What is multi-WAN?

The term for having more than one internet connection and some logic for how to use them is "multi-WAN." It covers everything from tethering your phone in a pinch to packet-level bonding with persistent IP addresses and enterprise-grade SD-WAN. This guide walks the full spectrum.

A WAN (wide area network) is just the path that connects your network to the internet - fiber, cable, DSL, a 5G modem, a satellite dish, a tethered phone, or even another Wi-Fi network used as a WAN input. Multi-WAN means you have more than one of those paths available, plus some kind of logic that decides how your traffic uses them.

That logic is where the differences start.

At the simple end: "If the cable modem stops responding, activate the LTE backup." At the advanced end: "Keep two tunnels alive simultaneously, split packets across both, and present one stable public IP to the outside world so that nothing I am doing ever notices the switch."

Both are multi-WAN. They are also completely different in terms of what they protect. Two connections alone do not make a setup resilient. What matters is what happens to your traffic, your active sessions, and your public-facing IP address when one of those connections degrades or disappears.

That distinction - between "I have a second internet connection" and "I didn't even notice one went down" - is the central idea of this entire guide.

Why multi-WAN matters more now than it used to

The internet used to be something you could live without for an afternoon. That is less and less true for more and more people.

Remote and hybrid work means more people running live meetings, VPN sessions, and cloud tools over residential connections - where a three-second drop can end a call or kill a file transfer. Small businesses are increasingly cloud-dependent: POS, booking, inventory, and communication tools all running over a single ISP line, with no fallback when it goes down. Households are running more concurrent devices with higher expectations: streaming, gaming, smart home systems, security cameras, all at once. Gamers and streamers feel micro-outages that nobody else in the house notices - a dropped lobby, a spike of packet loss, a stream that buffers for just long enough to lose viewers.

The result is that the gap between "my internet is down" and "something I care about just broke" has narrowed to almost nothing.

The multi-WAN spectrum: every common strategy, from scramble mode to true resilience

Multi-WAN is not one product. It is a spectrum of strategies that range from "grab my phone" to "full software-defined WAN orchestration."

The sections that follow explain each level in detail. We'll explain the technical details and some of the jargon later in this guide.

Level 1: No backup at all

What you are using: Whatever your ISP gave you. One modem, one router, one line.

One ISP. One modem. One router. If the line drops, everything drops. This is the most common setup in the world, and it works fine until it does not. If your outages are rare and your stakes are low, there is nothing wrong with it.

Level 2: Manual failover with your phone hotspot

What you are using: Your iPhone or Android hotspot, manually enabled when your home internet dies.

The classic scramble. Your home internet dies, you enable your phone hotspot, you reconnect the device that matters most, and you hope the call can resume before the other person hangs up.

This works in emergencies. It is not a strategy. It covers one device at a time, it requires you to notice the outage and act on it, and it does nothing for the other 15 devices in your house that are now offline. It is backup in the same way that a bucket is a plumbing system.

Level 3: Cheap carrier or ISP backup add-ons

What you are using: T-Mobile Home Internet Backup, Verizon LTE Business Internet, Comcast Connection Pro, or a similar add-on from your carrier or ISP.

Carriers and ISPs increasingly offer low-cost backup internet as an add-on or standalone product. These range from basic LTE modems that sit on standby to app-controlled backup services.

They are useful. They are better than nothing. But there are limitations worth understanding. These services almost always use low-priority SIM profiles, which means speeds and latency degrade significantly during network congestion - exactly when your primary ISP is most likely to also be struggling. They are also universally capped, either by gigabytes of data or by hours of backup usage per month. And they are single-network: if the one carrier they connect to has a problem, the backup has the same problem.

This tier is solid for people whose outage pain is "I had to wait five minutes" rather than "my client call dropped."

Level 4: Multi-carrier hotspot or roaming-SIM services

What you are using: TravlFi, Solis, Nomad Internet, RoamLink, or similar eSIM/vSIM-based devices.

The premise is one device, one plan, multiple carrier networks. The device picks the strongest available signal and connects to it.

