First page of Microsoft's 100,000-partner directory, sorted by responsiveness Microsoft Solutions Partner — Security, Modern Work, Infrastructure, App Innovation Microsoft partner since 2006 1,100+ organizations under management
Home/Services/Microsoft Entra Private Access VPN Replacement
ImplementationSecurity and Protection

Microsoft Entra Private Access VPN Replacement

Deployment of Microsoft Entra Private Access — the private-application side of Microsoft's Global Secure Access — to replace legacy VPN access to your internal applications with identity-centric, per-app Zero Trust access. IT Partner installs and load-balances the private network connectors, defines your application segments (specific apps, IP ranges, and FQDNs rather than a flat network), attaches Conditional Access with MFA and device checks per application, rolls out the Global Secure Access client to a pilot group and then the estate, and runs a deliberate coexistence period alongside your current VPN until you are ready to switch it off. $3,950 per project (estimate — confirmed in writing after discovery), typically 2 weeks.

Timeline 2 weeksService owner Roman SotnikMicrosoft Entra Private AccessMicrosoft Entra IDMicrosoft Global Secure Access

What this engagement is

A traditional VPN grants a network position: once connected, a user — or the malware on their laptop — can typically reach far more than the one application they needed. Microsoft Entra Private Access inverts that model. Each internal application becomes an app segment in Entra; users reach exactly the applications policy grants them, each connection is brokered through Microsoft's edge with Conditional Access evaluated per app — MFA, device compliance, risk signals — and nothing hands out a routable position on your network. For organizations already invested in Entra ID, it is the shortest path from 'VPN because we have always had one' to Zero Trust remote access that an auditor will recognize, and it pairs naturally with a broader Zero Trust architecture. The deployment has four moving parts, and we build all of them. Private network connectors — lightweight outbound-only services on Windows Server, deployed in pairs for resilience — link your private network to Microsoft's edge without inbound firewall holes. App segments define what is reachable: we start segment-by-segment with your named applications (by FQDN, IP, and port), and use broader Quick Access ranges only as a transitional device, because per-app segmentation is the entire point. Conditional Access policies attach identity requirements to each segment — who, from what device state, with what authentication strength. And the Global Secure Access client goes out to Windows and macOS (mobile platforms where scoped) through your management tooling, piloted with a real user group before anyone touches the estate. What this service is not: a rip-and-replace stunt. Your existing VPN stays up through a planned coexistence window while access moves segment by segment and the pilot proves out the patterns that matter — including legacy protocols, single sign-on to on-premises resources via Kerberos where needed, and the traffic types that deserve validation before commitment. When every access pattern has moved or been consciously dispositioned, you hold the evidence to decommission the VPN. The decommissioning itself — appliance retirement, contract termination, firewall cleanup — is your action, on your schedule, and we hand you the checklist that makes it safe.

Success criteria

01Private network connectors are deployed redundantly, healthy, and passing traffic for every in-scope application segment.
02Each in-scope internal application is reachable through Entra Private Access by its authorized users — and not by anyone else.
03Conditional Access is enforced per application segment with the agreed MFA and device requirements, verified in sign-in logs.
04The Global Secure Access client is deployed to the agreed user population with a working rollout method for the rest of the estate.
05The pilot group has worked through the agreed validation script — including any legacy-protocol and SSO patterns — with findings resolved or dispositioned.
06Coexistence with the legacy VPN is stable: users on either path can work, and the migration state of every access pattern is documented.
07Your team receives the design document, operations runbook, and the VPN decommission-readiness checklist.

