Wake a computer at home, and reach it from anywhere

The file is on the machine at home. You are not at home, and to get it you will have to wake a computer you cannot see. The obvious answers — email it to yourself before leaving, keep everything in the cloud — only work if you thought of it in advance, and the situation where you need this is precisely the one where you did not.

Reaching a computer at home from elsewhere is four separate problems stacked on top of each other, and people usually solve the wrong one first. In order: wake a computer that has gone to sleep, get into your home network, find the machine once you are in, and do the thing you came to do.

The four layers

Each one has a different answer, and skipping a layer is why setups fail.

Most guides start at the fourth layer. If the first three are not solved, the fourth cannot work — and the first three are where all the setup time goes.

The good news is that layers one to three are configured once and then never touched, while layer four is the part you actually use.

How do you wake a computer over the network?

To wake a computer is to send it a small network packet that its adapter listens for while the rest of the machine sleeps. A computer that is asleep can be woken this way; one that is fully off usually cannot, unless its firmware supports it and it is still receiving power.

To wake a computer reliably, three settings must agree:

• Enable wake-on-LAN in the firmware. BIOS or UEFI, often called "Wake on LAN" or "Power on by PCI-E", and off by default on most machines.
• Enable it in the operating system on the network adapter, which is a separate permission from the firmware setting.
• Use Ethernet if you can. Waking over Wi-Fi works on some hardware and not on much of it, and this single change is the difference between reliable and intermittent.

The packet that does the waking is addressed to the machine's hardware address rather than its IP, because a sleeping machine has no active IP — which is also why it must be sent from inside your network. An app such as Wakivo sends it from a phone; from outside the house it goes over the VPN in the next section, or from a device already at home.

Configure sleep rather than shutdown, and the ability to wake a computer becomes reliable: a machine that sleeps after an hour idle and wakes on demand uses very little power and is always available.

How do you get into your home network?

This is the layer people get wrong most expensively, and there is one right answer for almost everybody.

Run a VPN on your router. Most modern routers can host one, and several open protocols are straightforward to configure. Once connected, your phone or laptop behaves as though it is at home — the wake packet works, the file shares appear, and nothing else needs to be exposed.

What not to do: forward ports to individual services. Exposing remote desktop, a file server or a NAS interface directly to the internet is how home machines get compromised, and automated scanning finds an open port within hours. This advice appears in a great many older guides and has not been reasonable for years.

A third-party remote access service is a legitimate middle path: no networking to configure, and a company sitting in the connection. For most people that is an acceptable trade, made knowingly.

Finding your home from outside

Home connections usually have an address that changes, so the name you connect to has to keep up.

Dynamic DNS gives you a hostname that updates itself when your address changes. Many routers have a client built in; otherwise a small script on any always-on machine does it.

Check whether you have a public address at all. Some connections put you behind carrier-grade NAT, where no incoming connection is possible regardless of configuration. If your router's reported address does not match what a "what is my IP" page shows, this is your situation, and the answer is either asking the provider for a public address or using a relay-based service.

That check is worth doing first, before any other configuration, because it determines whether the direct approach is available at all.

What do you actually need to do once you are in?

Once the first three layers work, four options, and most people need the least complicated one.

1. File sharing. Mount the home folder and copy what you need. For "the file is on my desktop", this is the whole answer and no remote desktop is required.
2. Remote desktop. See and control the screen. Necessary when you need an application that only exists on that machine.
3. SSH. A terminal, for anyone comfortable with one. Efficient over poor connections in a way graphical access is not.
4. A file request to yourself. Sometimes the simplest path is a sync folder that was already running, which is why setting one up beats configuring anything above.

The fourth deserves the last word in this section: a synced folder solves this problem before it happens, requires no networking, and works when your home internet is down. For files you might need, it beats every other option here.

What this costs to run

Worth knowing before setting it up, because the running cost decides whether you leave it configured.

Power. A desktop asleep draws single-digit watts, against 60 to 100 while working. Leaving a machine asleep rather than off costs very little over a year and removes the whole problem of a machine that cannot be woken.

A router that can host a VPN. Many already can. Where one cannot, a small always-on device — a mini PC or a single-board computer drawing a few watts — serves the same purpose and costs less than replacing the router.

Dynamic DNS. Free tiers exist and cover this use entirely; paid options are a few pounds a year and mostly buy a nicer hostname.

Your time, once. Realistically 1 to 2 hours for all four layers if nothing fights you, and most of it is firmware settings and router configuration rather than anything difficult.

The comparison worth making is against the alternative: paying for cloud storage large enough to hold everything you might conceivably need, in order to avoid ever reaching the machine. For a photo or document library measured in hundreds of gigabytes, the home setup is cheaper within a year and keeps the files on hardware you own — which is the other half of why people do this.

Keeping it safe

Four rules, and they are not optional once a machine is reachable.

Nothing exposed directly. Everything behind the VPN, without exception.

Key authentication rather than passwords wherever it is offered, and no password authentication at all on SSH.

Keep the machine patched. A rarely-used machine that is remotely reachable is exactly the machine nobody updates.

Know what wake-on-LAN grants. The packet carries no authentication at all — anyone on your network can wake a computer on it. On a home network that is fine and worth knowing; on a shared network it is the actual security model.

More guides in guides, the tools in online tools, and workflow pieces in digital workflows. The Wake-on-LAN mechanism documents the packet itself if you want the detail.

The short version

Reaching a computer at home is four layers: wake a computer over the network, get into that network, find the address, then do the task. Solve them in that order, because the last one cannot work until the first three do.

Enable wake-on-LAN in both firmware and the operating system and prefer Ethernet; run a VPN on your router rather than forwarding ports to anything; use dynamic DNS if your address changes, after checking you have a public address at all. Then reach for file sharing before remote desktop — and consider that a synced folder would have solved it before you left the house.