This is genuinely useful for travelers, RV users, and mobile setups where carrier availability changes location to location.

But it is important to be precise about what these devices do. They typically connect to one carrier at a time and switch between carriers as conditions change. That is better coverage, not multi-WAN in the simultaneous sense. You are not running two links at once with failover logic between them. You are picking the best single link available at any moment.

There is also a priority issue worth understanding. The only way these devices can roam across multiple carrier networks is through MVNO (mobile virtual network operator) profiles. MVNOs lease access to carrier towers rather than owning them, and their traffic is universally lower priority than first-party subscribers on the same network. That means during network congestion - exactly the moments when reliability matters most - these connections are the first to slow down. Even when a device promises "premium" access, the underlying SIM profile determines priority, and MVNO profiles sit at the bottom of the stack.

That distinction matters. Better carrier selection reduces dead zones. It does not provide packet-level bonding, persistent IP, session continuity, or carrier-grade network priority.

Level 5: Dual-WAN router - failover, load balancing, and policy routing

What you are using: A dual-WAN router - anything from a GL.iNet Flint or TP-Link Omada to a UniFi gateway, pfSense box, or Peplink - plus a secondary WAN like a cheap cellular SIM or a second ISP connection.

This is where the real technical decisions begin. Most serious multi-WAN routers can do failover, load balancing, and policy-based routing, but the complexity scales with how much of the capability you actually configure and use.

Cold failover is the simplest configuration. One primary connection, one backup. The router monitors the primary, and if it detects failure, it activates the backup. For general browsing, streaming, and non-real-time work, this is often enough.

Load balancing distributes different sessions across both WANs simultaneously - your Google Docs session over fiber, your kid's YouTube stream over LTE. More total capacity, but each individual session still rides one link. If that link dies, the session dies. Load balancing helps busy networks. It does not make any single session more resilient.

Traffic rules let you get granular: work laptops always on fiber, IoT on LTE, gaming on the most stable path. Powerful, but rules interact and edge cases appear - the more you configure, the more you maintain.

Imagine you are on a VPN using your fiber modem's public IP (IP-A). Fiber drops. Your router fails over to LTE. Your new public IP is IP-B. Your router is online - but your company's VPN server is still expecting traffic from IP-A. The tunnel collapses. You log back in and reopen everything.

This is not a bug. It is how most session-based applications work. VPNs, online games, VoIP, livestreams - all behave this way. The failover time itself (typically 1-30 seconds) is only part of the problem. The IP change is often the bigger one.

For a lot of people, this is still the right answer. If your outages are infrequent and your tolerance for reconnecting is reasonable, a dual-WAN router with a $15-30/month backup SIM is a perfectly reasonable solution. Do not overbuy resilience you do not need.

Level 6: DIY Bonding

What you are using: A Peplink Balance router with SpeedFusion Cloud, OpenMPTCProuter on an OpenWrt box with a VPS relay, or Speedify on your laptop - plus two or more SIM plans and possibly an external antenna.

This is the first point on the spectrum where true bonding enters the picture. Unlike failover or load balancing, bonding splits traffic at the packet level across multiple WANs through a tunnel that presents one stable, persistent IP to the outside world. If one WAN fails, the tunnel continues over the remaining path(s) and your sessions never see an IP change. That is the capability that makes VPNs, game sessions, and VoIP survive an outage without dropping.

The tools to build this yourself are real - OpenMPTCProuter, Speedify, Peplink with SpeedFusion Cloud, Raspberry Pi boards with custom firmware. Some of these can genuinely achieve packet-level bonding with session continuity.

The problem is everything else. You need compatible hardware, multiple SIM plans, a relay server (your bonded speed is capped by the server's throughput), and enough knowledge to configure routing, failover policies, firewall rules, and tunnel logic. You need to maintain it all when firmware updates break things, carriers change behavior, or the relay server hiccups. And when something goes wrong, you are the support team - the carrier says it is the router, the router vendor says it is the VPS, and you are the one on a forum at midnight reading a thread from 2019.