What you receive

Discovery and access-pattern inventory — which internal applications your remote users actually reach, over which protocols, from which platforms; this is what turns the estimate into a fixed quote.
Zero Trust access design document — connector placement, app segment model, Conditional Access matrix per segment, client rollout plan, and coexistence strategy.
Private network connector deployment — installed in pairs on your Windows Server hosts, registered, health-verified, and grouped for the segments they serve.
App segment configuration — enterprise-app segments for named applications (FQDN/IP/port), with any transitional Quick Access ranges explicitly marked for later tightening.
Conditional Access policies per segment — authentication strength, device compliance, and session controls matched to each application's sensitivity, deployed report-only first.
Global Secure Access client rollout — packaged and deployed through your endpoint management (Intune where present) to pilot, then production rings.
Kerberos SSO configuration for on-premises resources where in scope, so file shares and legacy apps do not become the reason the VPN survives.
Pilot execution with a written validation script and a findings log, followed by the production cutover of agreed segments.
Operations runbook — monitoring connector health, reading traffic logs, adding a new app segment, troubleshooting client issues — plus the VPN decommission-readiness checklist.

How the work unfolds

1. Discovery and design

Inventory remote-access patterns, applications, protocols, and client platforms; assess licensing position; produce the design document and confirm the fixed quote. Anything that will not fit Private Access is named here, not discovered later.

2. Foundation

Connector hosts prepared and connectors deployed in redundant pairs; tenant Global Secure Access settings configured; first app segments built; Conditional Access policies staged in report-only mode.

3. Pilot

Client deployed to the pilot group; each access pattern exercised against the validation script — including legacy protocols, printing, file shares, and admin tools; policies switched from report-only to enforced for the pilot; findings fixed or dispositioned.

4. Segment-by-segment rollout

Client rollout expands in rings; application segments cut over in agreed order; the legacy VPN remains available as the fallback throughout, with usage monitored so you can watch it drain.

5. Handover and decommission readiness

Runbook walkthrough with your team, final policy review, coexistence state documented, and the VPN decommission-readiness checklist delivered. The VPN switch-off decision and execution stay with you.

Prerequisites

Microsoft Entra ID P1 or P2 in place for users in scope — it is the licensing prerequisite for Global Secure Access, and it is included in Microsoft 365 Business Premium, E3, and E5.
Entra Private Access licensing for users in scope — sold per user as a standalone add-on or as part of the Microsoft Entra Suite; we confirm your cheapest compliant path during discovery, and Microsoft's pricing at contracting time governs.
One or preferably two Windows Server instances (physical or virtual, on-premises or in Azure) with outbound internet access to host the private network connectors.
Administrative access to the Microsoft 365 tenant and to endpoint management for client deployment — we request granular, time-bound access that you approve.
An inventory of internal applications for remote access, or the willingness to build it with us in discovery — including any that use legacy or unusual protocols.
A pilot group of real users (mixed roles beat IT-only pilots) and a named contact who can approve segment cutovers.
Continued operation of your existing VPN during the coexistence window — replacement is staged, not same-day.

Who does what

IT Partner

  • Design the app-segment and Conditional Access model and deliver the design document.
  • Deploy and verify connectors, app segments, policies, and the client rollout packaging.
  • Run the pilot with a written validation script and resolve or disposition every finding.
  • Execute the segment-by-segment cutover with your approval per segment.
  • Configure Kerberos SSO for in-scope on-premises resources.
  • Deliver the runbook and the VPN decommission-readiness checklist, and walk your team through both.

Your team

  • Provide licensing, connector hosts, tenant access, and the application inventory inputs.
  • Nominate the pilot group and make application owners available for validation.
  • Keep the legacy VPN operational through the agreed coexistence window.
  • Approve each segment cutover and communicate the rollout to users through your channels (we draft, you send).
  • Decide and execute the eventual VPN decommission — appliance, contracts, and firewall cleanup are yours.
  • Operate the platform after handover, using the runbook and our support options if you want us to stay involved.

