10 min read

Site-to-Site VPN Explained: How It Connects Offices (2026)

LiMP VPN Editorial TeamLiMP VPN Editorial TeamBlog & news editorial team
Site-to-Site VPN Explained: How It Connects Offices (2026)

When an employee works from home or a cafe, protecting their traffic only takes a regular client VPN like LiMP VPN — install the app, turn it on, turn it off. But when you need to connect a head office and a branch so that an internal server, a shared database, or a network printer is reachable from both locations without installing a VPN client on every computer, that is a different job, and a site-to-site VPN is what solves it.

Before configuring the tunnel, check whether the two office subnets overlap: if both use the same range, for example 192.168.1.0/24, routing between the networks will not work. The LiMP VPN subnet calculator quickly works out the mask, address range, and host count for a subnet — a free online tool, nothing to install.

Below: how a site-to-site VPN differs from a regular device VPN, which protocols fit the job, how to plan office addressing without conflicts, a step-by-step WireGuard setup with a sample configuration, and an honest answer to whether LiMP VPN can be used for this task.

What a Site-to-Site VPN Is and How It Differs from Remote Access VPN

A site-to-site VPN is a tunnel between two network gateways (routers or servers) that runs permanently in the background and connects entire networks: devices in office A reach devices in office B by their internal IP addresses, as if they were on the same local network. A remote access VPN (regular client VPN) is a tunnel between ONE device and a VPN server, which a person turns on and off manually whenever they need to protect their connection or reach a specific resource.

AspectSite-to-Site VPNRemote Access VPN (Client)
What it connectsNetwork to network (office to office, office to data center)One device to a server
Who sets it upA network administrator, on both gatewaysThe user, by installing an app
Operating modeAlways on, in the background, no employee action neededTurned on and off manually as needed
Typical scenarioConnecting offices, branch to headquarters, office to cloudProtecting traffic on public Wi-Fi, in a cafe or hotel
What each office needsA router or server with VPN support on each sideJust an app on a phone or laptop
ProsTransparent access to the whole network, no client needed on every machineQuick to install and turn on, no admin required
ConsHarder to set up, needs a static or public address on at least one sideDoes not connect whole networks, protects one device at a time

When You Actually Need a Site-to-Site VPN

A site-to-site VPN makes sense when the goal is connecting networks, not protecting one person's traffic.

  • A head office and a branch constantly exchange files or work in a shared database, and access needs to work without a VPN client on every workstation.
  • You need access to an internal server (file storage, a video surveillance system, a phone system) from another location, as if that server were physically in the same building.
  • The company rents a server or virtual machine in a data center or cloud provider and wants it to behave like part of the office network rather than a separate external resource.
  • Two companies are merging, or one is opening a new branch, and two previously independent networks need to be joined into one.
  • An office has equipment (a printer, a terminal, an IP phone system) that staff in another location need to reach without exposing ports directly to the public internet.

Which Protocols to Use for Site-to-Site VPN

Most site-to-site VPNs are built on one of three protocols — WireGuard, IPsec/IKEv2, or OpenVPN — and the choice depends on what equipment is already in the offices and how much performance matters.

ProtocolHardware SupportSetup ComplexityPerformance
WireGuardModern router firmware (OpenWrt, MikroTik RouterOS 7+, pfSense/OPNsense)Low — a compact, few-line config on each sideHigh, minimal encryption overhead
IPsec/IKEv2Supported by nearly all office routers and firewalls across vendorsMedium-to-high — more parameters: IKE phases, cipher choice, PFS, key lifetimeHigh, many devices have hardware acceleration
OpenVPNSupported almost everywhere in software, even on older hardwareMedium — its own configuration and certificate formatLower than WireGuard or IPsec, since it runs in userspace

If you are building infrastructure from scratch and speed and simplicity matter most, WireGuard is usually the pick. If you need to connect mixed-vendor hardware in an already-established corporate network, IPsec/IKEv2 is often chosen as the more universal industry standard. Picking a specific router brand or firmware is a separate decision for your IT team — here we are only describing classes of protocols and devices, not recommending a purchase.

How to Plan Office Addressing So Subnets Do Not Overlap

The single most common reason an already-configured site-to-site tunnel "does not route" traffic is overlapping office subnets, not a flaw in the VPN protocol itself.

If office A uses 10.10.0.0/24 and office B uses 10.20.0.0/24, there is no conflict: devices in one network route unambiguously into the other through the tunnel, and every IP address is unique across the combined network. The problem appears when both offices are configured, by default, with the same range — for example 192.168.1.0/24, which is the most common factory setting on consumer and some office routers. In that case the gateway cannot tell which office a packet addressed to 192.168.1.50 belongs to, and traffic either fails to arrive or goes to the wrong place.

There are two ways to resolve overlapping subnets: renumber one of the networks onto a unique range (for example, move the branch to 192.168.20.0/24) — the most reliable fix, though it requires reconfiguring DHCP and some devices; or configure NAT on the VPN gateway, which remaps one network's addresses to a separate technical range only for the tunnel, without touching device settings inside the office — harder to debug, but it does not require changing local addressing. Before choosing an approach, work out the actual ranges and free addresses of both networks with the LiMP VPN subnet calculator — faster than estimating masks by hand.

How to Set Up a Site-to-Site VPN with WireGuard — Step by Step

Below is a minimal setup for a tunnel between two gateways using WireGuard; the commands are the same whether you are on a Linux server or most router firmware with WireGuard support.