DIY bonding can work well. It also needs to be a hobby to be feasible, and it gets expensive quickly. Beyond the cost of the hardware, software, cloud servers, and ISPs - it is your time, your attention, and the reliability gap between "it works on a good day" and "it works when I need it most."

Level 7: Waveform Internet - managed bonding with hot failover

What you are using: A pre-configured Peplink system with high-priority 5G SIMs, SpeedFusion bonding, intelligent routing policies, proactive monitoring, and full-service support - shipped ready to plug in.

This is where multi-WAN crosses from "I have redundancy" to something fundamentally different: internet that does not break, even when the connections underneath it do, without you even needing to think about it.

Until recently, getting this kind of setup meant building it yourself. Waveform Internet is built on the same enterprise-grade Peplink hardware and SpeedFusion bonding, but every layer of complexity is handled for you. The hardware ships pre-configured with first-party, high-priority 5G SIMs - not MVNOs, not resold plans, but the same priority tier as (or higher than) the carrier's own home internet products. A support team handles carrier selection, band optimization, routing configuration, bonding tunnel management, firmware updates, and proactive monitoring - if a carrier degrades, the team can switch networks or adjust routing before you notice. Every system runs forward error correction and packet duplication tuned to each customer's WAN conditions. And the routing policies that determine which traffic gets bonded and which rides the fastest WAN directly - the part that is genuinely hard to get right and harder to maintain - are built and managed for you.

It can work alongside your existing ISP as a managed backup and bonding layer (Waveform Backup Internet), or it can replace your ISP entirely using dual 5G paths (Waveform Internet). The full breakdown of tiers, setup, and pricing is covered later in this guide.

For homes and small-businesses, nothing else we are aware of packages all of this into one managed product. Enterprise options exist (Bigleaf, WiLine, SD-WAN broadly), but they are priced and designed for multi-site deployments. This is the point on the spectrum where you plug something in and your internet just works.

Level 8: Enterprise SD-WAN and managed WAN services

What you are using: Cisco Meraki, Fortinet, VeloCloud, Cato Networks, Bigleaf, WiLine, or similar - typically deployed across multiple sites with centralized IT management or an MSP.

SD-WAN is often mentioned alongside multi-WAN, and there is real overlap - many of these platforms provide bonding, failover, and traffic optimization. But SD-WAN is fundamentally about managing WAN traffic across multiple sites with centralized policy, application-aware routing, and branch-wide orchestration. Services like Bigleaf and WiLine provide similar bonding and failover capabilities for individual business locations, but are priced and supported for companies with IT teams or MSP relationships, not for homes or small businesses buying direct.

SD-WAN is the right answer for a different buyer with a different problem. It is useful as the enterprise benchmark, and it helps frame where simpler multi-WAN strategies fit. But for the core audience of this guide, Levels 5 through 7 are where the real decisions happen.

Here is the full spectrum at a glance. The key differentiators are session continuity (does the thing you care about survive a WAN failure?), management burden (who is responsible when something breaks?), and cost (how much are you paying, and what do you get?).

Failover vs load balancing vs bonding: what these terms actually mean

These three terms are the most commonly confused concepts in multi-WAN. Here is the clearest version.

Failover means your system detects a WAN failure and responds to it. How it responds - and what survives - varies enormously.

Cold failover waits for a failure, detects it, then switches. The backup WAN is offline until needed. Failover typically takes 10-30 seconds on consumer routers. When the switch happens, your public IP changes, and any session bound to the old IP breaks.

Warm failover keeps backup WANs active, but does not bond. When the primary drops, traffic is rerouted quickly - faster than cold failover, but your IP still changes and session-sensitive applications will still drop.

Hot failover keeps a bonding tunnel established across multiple WANs before anything goes wrong. When a WAN fails, there is no activation delay because the other path was already carrying traffic inside the tunnel. Your public IP (the tunnel's persistent IP) never changes. Sessions survive.