What's not included

Full SASE or network-architecture redesign — routing overhauls, SD-WAN, firewall replacement, and site-to-site connectivity changes are separate engagements; this project changes how users reach applications, not how your sites reach each other.
Microsoft Entra Internet Access at production scale — the web-filtering and internet-traffic side of Global Secure Access is a natural add-on and we will gladly scope it, but it is its own design exercise, not a line item hidden inside this one.
Physical or contractual decommissioning of your VPN — we deliver decommission readiness; switching it off, returning appliances, and ending vendor contracts are your actions.
Licensing purchases — Global Secure Access licensing is billed by Microsoft; we identify the exact requirement and your cheapest compliant path, and you buy through your existing channel (or through us as a CSP if you choose — your call either way).
Application remediation — apps whose traffic patterns genuinely do not fit Private Access (identified in discovery or pilot) get options documented, not silent workarounds; implementing those options is scoped separately.
Site-to-site or branch-office replacement for the legacy VPN's network-to-network functions — Private Access replaces user-to-application access; network-to-network links need their own design.
Ongoing operations and monitoring after handover — available separately as a managed service; the runbook makes self-operation realistic.

Limitations & technical notes

!Private Access is user-to-application access. If your VPN also carries site-to-site traffic, server-to-server flows, or vendor network links, those functions need separate treatment and we say so in the design document — a VPN doing three jobs is not replaced by a product doing one, and pretending otherwise is how replacements fail.
!Certain traffic patterns deserve pilot-stage validation before commitment: connections initiated from the network toward the client, latency-sensitive real-time protocols, and applications with hard-coded network assumptions. The pilot's validation script exists precisely to find these while the VPN is still the fallback.
!The $3,950 estimate covers a typical SMB deployment — redundant connectors at one primary site, an agreed initial set of application segments, Conditional Access, client rollout, pilot, and coexistence support. Multiple sites, large application estates, or complex legacy-protocol work move the number; the figure is confirmed as a fixed written quote after discovery, before work begins.
!Licensing is Microsoft's meter, not ours: Global Secure Access requires Entra ID P1 or P2, and Entra Private Access is licensed per user — standalone or within the Microsoft Entra Suite — per Microsoft's published pricing as of this writing. We verify current requirements and prices at design time and put the licensing position in writing.
!The Global Secure Access client and platform capabilities evolve quickly under active Microsoft development; supported platforms and features are verified against Microsoft's current documentation at design time rather than assumed from older material.
!Coexistence is a feature, not a failure: plan for the legacy VPN to run alongside Private Access for the agreed window. The decommission date is an outcome of validated migration, not a promise made before discovery.

Frequently asked questions

What is Microsoft Entra Private Access, in one paragraph?

It is the private-application half of Microsoft's Global Secure Access (the other half, Entra Internet Access, handles web traffic). Instead of connecting users to your network the way a VPN does, it brokers each connection to a specific internal application through Microsoft's edge, with Entra ID authentication and Conditional Access evaluated per app. Users get the applications they are authorized for; they never get a network position.

How is this actually better than our VPN?

Three concrete differences. Blast radius: a compromised VPN account exposes whatever the network reaches; a compromised Private Access account exposes only its assigned app segments, each still gated by MFA and device checks. Policy: access rules live in Conditional Access per application, auditable in one place, instead of in firewall rules and VPN group mappings. Operations: no inbound firewall holes, no appliance patching cycle, and adding an application is a policy change rather than a network change. Whether those advantages justify the move for you is what discovery establishes.

What does the $3,950 cover, and when could it change?

It is an estimate for the typical scope: redundant connectors at one site, an agreed initial set of app segments, per-segment Conditional Access, client rollout with a pilot, Kerberos SSO where needed, coexistence support, and the runbook. Discovery confirms it as a fixed quote in writing before work begins — more sites, many segments, or heavy legacy-protocol work are what move it. You pay after you approve delivery.

What licensing do we need?

Two layers, both Microsoft's meter: Entra ID P1 or P2 as the platform prerequisite (already included in Microsoft 365 Business Premium, E3, and E5), plus Entra Private Access per user — available standalone or as part of the Microsoft Entra Suite, which also bundles Internet Access, ID Protection, and ID Governance. Which path is cheapest depends on what you already own and what else you plan to deploy; we put the exact requirement and options in writing during discovery, against Microsoft's current published pricing.

