This article isn't about what a VPN is in general; for that, see our plain-language explanation of VPN technology. It's about the mechanics: what actually happens to a data packet between the moment you tap "Connect" and the moment a website's response shows up on your screen. If you want to see how this works in a ready-made app for Android, Windows and Chrome, check the LiMP VPN features page. Here, we focus purely on the process: authentication, encryption, packet encapsulation, routing, the real difference between full and split tunnel, and how a VPN tunnel actually differs from plain HTTPS.
What happens when you turn on a VPN
Tapping "Connect" triggers a sequence of four steps, and each one adds its own layer of protection: first the device and server confirm they're really talking to each other, then data gets encrypted, wrapped into a new packet, and routed onto the internet.
- AuthenticationThe app on your device and the VPN server exchange keys and verify each other: the device makes sure it's connecting to the real VPN provider's server and not an impostor, while the server checks that the client has a valid subscription or access certificate.
- Encrypting the packetBefore anything leaves your device, each outgoing packet is encrypted locally, using the key agreed on during authentication. At this stage the data becomes a block of bytes that's meaningless to anyone without that key.
- Encapsulation: wrapping it in a new packetThe encrypted data isn't sent out as-is: it's wrapped inside a brand-new network packet whose header lists the VPN server as the destination. This is what tunneling actually means: the original packet, including its real destination, becomes "cargo" inside a new, outer packet.
- RoutingThe outer packet travels across the regular internet to the VPN server like any other traffic, and every node along the way only sees the server's address. At the server, the packet is unwrapped and decrypted, and the real request goes out to the target website on the server's behalf. The response travels back the same way, in reverse.
To any observer between your device and the server (your ISP, café Wi-Fi equipment, any hop in between), all that's visible is an encrypted connection to one consistent VPN server IP. Which site you're actually requesting, and what you're sending, stays hidden inside the outer packet.
How traffic encryption actually works
Encryption relies on two different mechanisms that solve two different problems: asymmetric encryption protects the initial key exchange, and symmetric encryption protects the entire stream of traffic afterward.
During the authentication step above, the device and server use a public/private key pair to agree on a shared secret without ever sending that secret in the open. Even if someone captures the entire exchange, deriving the actual key from it is practically infeasible. That's the asymmetric part, and it only runs once, at connection setup.
After that, every individual packet of traffic is encrypted using symmetric encryption, typically AES-256 or ChaCha20: both ends use the same key, which is considerably faster than asymmetric algorithms and well suited to a continuous real-time data stream. Modern protocols add forward secrecy on top of that: a new key is generated for every new session, so compromising one key never exposes past or future traffic.
The practical effect for a user is simple: anyone watching the network sees nothing but a continuous stream of uniform encrypted bytes, with no way to pick out individual requests, passwords, or page content.
Full tunnel vs. split tunnel: what's the difference
Full tunnel sends all of a device's traffic through the VPN with no exceptions; split tunnel sends only selected apps or destinations through it, letting the rest of the traffic travel the normal network directly.
In full-tunnel mode, absolutely everything (browser traffic, email, background app updates) gets encrypted and routed through the tunnel. That gives maximum protection, but it can use more of the connection's capacity and occasionally slow down traffic that doesn't objectively need encryption, like loading content from a device on your own local network.
In split-tunnel mode, the user or app explicitly defines what goes through the VPN and what doesn't. For example, you might route only a browser and mail client through the tunnel while leaving a game or a local printer's traffic outside it. That reduces the load on the connection, but it also means part of the traffic isn't protected by VPN encryption at all, and that needs to be treated as a deliberate choice, not an afterthought.
| Tunnel type | What goes through the VPN | Typical scenario | Pros / cons |
|---|---|---|---|
| Full tunnel | All device traffic, no exceptions | Public Wi-Fi, handling sensitive data while traveling | Maximum protection; can be overkill for local traffic |
| Split tunnel | Only selected apps or destinations | One app needs protection, the rest is local services | Lighter load on the connection; part of traffic is unprotected |
For most everyday scenarios, such as café, airport, or hotel Wi-Fi, full tunnel is the sensible default: it needs no extra configuration and protects everything at once, with no risk of forgetting to add some app to an exceptions list.
Which protocols make the tunnel possible?
A protocol is the rulebook that the device and server use to agree on encryption, authentication, and the packet-wrapping format described above; the protocol affects speed and stability, not the underlying four-step logic itself.
