Why Double VPN Slowed Everything Down—and the Single Route That Kept My Call Alive

At 9:27, three minutes before a client call, my presentation stopped uploading. The progress bar moved one percentage point, paused, then returned to “reconnecting.” I blamed the rented flat’s Wi-Fi and switched my laptop to my phone’s hotspot. The upload remained stuck, and the meeting preview showed my camera responding several seconds after I moved. Only then did I notice the double-arrow symbol beside the VPN connection.

I had enabled Double VPN the previous evening because it sounded like the setting a privacy-conscious person was supposed to choose.

That instinct had become easier to understand in the UK. After stronger online age checks took effect in July 2025, VPN interest rose sharply. Ofcom research later showed a substantial increase in daily VPN use, while another study found that privacy concerns and distrust of age-verification services were major parts of the public discussion. (RFC 9000)

My own reason was ordinary. I had reached an age-verification prompt while browsing and disliked the idea of handing over an identity document or face scan for a routine online session. I already trusted a large, established VPN provider, so I opened its advanced settings and selected Double VPN.

The feature sent my traffic through two VPN servers instead of one. The provider had a long history, a broad network and detailed privacy tools. Two encrypted stages looked more reassuring than one.

I switched it on and went back to work.

The problem only became visible when the connection had to do something demanding.

Webpages still opened. Email still arrived. Those small successes convinced me that the flat’s Wi-Fi was merely temperamental. The client call exposed the real cost. My video preview stuttered, the presentation upload timed out, and every action inside the shared workspace seemed to pause before taking effect.

I opened the VPN app and moved the exit server closer to me.

The delay improved slightly. The upload did not.

Then I changed the entry server, trying to build a shorter-looking route. That created another reconnect, another interrupted upload and another minute lost before the meeting.

I was changing the map without changing the basic journey.

A normal VPN sends traffic through one VPN server before it reaches the destination. Double VPN adds a second server to that route. The data travels farther, passes through another potential queue and waits for two VPN locations instead of one.

That extra distance is the feature, not an accidental side effect.

Providers offering multi-hop connections acknowledge the same trade-off. Mullvad warns that a second server can reduce speed and increase latency. Proton recommends turning off its multi-server Secure Core mode when connection performance matters.

My Double VPN connection was not broken. It was doing exactly what I had asked it to do: taking the longer route.

For casual browsing, the delay had been easy to overlook. A live call was less forgiving. Audio and video had to arrive continuously. Screen sharing depended on frequent updates. The cloud workspace kept sending small requests in both directions.

The connection did not simply feel slower. It felt hesitant.

That is also the frustration behind many public Double VPN discussions: people enable it because the extra hop sounds like a stronger default, then discover that calls, games, streams and large transfers absorb the cost first.

At 9:29, I stopped trying to optimise the pair.

I disabled Double VPN and connected to one nearby server through the same provider.

The presentation immediately began uploading faster. My meeting preview became usable again. Removing the second hop had recovered most of the lost performance.

Then the Wi-Fi weakened.

My laptop switched back to the phone hotspot, the VPN disconnected during the change, and the upload paused once more. The single route was faster, but keeping it alive on an unstable connection still required intervention.

I could not observe the provider’s internal server load and routing decisions, so I could not assign every delay to one exact cause. The practical result was clear: the second hop made the urgent task worse, and removing it restored the connection.

Now I needed something slightly different.

Not a more elaborate route. A simpler route that could recover when the network underneath it changed.

I closed the first app and opened OnlydogVPN, which I had previously installed as a smaller backup. Instead of asking me to choose an entry country, an exit country and a protocol, it organised the connection around the situation I was trying to handle.

I selected the relevant preset and connected.

The presentation upload resumed.

Then the meeting opened.

My camera preview moved normally. The client’s voice arrived without breaking apart. Screen sharing started on the first attempt.

Halfway through the call, the flat’s Wi-Fi weakened again and the laptop moved back to the phone hotspot. The video softened briefly, then recovered. I stayed in the meeting. The upload continued in the background.

That was the result I needed.

The smaller app uses HTTP/3-based transport, which runs over QUIC and is designed to handle changes in the underlying network path more smoothly. In practice, that meant the connection could survive the move between Wi-Fi and mobile data without forcing me to rebuild the session.

The technical explanation was brief. The practical difference was not.

My work continued.

More importantly, I had stopped forcing an ordinary client call through two separate VPN servers simply because “double” sounded more secure.

Multi-hop still has a purpose. Someone facing a serious surveillance risk may accept lower speed in exchange for separating the entry and exit points. That is a specific protection for a specific threat.

It was not the problem I was solving at 9:30 on a Tuesday morning.

I wanted to keep my traffic away from the local rental network while uploading a file and speaking to a client. One dependable encrypted route handled that task. The second hop added delay without giving the meeting anything it needed.

Once the call ended, I left the smaller app connected and finished the upload. I moved between the desk and the kitchen while the laptop alternated between the building Wi-Fi and my hotspot. I no longer watched the progress bar because it no longer demanded my attention.

The service has fewer server locations, a shorter public history and fewer independent reviews than the established provider. Someone who needs a particular pair of entry and exit countries may prefer the older, larger network.

I did not need a carefully constructed pair.

I needed the fastest reasonable path from an unreliable connection to a completed piece of work.

That difference changed how I thought about advanced VPN settings. I had treated Double VPN as an upgrade that should remain enabled unless something went wrong. In reality, the setting was a specialised trade: more routing separation in exchange for more latency and more opportunities for the connection to slow down.

The extra protection only makes sense when the threat justifies the cost.

For reading sensitive material under an elevated surveillance risk, I might accept that cost. For a live meeting, cloud upload or time-sensitive workspace, it was working directly against the result I needed.

The established provider gave me sophisticated routing controls, and I had activated them without asking whether the task required them.

The smaller app made the useful decision easier. It gave me one route, recovered when the network changed and stayed out of the meeting once the connection was established.

The presentation arrived. The call remained connected. The VPN stopped being the slowest participant in the room.

For that morning’s work, a resilient single route protected more of what mattered than a second hop I could not afford to wait for.