Does the VPN get switched off at the end of the two weeks?

No, and be wary of anyone who promises that. The engagement ends with Private Access serving the migrated segments, the VPN running as fallback, and a decommission-readiness checklist showing exactly what still depends on it, if anything. You switch the VPN off when the checklist is clean and you are comfortable — for many organizations that is weeks after go-live, deliberately.

What happens to file shares, printers, and legacy apps that expect to be 'on the network'?

They are the honest test of any VPN replacement, so they go into the pilot's validation script explicitly. SMB file shares and many legacy applications work through Private Access as app segments, and single sign-on to Kerberos-protected on-premises resources is configurable — we set that up where in scope. Patterns that genuinely do not fit (a rare few do exist, typically apps that initiate connections toward the client) are identified during pilot and given documented options rather than quietly left broken.

Which devices does the Global Secure Access client support?

Windows and macOS carry most deployments, with mobile client support available for scoped platforms — the exact platform matrix is verified against Microsoft's current documentation at design time, because the client is under active development. Rollout goes through your endpoint management (Intune where you have it), pilot ring first. Unmanaged personal devices are a policy conversation: Conditional Access can require compliant devices per segment, which is often the moment organizations tighten BYOD access deliberately.

Do all our users have to move at once?

No — the rollout is ringed twice over. The client goes out in deployment rings (pilot, then expanding groups), and applications cut over segment by segment in an order you approve. Any user not yet migrated keeps using the VPN; both paths work during coexistence. This is also the honest reason the project succeeds: nothing forces a big-bang moment where everything must work simultaneously.

Can Private Access replace our site-to-site VPN links too?

Not in this scope, and mostly not as a product: Private Access replaces user-to-application access. Site-to-site links, branch connectivity, and server-to-server flows are network architecture — if your current VPN concentrator carries both duties, the design document separates them explicitly so the user side can move to Zero Trust while the network side gets its own plan. We can scope that separately.

How does this relate to Conditional Access policies we already have?

It builds on them rather than replacing them. Each app segment becomes an object Conditional Access can target, so your existing patterns — MFA enforcement, compliant-device requirements, named locations — extend to internal applications that policy could never see behind the VPN. New policies deploy in report-only mode first, and if your Conditional Access baseline itself needs work, our dedicated Conditional Access implementation covers that foundation.

Is this Zero Trust, or just a different VPN?

It implements the network-access slice of Zero Trust properly: per-application access, identity and device verification on every connection, no implicit trust from network position. It is not the whole architecture — data protection, endpoint hardening, and monitoring are their own pillars. If you are working toward a full Zero Trust posture, this project slots into our broader Zero Trust architecture implementation as the remote-access component.

What can go wrong, honestly?

The classic failure modes are: an application inventory that missed something users depend on (found by our discovery and pilot, which is why both exist), a legacy traffic pattern that does not fit the model (found by the validation script, dispositioned with options), connector hosts that were undersized or non-redundant (we deploy pairs and check health), and a VPN switched off before the checklist was clean (which is why decommission is your deliberate act, not our closing stunt). The two-week timeline holds when the prerequisites are ready; discovery tells you if your estate needs longer.

Who runs it after handover?

Your team, with the runbook: connector health monitoring, traffic logs, adding new app segments, and client troubleshooting are all documented procedures rather than tribal knowledge. Adding an application later is minutes of policy work, not a project. If you would rather we operate or monitor it ongoing, that is available separately — no obligation, and the runbook is written so you never need us for routine work.

Didn’t find your question?

Ask it here. A real engineer answers by email within one business day — and if it’s a good one, it becomes part of this page so the next person finds it.

Answered by a person, one time, to your inbox. Nothing you type here is published without a human reviewing and anonymizing it first.

Often combined with

$3,950 per project
2 weeks
Plan my VPN replacement