Load balancing distributes separate sessions across multiple WANs. A Microsoft Teams session over fiber, your partner's work VPN over 5G. Both links can be active simultaneously, and you get more total capacity, but each individual session is pinned to one WAN. If a WAN fails, sessions on that WAN die. Load balancing is about improving performance on busy networks.

Bonding splits traffic at the packet level across multiple WANs, through a tunnel that presents one persistent IP address. If one WAN fails, the tunnel keeps working through the remaining WAN(s). Active sessions stay alive because they never see an IP change.

Failover is about availability: getting you back online. Load balancing is about distribution: using all your connections at once. Bonding is about continuity: keeping live sessions alive when a connection fails. The best multi-WAN setups combine all three - and the difference between them is what separates "my internet came back" from "I didn't even notice it went down."

How bonding actually works - and why it matters

Bonding is the most misunderstood concept in multi-WAN. Most people hear "bonding" and think "two connections combined into one faster pipe." The reality is more nuanced and more useful than that. This section covers why sessions break during failover, how bonding prevents it, what actually happens inside a bonding tunnel, and what it costs in throughput and latency.

The IP change problem - and the concept that solves it

The root cause is not speed, bandwidth, or failover time. It is that most session-based applications are bound to the public IP address they started on.

When your WAN fails and your public IP changes, the remote server treats the session as broken. Your VPN collapses, your game kicks you, your VoIP call drops. This is why even a "1-second failover" can still mean a full disconnect. The failover speed is not the bottleneck. The IP change is.

Video call clients like Zoom and Teams handle IP changes better than most and will often reconnect automatically. But VPNs, online games, VoIP, remote desktop, and livestreams will kick you outright.

The solution is persistent IP - and it is the most under-discussed concept in backup internet. A persistent IP is maintained by the bonding tunnel. No matter which WAN is carrying traffic at any given moment, the outside world sees one stable IP address. When a WAN drops, the tunnel persists through the other, and sessions never see a change.

Every cold failover system - no matter how fast - changes your IP when it switches. That IP change is what kills your sessions. Persistent IP is what prevents it. The distinction between "has failover" and "has bonding with persistent IP" is the entire difference between "my backup finally kicked in" and "I didn't even notice."

What happens inside the tunnel

Persistent IP is what keeps sessions alive. But bonding does more than just maintain a stable address - the tunnel itself actively cleans up the connection quality of the underlying WANs.

Not all bonding is created equal. There are many ways to configure a bonding tunnel, and the difference between a good setup and a bad one is significant. The following are the key mechanisms that make a well-configured bonding tunnel actually work.

Packet splitting and the reordering buffer. Packets are distributed across your WANs and reassembled in order by a small buffer. That buffer is where the 5-10ms overhead comes from - but it is also why jitter decreases. It absorbs the variance in arrival times across paths. Two jittery connections go in, one smooth stream comes out.

Dynamic path selection. The bonding controller constantly measures latency, jitter, and packet loss across all paths and routes each packet toward whichever WAN is performing best at that moment. This is the real mechanism behind "bonding stabilizes latency" - the system always picks the fastest available path and absorbs the other path's bad moments.

Forward error correction (FEC) and packet duplication. FEC adds redundant data so lost packets can be reconstructed without retransmission. Packet duplication sends the same data across both paths, so a drop on one is instantly covered by the other. A single lost TCP packet triggers retransmission that adds 100-300ms of delay, so eliminating packet loss at the cost of a few milliseconds of overhead is an excellent trade.

The relay server. Bonded traffic routes through a relay server where packets are reassembled. Geographic distance to that relay is the biggest contributor to baseline tunnel latency - a nearby relay keeps overhead at 5-10ms, a distant or underpowered one can push it much higher. This is one of the things that are hard to optimize, and DIY setups often get wrong.

The net result is not necessarily faster internet. It is smoother, less fragile internet, especially on "dirty" cellular links where the problem is inconsistency more than raw speed. A 5G connection that fluctuates wildly on its own becomes stable and usable when bonded across a second path with dynamic path selection, FEC, and duplication handling the inconsistencies.

