's claim.

Read More