  1. Install WireGuard on both gatewaysOn Linux distributions (Debian/Ubuntu), via the system package manager; on a router, through built-in firmware support (OpenWrt, MikroTik RouterOS 7+, pfSense/OPNsense).
  2. Generate a key pair on each sideThe wg genkey and wg pubkey commands create a private and public key for office A's gateway, and separately for office B's.
  3. Confirm the office subnets do not overlap(see the section above) — there is no point configuring further until this is settled.
  4. Configure the interface on office A's gatewayExample configuration:
    [Interface]
    PrivateKey = <office A private key>
    Address = 10.0.0.1/30
    ListenPort = 51820

    [Peer]
    PublicKey = <office B public key>
    Endpoint = <office B public IP>:51820
    AllowedIPs = 10.20.0.0/24
    PersistentKeepalive = 25
  5. Configure a mirrored interface on office B's gateway— with its own private key, office A's public key, and AllowedIPs = 10.10.0.0/24 (office A's network).
  6. Add a static routeon each gateway so traffic to the other office's network goes through the WireGuard interface rather than the default route to the internet.
  7. Bring the interface up and check the handshakewith wg show — the latest handshake field should keep updating, confirming both gateways see each other.
  8. Check the MTU if the connection is unstableDropped large file transfers or video calls over the tunnel are often caused by packet fragmentation — see our guide on VPN protocols for background on how tunneling overhead affects throughput.

PersistentKeepalive = 25 matters especially if at least one side is behind a carrier NAT without a static public IP: without periodic keep-alive packets, the tunnel will go idle and only re-establish when new outbound traffic appears.

How to Verify the Tunnel Between Offices Is Actually Working

After setup, check connectivity at three levels: the tunnel itself, routing, and reachability of specific services.

  • Tunnel check. On WireGuard — the wg show command, looking for a fresh latest handshake. On IPsec — check the security association (SA) state through your gateway's tools.
  • Routing check. ping from a device in office A to an internal IP in office B — if there is no reply despite a live tunnel, the usual cause is a missing static route or a firewall rule blocking traffic between the subnets.
  • Hop-by-hop check. traceroute (or tracert on Windows) shows where traffic stops if a direct ping fails.
  • External port reachability. The tunnel will not come up if an ISP or firewall blocks the needed port: UDP 51820 for WireGuard (or whichever port you configured), UDP 500 and UDP 4500 for IPsec/IKEv2. You can check whether a port is reachable without touching the gateway itself using the LiMP VPN port checker, which shows whether a given port is open on an external address.

Can LiMP VPN Be Used to Connect Two Offices?

Honestly: no. LiMP VPN is a consumer client VPN for individual devices on Android, Windows, and Chrome — not hardware, and not a site-to-site tunneling service between offices. It protects one user's traffic, an employee's laptop at a cafe or phone while traveling, but it does not join entire networks together and does not replace the gateway-level configuration described above.

Where LiMP VPN genuinely helps in a business context is protecting individual remote employees on untrusted networks: a cafe, an airport, a hotel on a business trip. Building a permanent channel between two offices requires an actual site-to-site tunnel on network hardware, as described in this article, not a client application.

Frequently asked questions

How is a site-to-site VPN different from a regular VPN app on a phone or laptop?

A site-to-site VPN connects two entire networks through gateways (routers or servers) and runs permanently in the background. A regular client VPN connects a single device to a server and is turned on manually whenever you need to protect your connection.

Which protocol is better for site-to-site VPN — WireGuard or IPsec?

For new infrastructure where speed and simple configuration matter, WireGuard is usually the choice — the config is more compact and the encryption overhead lower. For compatibility across mixed-vendor hardware in an existing network, IPsec/IKEv2 is often picked as the more universal industry standard.

What if both offices use the same subnet, like 192.168.1.0/24?

There are two options: renumber one office onto a unique range (more reliable, but requires reconfiguring DHCP) or configure NAT on the VPN gateway, which remaps one network's addresses for the tunnel without changing the local addressing of devices inside the office.

Which ports need to be open for a site-to-site VPN?

For WireGuard, usually UDP 51820 (configurable). For IPsec/IKEv2, UDP 500 and UDP 4500. If the tunnel will not come up, the first thing to check is whether an ISP or firewall is blocking exactly those ports.

The tunnel is up, but devices in different offices still can't see each other — why?

The cause is usually not the VPN itself, but a missing static route into the other office's network, or a firewall rule blocking traffic between the subnets even inside an otherwise healthy tunnel.

Do you need a static public IP address for a site-to-site VPN?

At least one side typically needs a static, or at minimum predictable, public address (or dynamic DNS) so the other side can initiate a connection to it as the Endpoint. Setting up a tunnel when both sides sit behind dynamic carrier NAT is considerably harder without extra configuration.

Can you connect an office and a cloud server through a site-to-site VPN?

Yes, this is one of the typical scenarios: a virtual machine in the cloud is configured as a second "office" with its own subnet, and the tunnel comes up between the cloud gateway and the physical office's router the same way it would between two offices.

Does LiMP VPN work for connecting two office networks?

No — LiMP VPN is a client VPN for individual devices (Android, Windows, Chrome); it does not build site-to-site tunnels between networks. Connecting offices requires configuration on network gateways using WireGuard, IPsec/IKEv2, or OpenVPN, as described above.

Author

LiMP VPN Editorial Team

Blog & news editorial team

The LiMP VPN editorial team publishes articles and news about VPNs, privacy and digital security: setting up a VPN on different devices, protocols and encryption, bypassing blocks, staying safe on public Wi-Fi and current threats. Every piece is reviewed by the LiMP VPN service team.

Editorial e-mail: content@limpvpn.com

Read also

Secure your connection in a minute

Download LiMP VPN for free and feel the difference within a minute.