The tradeoffs: throughput and latency

Bonding is not free. Tunneling traffic adds overhead from encryption, encapsulation, and packet management. There are two tradeoffs to understand.

Throughput: Bonding imposes a speed limit that depends on the router doing the encryption and the relay server handling the tunnel. On the Peplink B One 5G used in Waveform Internet, the hard cap with encrypted SpeedFusion is roughly 200 Mbps; realistic working throughput is around 150 Mbps. A much more powerful router with a dedicated SpeedFusion server could bond at multi-gigabit speeds, but the hardware costs tens of thousands of dollars. For most home and small business use cases, the practical ceiling is 150-200 Mbps.

With Waveform Internet, non-bonded traffic (on a selective bonding setup) bypasses the tunnel entirely and runs at full WAN speed. This is critical, because the applications that benefit most from bonding - VPNs, VoIP, online games, video calls - do not need or benefit from raw speed. A Zoom call uses 3-5 Mbps. A VPN session rarely exceeds 50 Mbps. An online game uses well under 1 Mbps. These are the applications where stability, latency, and session persistence matter, not throughput. Everything else - streaming, downloads, browsing, large file transfers - can ride the fastest available WAN directly at full speed, with warm failover standing by.

Latency: The bonding tunnel adds roughly 5-10 ms of latency depending on geography. For Zoom, VoIP, and virtually all real-time applications, that is imperceptible. For competitive online gaming where single-digit milliseconds matter, it is theoretically noticeable - but the more important effect is that bonding stabilizes latency. The bigger enemy of a good gaming session is not 10 ms vs 15 ms average latency. It is a 15 ms connection that spikes to 200 ms every few seconds because a single cellular path hit congestion. Bonding smooths that out by spreading traffic across paths and absorbing spikes.

The combination - slightly lower peak throughput and slightly higher baseline latency, in exchange for dramatically more stable, interruption-proof performance - is the right tradeoff for a lot of people. And selective bonding (bonding only the traffic that needs it) means you never feel the throughput cost at all.

Building it yourself: what DIY multi-WAN actually takes

The DIY path is genuinely appealing, and for the right person, genuinely rewarding. Buy a Peplink Balance router. Add a couple of SIM plans. Subscribe to SpeedFusion Cloud for bonding. Maybe add an antenna. Tune the failover policies. There is real satisfaction in building a system that works exactly the way you designed it.

Here is what the project actually involves:

The hardware research is substantial. Router compatibility, modem selection, antenna matching, SIM provisioning, carrier band support, firmware versions. Each choice constrains others. The "simple" dual-WAN setup often becomes a three-weekend research project before anything is plugged in.

The configuration goes deep. Failover policies, bonding tunnel settings, WAN health checks, traffic rules, firewall configuration, FEC and packet duplication settings, DNS behavior, and the interaction between all of them. Figuring out which traffic to bond and which to let ride directly - and writing the routing rules to make that work reliably - is its own project.

You are the support team. When something breaks, the carrier says it is the router. The router vendor says it is the relay server. The server says it is the carrier. Nobody owns the whole system. You do.

And it's not even cheaper! A common DIY bonding setup - Peplink router ($650-1000), two unlimited SIM plans ($65/mo each), SpeedFusion Cloud ($20/mo), PrimeCare license ($100/yr/device) - runs roughly $2,865 in the first year, before accounting for your time. For comparison Waveform Internet on the Smart tier runs $2,480 in year one for the same setup, but plug-and-play with world class support.

For people who enjoy network engineering, this is a legitimate and rewarding path. But go in with clear expectations: you are not just buying resilience. You are taking on operational responsibility for a system that needs ongoing attention.

Which multi-WAN setup is right for you?

The right setup depends less on how many connections you have, and more on what happens to your life, your work, or your business when a session drops.

If outages are rare and low-stakes

A phone hotspot or cheap carrier backup line is fine. If you want something a step up, a dual-WAN router (GL.iNet Flint, TP-Link Omada, UniFi gateway, or similar) with a $30-40/month backup SIM gives you real automated redundancy for not much money. Spend $150 once on the router, configure the failover over a weekend, and move on.

