In many organisations, networking and security have developed separately. Branch offices run on SD-WAN or MPLS, remote staff connect through a VPN, web traffic passes through a proxy and SaaS applications are managed from another console. Each tool may work well on its own. Together, they can create overlapping policies, multiple endpoint agents and traffic routed back through a data centre simply to be inspected.
Secure access service edge (SASE) addresses this complexity by bringing networking and security together in an architecture delivered largely from the cloud. The aim is to apply consistent policies whether someone connects from a branch office, home or on the move.
This guide explains how SASE works, where it helps, what its limitations are and what to consider when comparing providers or planning a migration.
What is SASE and how does it work?
SASE combines wide area networking with a range of security services, delivered mainly through a provider’s cloud points of presence (PoPs) rather than appliances installed at every site. Users and branch offices connect to a PoP, where traffic is routed and security policies are applied before it reaches the internet, a SaaS application or a private data centre.
The main components typically include:
- SD-WAN (software-defined wide area networking): Routes branch traffic across broadband, leased lines, 4G/5G or MPLS according to application requirements and network conditions.
- ZTNA (zero trust network access): Gives users access to specific applications rather than the wider network, based on identity, device posture and policy. It can replace remote access VPN for compatible applications or reduce reliance on it.
- SWG (secure web gateway): Filters and inspects web traffic to enforce acceptable-use rules and protect against threats.
- CASB (cloud access security broker): Provides visibility and control over SaaS use, including applications that haven’t been approved by IT.
- FWaaS (firewall as a service): Delivers firewall controls from the cloud without requiring a physical firewall at every site.
Many platforms also include data loss prevention (DLP) across web, SaaS and private application traffic.
In its Magic Quadrant for SASE Platforms, published in July 2025, Gartner defines single-vendor SASE as SD-WAN, SWG, CASB, network firewalling and ZTNA delivered by one vendor through a cloud-centric architecture.
What is the difference between SASE and SSE?
Security service edge (SSE) covers the security services within the wider SASE architecture. These typically include SWG, CASB and ZTNA, with some platforms also offering FWaaS and DLP. SSE does not include SD-WAN.
That distinction matters when comparing options. An organisation with a relatively new SD-WAN deployment may be able to add SSE while keeping its existing networking platform, rather than replacing both at once.
What makes a SASE platform different?
An organisation could buy SD-WAN, a web proxy, a CASB and a ZTNA product from four suppliers. What distinguishes a SASE platform from that collection is how well the components work together.
A well-integrated platform can offer:
- Shared policies: Rules defined once can apply across branch, home and mobile users, rather than being recreated in several consoles.
- A consolidated client: One endpoint agent may replace separate VPN, proxy and DLP software, although some platforms still require multiple agents.
- Centralised logging: Keeping telemetry in one place can simplify investigation and reporting.
- Single-pass inspection: Traffic can be processed at a PoP without passing through several separate appliances or services in sequence.
These are potential benefits, not guaranteed features of every product marketed as SASE. Integration varies between providers, making it an important part of any evaluation.
Traffic inspection also depends on configuration. Decryption policies need to account for the platform’s capabilities, the type of traffic, privacy requirements and applications that may not tolerate interception.
Benefits and limitations of SASE
Where integration is strong, consistent policies can reduce gaps created by managing security rules in several places. Connecting users to a nearby PoP can also avoid routing traffic back through a central data centre, potentially improving SaaS performance. The actual benefit depends on PoP locations relative to users and applications, as well as how traffic is routed beyond them.
Fewer consoles and endpoint agents can simplify day-to-day operations. New sites may also be brought online without shipping and installing a stack of security appliances.
There are trade-offs, however.
Greater dependency on one provider. An outage or misconfiguration can affect multiple sites and users at once. Resilience, failover and recovery arrangements need to be understood before deployment.
Uneven feature maturity. A provider with strong SD-WAN capabilities may be less developed in CASB, for example, or vice versa. Assess each component rather than assuming the entire platform is equally capable.
Vendor lock-in. As more networking, security policies and operational workflows move onto one platform, the cost and complexity of switching can increase. This can also weaken your negotiating position at renewal.