- WireGuard. A modern protocol with a compact codebase that usually delivers the highest speed and reconnects fastest after a network change, for example when switching from Wi-Fi to mobile data.
- OpenVPN. A time-tested, open-source protocol; slightly slower than WireGuard, but it tends to get through restrictive or unusually configured networks well, especially over TCP.
- IKEv2/IPsec. A protocol built for mobile stability: it re-establishes the tunnel quickly when the network changes, which suits phones that frequently switch between Wi-Fi and cellular data.
A detailed comparison of each protocol's speed, security, and best-fit scenarios is its own topic, covered in VPN protocols explained. For most users, modern apps pick a suitable protocol automatically, so there's no need to decide manually.
How is a VPN tunnel different from HTTPS?
HTTPS only encrypts the connection between your browser and one specific website, and that site's address is still visible to your ISP; a VPN encrypts all of a device's traffic and hides the addresses of every site you visit from the ISP entirely. These are two layers of protection with very different scope, and neither replaces the other.
When you open an HTTPS site without a VPN, your ISP can't see the page's content, but it can see which domain you're connecting to: that's exposed through connection metadata such as SNI in the TLS handshake. With a VPN on, the ISP only sees an encrypted connection to the VPN server's IP and can't tell which site is actually being requested inside the tunnel.
The practical takeaway: HTTPS protects the content of one page from interception along the way, while a VPN additionally hides which sites you're visiting at all, and protects a device's entire traffic, not just the browser. Running both at once costs you nothing, since they work independently and complement rather than duplicate each other.
Where a VPN is actually used
In practice, the tunnel described above solves three recurring problems: protection on untrusted networks, privacy from your ISP, and securely handling work data while traveling.
- Open Wi-Fi in cafés, airports, hotels. These networks are rarely protected by an individual password, and traffic on them can theoretically be seen by any other client on the same access point; the encryption step above closes exactly that risk.
- Home networks and routers. Turning on a VPN at the router level extends the entire process described above to every device on the network (phones, laptops, smart speakers) without configuring each one separately.
- Remote work. Sending work messages, documents, and accessing internal company systems over a random network while traveling is a scenario where the extra layer of encapsulation and encryption reduces the risk of interception along the way.
- General privacy from your ISP. Even on a home network, a VPN limits what an ISP can technically learn about the domains you visit, which is basic digital hygiene rather than a reaction to a specific threat.
In all of these cases a VPN is solving a privacy and connection-security problem; it isn't marketed or treated here as a way to reach any particular site or service.
What a VPN doesn't protect against
An honest assessment matters here: a VPN encrypts and routes traffic between your device and the server, but it isn't responsible for everything else happening on the device itself or on the service you're using.
- Phishing and social engineering. If you type your password into a fake site yourself, tunneling your traffic won't stop that, since the page still receives whatever you entered.
- Malware already on the device. Encrypting the connection doesn't stop malware that's already installed from reading data straight off the screen or out of memory.
- Weak passwords and no two-factor authentication. A breach through a compromised account has nothing to do with connection encryption; that calls for separate password hygiene.
- Full anonymity. A VPN hides your IP address and traffic route, but it doesn't make you completely unidentifiable: accounts, cookies, and the services themselves keep working as usual.
In other words, a VPN covers exactly the segment of the chain described in the authentication, encryption, encapsulation, and routing steps above. It doesn't turn a device into something universally secure on its own.
How to check that the tunnel is actually working
You can check whether the tunnel described above is really active through three indirect signs: whether the IP address sites see has actually changed, whether your real DNS address is "leaking" around the tunnel, and whether your browser reveals your real address via WebRTC.
Walking through each of those checks in detail here would be redundant, since there's a separate step-by-step procedure covering 6 tests, laid out in how to check if your VPN is working. The short version: an app showing "Connected" doesn't by itself prove that all your traffic is actually going through the encryption and packet-wrapping process described above. Only a dedicated check gives you real certainty.
Bottom line
A VPN works as a four-step process: authentication, on-device encryption, wrapping the packet inside a new envelope addressed to the server, and routing through that server onto the internet. Full tunnel sends all of your traffic through this process; split tunnel sends only part of it. HTTPS, unlike a VPN, only protects a single connection to one specific site rather than the whole device. If you want a ready-made app with this kind of tunnel for Android, Windows, or Chrome, the LiMP VPN features page has the full list, and current subscription terms are on the LiMP VPN pricing page.