"My internet drops maybe twice a year and I can rejoin a call in 30 seconds." This is your tier. You do not need bonding. You do not need anything fancy. Save your money for a problem that actually matches the solution.

If you want automated failover and peace of mind

You have two good options.

• The DIY path: a dual-WAN router with a backup SIM, configured yourself. You source the SIM, set up the health checks, and accept that VPNs, games, and VoIP will drop during the switch.
• The managed path: Waveform Backup Internet Essential ($65/month) gives you the same automated failover on enterprise-grade Peplink hardware with a high-priority 5G SIM, pre-configured, managed, and optimized for you - no setup, no maintenance, and better routing policies than most people would configure on their own.

Both are perfectly reasonable answers. The question is whether the convenience and hardware quality are worth the monthly cost.

"I already have fast fiber, but it drops at the worst times." If the drops are rare and you can tolerate a brief interruption, either of the above is the right setup. If those drops tend to hit during live sessions and the reconnect disruption is a real problem, Waveform Backup Internet's Smart tier adds bonded failover for the traffic that actually hurts when it breaks, without you configuring anything.

"My family just wants internet that works and support that actually helps." A managed solution removes the troubleshooting entirely. No settings pages, no calling your carrier and your router vendor separately and hoping one of them owns the problem.

If you enjoy networking and want full control

Policy-based routing on pfSense, OPNsense, Ubiquiti, or MikroTik gives you granular control over how traffic uses each WAN - prioritize work devices, isolate IoT, tune failover behavior per-device. The tradeoff is real: initial setup typically takes a weekend or more, and when something misbehaves you are debugging firewall logs yourself. If that sounds like a good weekend to you, this is a perfectly reasonable path.

"I already run pfSense with VLANs and custom firewall rules, and I like it that way." Keep it. You do not need a managed solution. Peplink with SpeedFusion Cloud, or OpenMPTCProuter on a secondary box, will give you bonding capability without giving up the control.

That said - if this level of network management sounds useful but you are tired of doing it yourself, Waveform Internet can deliver all of it. You just text us what you want and we configure it.

If dropped sessions cost you real time, money, or trust

This is where bonding matters. Remote professionals whose VPN drops mid-transfer. Small businesses where a POS outage during the lunch rush means lost sales. Gamers who are tired of disconnect penalties. Anyone who needs their sessions to survive a WAN failure, not just their connectivity.

You can build it yourself (Peplink + SpeedFusion Cloud, OpenMPTCProuter, Speedify) or get it managed through Waveform Internet Smart. The DIY path costs about the same and gives you full control but full operational responsibility. The managed path costs about the same and removes all of that.

"My store cannot lose internet during checkout." Bonding with persistent IP eliminates the interruption entirely - no re-taps, no re-swipes, no line backing up.

"I keep getting kicked from ranked lobbies." This is bonding territory. Game servers treat an IP change as a disconnect, full stop. Cold failover will not solve this. Router-level bonding protects every device and every session on the network.

"I am on Zoom all day and cannot risk disconnects." Zoom is fairly graceful about handling IP changes - cold failover is often enough if Zoom is your only concern, though it will not survive every time. But if you are also on a VPN, or your partner and kids are on calls too, router-level bonding protects everything without per-device configuration.

"My DSL is weak and 5G is inconsistent, but together they might work." A DIY Peplink setup with SpeedFusion bonding can aggregate both but is finicky to configure well. Waveform Internet Smart or Complete does the same thing out of the box, with carrier selection, bonding config, and ongoing tuning handled for you.

If nothing can drop, ever

Full bonding - all traffic through the tunnel, all the time. DIY is possible but demanding. Waveform Internet Complete is the consolidated version. The tradeoff either way is realistic throughput around 150 Mbps across all traffic.

If you manage multiple sites

SD-WAN or enterprise managed WAN services (Bigleaf, WiLine, Meraki, etc.). Different buyer, different problem, different price bracket.