Changes to team responsibilities. Network and security teams may need to share ownership of policies and agree who approves changes, investigates incidents and manages exceptions.
SASE also has limits as a security architecture. It can support a zero trust strategy, but the two are not interchangeable. Adopting SASE does not, by itself, establish zero trust.
SASE and zero trust: what is the difference?
The NCSC’s zero trust architecture design principles include knowing your users, devices, services and data; assessing device and service health; authorising requests through policy; and avoiding implicit trust in any network, including your own.
ZTNA and inline traffic inspection can support several of these principles for connections between users and applications.
Zero trust extends further, though. It relies on strong identity foundations, such as a carefully planned migration from passwords to passkeys, alongside device management, data protection and network segmentation.
SASE controls how users connect to applications. It does less to protect traffic moving between workloads inside an organisation’s own environment. Securing east-west traffic requires its own visibility and segmentation controls.
Single-vendor, multi-vendor or SSE alongside SD-WAN?
SASE can be assembled in different ways, each balancing integration, flexibility and operational responsibility differently.
Gartner’s July 2025 Magic Quadrant forecast that by 2028, 70% of SD-WAN purchases would form part of a single-vendor SASE offering, up from 25% in 2025.
That points towards greater convergence, but it doesn’t mean one architecture suits every organisation. Capabilities still vary between providers, as the vendor-by-vendor assessments in Gartner’s July 2026 Magic Quadrant for SASE Platforms show. Market forecasts provide useful context, but existing contracts, internal skills and appetite for risk should carry more weight in the decision.
The main approaches are:
- Single-vendor SASE: Networking and security come from one provider, potentially simplifying policy management and integration.
- Multi-vendor SASE: Components from different providers are connected to retain specialist capabilities or avoid replacing existing products. This can offer flexibility, but integration and policy consistency need careful management.
- SSE alongside existing SD-WAN: Security services are introduced while the current networking platform remains in place. This can reduce disruption when SD-WAN is relatively new or still under contract.
The best option depends on the capabilities you need, how well your existing tools integrate and whether consolidation delivers a clear operational benefit.
Managed, jointly managed or self-service SASE?
Choosing the architecture is only part of the decision. You also need to decide who will operate the platform, maintain policies and manage changes as requirements evolve.
Managed SASE is available through managed service providers (MSPs), telecoms providers and, in some cases, the SASE vendors themselves. Three operating models are common:
- Managed: The provider operates the platform and makes policy changes on your behalf, usually through service requests.
- Jointly managed: Responsibilities are shared. Your team typically defines and approves policies while the provider handles deployment, monitoring and day-to-day operations.
- Self-service: Your team operates the platform directly, while the vendor maintains the underlying cloud service.
Some responsibilities remain with your organisation regardless of the model. You are still accountable for access requirements, data classification, risk appetite and regulatory obligations, even when platform administration is delegated.
Read service levels carefully. A platform availability SLA is different from an MSP’s commitments for change requests and incident response. If a provider resells another company’s platform, establish which SLA covers each type of failure.
Agree exit terms before signing. Ask whether policies and configurations can be exported in a usable format, what notice periods apply, what transition support is included and how logs and other data will be returned or deleted.
Planning a migration from SD-WAN or VPN
For most organisations, adopting SASE means a phased migration rather than replacing networking and security tools all at once.
Should you replace SD-WAN or add SSE?
If your SD-WAN is relatively new, has several years left on its contract and integrates properly with your preferred SSE service, adding SSE alongside it may be the sensible option.
Replacing SD-WAN becomes more attractive when a renewal or hardware refresh is approaching, integration is limited to basic tunnels, or a shared policy across networking and security is a priority.
Moving from VPN to ZTNA
Replacing remote access VPN with ZTNA can be a practical early step because it can be phased without changing branch networking. Granting access to specific applications can reduce unnecessary network exposure, provided those applications are compatible and access policies are configured correctly.
Two common approaches are:
- By user group: Start with a defined population, such as contractors or a pilot team, then expand.
- By application: Begin with web applications before assessing those that use other protocols.
Some applications may need exceptions. Legacy systems, server-initiated connections and certain voice or administrative tools may not work well through ZTNA.