Waveform Internet: tiers, setup, and pricing

We have been referencing Waveform Internet throughout this guide because we genuinely think it is an excellent solution for a lot of people in the middle of the multi-WAN spectrum. This section is where we explain the product properly - consider it a light sales pitch, earned (we hope) by the guide that preceded it.

What you're getting

Before talking tiers or pricing, it is worth understanding what the service actually includes, because this is not a router you buy and a SIM you activate.

Every Waveform Internet system ships with enterprise-grade Peplink hardware - the same routers trusted by emergency services, maritime fleets, and field hospitals. Each device arrives pre-configured: first-party, high-priority 5G SIMs installed (not MVNO - first-party profiles on each carrier, with the same or higher network priority as the carrier's own home internet products), settings tuned for your specific location, and monitoring enabled from day one. You do not access the router admin interface. You do not configure anything. You plug it in and it works.

What makes Waveform Internet different from buying the same hardware and doing it yourself is not just the configuration - it is the ongoing management of the entire system. The bonding tunnel and routing policies are the core of what makes this work, and they are where most of the complexity lives. We've developed our routing policies to handle the hard part of selective bonding: determining exactly which traffic gets bonded (VPNs, VoIP, online games, video calls) and which traffic rides the fastest WAN directly (streaming, browsing, downloads, large transfers). Getting those policies right - and keeping them right as applications, carriers, and network conditions change - is a meaningful technical challenge. On top of that, every system runs FEC and packet duplication tuned to the customer's specific WAN conditions, further smoothing out the inconsistencies that make raw cellular connections unreliable.

After setup, we monitor the system continuously. If a carrier degrades, the team can switch networks or optimize band selection remotely. If performance shifts, they adjust routing. If something breaks, we usually know before you do. Configuration changes - WiFi name, password, bridge mode, VLANs, port forwarding, traffic prioritization - are submitted to support and pushed remotely, typically within 15 minutes during business hours (2-hour maximum).

That level of infrastructure and hands-on management is what makes the pricing ($65-$185 per month) land differently than you might expect.

What setup looks like

Backup Internet (you already have a primary ISP): One device ships - a Peplink B One 5G router with a pre-installed 5G SIM. You connect your primary ISP (fiber, cable, DSL, Starlink, or whatever you have) into the router's WAN port. Your home network then connects downstream of the Waveform router - either directly or through your existing WiFi mesh/access points. The Waveform router sits between your existing ISP and your network, managing failover and bonding transparently.

Waveform Internet (Waveform provides the primary connection): Two devices ship - a B One 5G router plus a BR1 Mini 5G modem, both with pre-installed 5G SIMs. Two independent 5G paths from the start. You can add additional WANs, and your network connects downstream of the B One router the same way.

Either way, the hardware arrives labeled with step-by-step instructions. Most customers are online in 10 to 15 minutes.

Essential vs Smart vs Complete

Our approach to bonding is not one-size-fits-all, but we think three sizes cover about 90% of use cases. Each tier is built around a different philosophy of how much traffic needs bonding protection, and the answer for most people is "less than you think."

Beyond the standard tiers, we can customize routing for specific needs. On any tier that includes bonding, we can route traffic by device MAC address - so if you have security cameras whose feeds should always be bonded, or a device that needs a static IP from a single WAN, we can configure that. If your setup does not fit neatly into one of the three tiers, talk to us.

Essential is automated cold failover. If your primary WAN dies, the 5G backup activates. Active sessions that are sensitive to IP changes (VPNs, games, VoIP) will drop and reconnect within 10 seconds. Good baseline protection on enterprise hardware with a managed, high-priority 5G backup - significantly better than configuring a multi-WAN router yourself, but not the right fit if your pain point is dropped sessions rather than dropped connectivity.

Smart is selective bonding - and the heart of what makes Waveform Internet different. Critical traffic gets bonded through a SpeedFusion tunnel with hot failover and persistent IP. Everything else uses the fastest available WAN directly at full speed, with warm failover standing by. This is the sweet spot for most remote workers, gamers, prosumers, and SMB users: it protects the traffic that actually breaks during imperfect failover - VPNs, games, VoIP, video calls, none of which need more than 10-20 Mbps anyway - while everything else (browsing, streaming, large downloads) bypasses the tunnel entirely and runs at full WAN speed.