Run VPN and ZTNA in parallel during the transition. Define a rollback route for each phase, and retire VPN access only when every dependent application has been accounted for.
Migrating branch sites
Branch migration usually works best site by site, starting with lower-risk locations. Existing firewalls can remain in place alongside cloud-based inspection during the transition.
Measure latency and application performance before and after each phase. That gives you evidence from your own environment rather than relying on headline performance claims.
How to evaluate SASE providers: a buyer checklist
Migration requirements should shape your provider assessment. Beyond security features, establish whether a platform fits your existing systems, meets performance requirements and can be operated effectively.
- Application performance: Test real response times, latency and routing from representative branch sites and remote users. Compare results with your current network, including during peak periods.
- Resilience: Establish what happens when a tunnel or PoP becomes unavailable, how quickly traffic fails over and where it is routed.
- Integration depth: Check whether the platform genuinely offers one console, one policy model and one endpoint agent, or several products under a single brand.
- Security capabilities: Assess SWG, CASB, ZTNA, FWaaS and DLP individually rather than treating them as a single bundle.
- SD-WAN capability: If networking is included, compare routing, resilience and application performance with your existing platform.
- Identity integration: Confirm support for your identity provider and device-posture signals.
- UK and EU data residency: Establish where traffic is inspected and logs are stored, then review the implications with your data protection team.
- Migration support: Check coexistence with existing VPN, SD-WAN and firewalls, including rollback options.
- Operating model and SLAs: Clarify whether the service is managed, jointly managed or self-service, and document responsibilities for each party.
- Total cost of ownership: Include direct costs, implementation, savings from retired tools and any caps on renewal increases.
- Roadmap and availability: Establish which features are generally available today. Treat promised capabilities as roadmap items until you can test them.
Don’t assume a nearby PoP automatically means better performance. Traffic may still be routed inefficiently after reaching it. During a pilot, record which PoP each site and remote user connects to, then test the applications people use every day against a baseline from your current network.
It’s also worth asking whether the provider owns or leases the backbone between PoPs, and how that affects routing, resilience and visibility.
Frequently asked questions
Is SASE suitable for every organisation?
Not necessarily. SASE is worth evaluating when networking and security tools have become difficult to manage together, remote access needs have grown or a network refresh creates an opportunity to consolidate. Organisations with stable, well-integrated tools may gain little from replacing them all at once.
Should we choose SASE or SSE?
SSE may be sufficient if your main priority is securing remote users and SaaS access, particularly when your existing SD-WAN is relatively new and well supported. Full SASE becomes more relevant when branch networking is also due for change and a shared policy across networking and security is a priority.
Can SASE replace VPN completely?
For many users and applications, ZTNA can replace remote access VPN. Complete replacement depends on application compatibility. Some organisations retain a reduced VPN service for legacy systems, administrative access or server-initiated connections.
Should we replace existing firewalls and networking tools?
Not automatically. Keeping SD-WAN, firewalls or specialist tools that are relatively new, well integrated and still under contract may reduce risk and avoid unnecessary expenditure. Running them alongside SASE can be a long-term design choice, not just a temporary migration stage.
Is managed SASE better than self-service?
That depends on your team’s capacity, out-of-hours coverage and how quickly policy changes need to be made. A jointly managed model may suit organisations that want to retain control of policy decisions without taking on all the day-to-day operational work.
Choosing the right SASE approach
SASE is less a single product than a series of decisions: how closely to integrate networking and security, who should operate the result and how quickly to move away from existing tools.
Start by identifying the gaps in your current setup, then compare architectures and providers against your own requirements. A phased migration, backed by performance testing and clear ownership, can reduce the risk of replacing several established services at once.
Conversations with relevant providers can help you test those requirements against real implementation approaches. The Elevate.Tech Summit, taking place on 10 November 2026 at Hilton London Canary Wharf, brings senior technology decision-makers together with providers through pre-arranged 30-minute one-to-one meetings based on participants’ requirements.
Qualifying buyers can attend on a funded basis, subject to eligibility and a commitment to the full-day itinerary. Find out how to attend as a buyer.
Image credit: https://unsplash.com/photos/white-clouds-photography-xa-X3YICs3s