Complete is full bonding for all traffic. Maximum continuity. The tradeoff is that all traffic runs through the encrypted bonding tunnel, with realistic throughput around 150 Mbps (hard cap ~200 Mbps). Complete is the right answer for surveillance systems, many businesses, and anyone whose requirement is "nothing drops, ever."

Pricing

All plans are month-to-month. No contract. Hardware is separate from service and follows one of two models: pay upfront ($499 for Backup, $849 for full Waveform Internet) and own it after 4 months of service, or rent-to-own ($44/mo or $74/mo) and own it after 12 months. Cancel any time; if you cancel before ownership transfers, return the hardware for a full refund of hardware charges.

Who Waveform Internet is NOT for

Waveform Internet is not for everyone, and we would rather you know that now than find out after buying.

If you are mobile or nomadic, Waveform Internet is a fixed-location service. It does not work in an RV, a truck, or a boat. If you need connectivity on the move, look at multi-carrier hotspot devices (Level 4) or mobile-oriented hardware with an external antenna.

If you are in an area with no usable 5G coverage, Waveform Internet runs on 5G cellular. If none of the supported carriers provide adequate signal at your address, the service will not perform well. We check coverage before shipping hardware, and we're able to provide coverage with good results in locations where most others cannot - thanks to granular band optimization and external antenna options that dramatically improve signal in marginal areas. But 5G availability is still the hard constraint. If there is genuinely no signal, we'll let you know and refund you immediately.

If your outages are rare, low-stakes, and you can tolerate a 30-second reconnect, you might not need bonding or anything fancy. A consumer multi-WAN router with a cheap backup SIM is a reasonable solution at a fraction of the cost.

If you want full control over your router and network configuration, Waveform Internet is managed. You do not get access to the router admin interface. All configuration changes go through support (average response: 15 minutes, maximum: 2 hours). For most people that is a feature - they never want to see a settings page. For someone who wants to write their own firewall rules, configure VLANs by hand, and tune every routing policy, it is a constraint. If that is you, you'll probably be happier building your own Peplink stack, or running OpenMPTCP.

If you need speeds well above 150 Mbps on all traffic simultaneously, full bonding caps throughput at roughly 150 Mbps realistic (200 Mbps hard cap on the B One 5G hardware). The Smart tier avoids this for most traffic by only bonding what needs it. But it is worth asking whether you actually need that much speed everywhere - your Zoom calls use 3-5 Mbps, your VPN rarely exceeds 50 Mbps, and an online game uses well under 1 Mbps. The applications that need resilience are not the ones that need raw speed. If you genuinely need 500 Mbps to every device all the time, bonding is the wrong architecture. Most people do not.

FAQ

What to do next

If you've read this far, you understand multi-WAN better than most network engineers will ever explain it to you. You know the difference between failover and bonding - between a backup that gets you back online and one that never lets you go down in the first place. You know what persistent IP is, and why it matters more than failover speed. You know what the DIY path actually takes.

That puts you in a good position to make the right call for your situation.

And if you've read through the spectrum and found yourself thinking that is exactly what keeps happening to me - the dropped VPN, the kicked lobby, the frozen call at the worst possible moment - then you know where on the ladder the problem actually gets solved.

We built Waveform Internet because we spent a decade helping people fix their internet with antennas, diagnostics, and hardware, and we kept running into the same wall: the hardware was not the problem. The internet itself was the problem. So we built a service that makes internet unbreakable - professional-grade bonding, high-priority optimized 5G, intelligent routing, and real support, shipped in a box that takes 10 minutes to set up. No project. No maintenance. No hoping it works when it matters most.

If that sounds like what you need, try it risk-free for 30 days. If it is not the right fit, send it back - we cover return shipping and refund everything.

Compare Waveform Internet tiers →