Two arrows merging into one

US Mobile Announces Multi-Network Plans

Earlier this week, US Mobile announced a closed beta for a new, multi-network offering. Here’s the core excerpt from the announcement that Ahmed Khattak, US Mobile’s CEO, shared on Reddit:

Over the past couple of months, I’ve been thoroughly testing the cellular switching functionality on my phone and have been genuinely impressed. Conducting granular speed tests in areas where I know network performance varies sparked an exciting idea: how do we bring this capability to our customers? In that spirit, I am thrilled to announce the launch of our Multi-Network Unlimited Plans, starting with a closed beta.

With these plans, you can use your unlimited data across multiple networks on a single device that supports DSDS (eSIM/eSIM or eSIM/pSIM). For an additional cost of $15, you can add a line from another network to your device and share your unlimited data seamlessly between both networks.

OS-Level Switching

As far as I can tell, US Mobile plans to rely on the cellular data switching capacities that are available on multi-SIM devices and built into Android and iOS. The approach differs from what we saw with Google Fi’s (now largely abandoned) switching technology that functioned with a single SIM card.

My impression, shared by many Reddit commenters, is that OS-level automated switching between SIM cards is not seamless or optimal. Network switching occurs reliably when an active network becomes unavailable. Switching may occur at other times as well, but the technology falls short of constantly testing both networks and reliably selecting the better network at every moment.

So far, I sound negative. But I’m thrilled about this announcement. While I don’t understand the underlying technology, I expect both Google and Apple will try to improve this technology as more people run muti-SIM setups. I’m glad US Mobile can piggyback on these behemoth companies instead of trying to do tough technical work themselves.

And US Mobile is still pulling off a possibly first-of-its-kind feat in logistics. While users have SIMs on two different networks, US Mobile’s systems can treat this service similarly to to a single plan that draws on one unified pool of data.

Getting Technical

In response to questions from Reddit users, Khattak edited the announcement to add technical details covering how the network switching works:

To optimize the performance of a device employing Dual SIM Dual Standby (DSDS) technology in network selection, the device employs a methodology based on assessing the viability of switching between primary and secondary networks. This assessment begins with the calculation of an estimated link capacity for both the primary and secondary networks. This estimation leverages the instantaneous link capacity data obtained through the device’s radio transceiver.

Subsequently, the estimated link capacity for each network is paired with the device’s data usage, resulting in a congestion ratio representing the ratio of data usage to link capacity for each network. Higher congestion ratio values signify elevated data usage on the respective network and/or comparatively lower link capacity. This indicates a higher likelihood of encountering slower data transfer speeds and diminished performance when utilizing that particular network.

The process of estimating link capacity and device data usage operates over a defined time window, adjustable to accommodate desired sensitivity levels in the estimates. Additionally, a moving average of the congestion ratio is continuously computed for each network, serving as an ongoing reference maintained by the device. This ensures a dynamic assessment of network performance, facilitating informed decisions regarding network switching to optimize the device’s connectivity experience.

I read over that several times, but I may not fully comprehend it. Three metrics are mentioned: (1) data usage, (2) link capacity, and (3) congestion ratio. Link capacity will differ between two networks. While data usage changes over time, I believe it should be the same for both networks at any given moment. Since the congestion ratio is just the ratio of the other two metrics, I think only link capacity matters at the end of the day.

Perhaps more importantly, the explanation seems to describe how switching technology could work in an idealized scenario. I’m not sure how relevant it is to how switching occurs today in the real world. US Mobile appears cognizant of that. In comments on the announcement, Khattak suggests latency may play a role in switching decisions. If latency is accounted for, that may occur outside of the process described in the excerpt above.

Network Switching Is Complicated

My hunch is that US Mobile doesn’t fully understand how OS-level switching decisions are made. And I don’t mean that as a dig. I consider network switching one of the most interesting technologies emerging in cellular, and I don’t understand how it works. The underlying mechanisms could even be changing as Google and Apple update their operating systems.

Anyhow, it’s great to see a new multi-network option hitting the market—particularly one that comes without an outrageous price point.

Network abstract

Network-As-A-Service Model

I periodically write about how the cellular industry in the U.S. might be more efficient if it moved away from the conventional model for network operators. Under the conventional model, network operators have their own hardware placed throughout the country. A typical cell phone user will almost exclusively use his or her operator’s hardware for connectivity.

Other models are possible. Google Fi already allows dynamic switching across multiple networks. Further unconventional models have promise. I love the idea of decentralized networks. What if phones could connect to any cellular base station in an area (regardless of the company that runs it)? Software could automate an instantaneous, auction-like process where each base station in range of a device competes to offer cell phone users the best rates or quality of service.

While there are serious obstacles in the way of a decentralized model, there are rumblings about a less-extreme, network-as-a-service model for the U.S. While it wouldn’t be decentralized, third-party companies with network hardware could rent access to operators. Mike Dano of Light Reading dived into the possibility yesterday:

Just like the tenants of an apartment complex, wireless network operators like AT&T and Verizon pay rent to cell tower landlords for space on their towers.

And one analyst firm [Cowen] suggests that, in the future, those network operators may also rent their 5G radios and antennas from cell tower operators, too…Some executives in the cell tower industry agree that it’s a topic of discussion.

The full article is worth reading.

Abstract photo representing wireless technology

Variable-Rate Pricing, Network Switching, and Mobile X

Urban planners have a joke: “You aren’t in traffic; you are traffic.”

While most people consider how long they’d have to wait in traffic if they travel, almost no one thinks about how much worse they’d make traffic for everyone else.

Conventional tolls charge road users the same rates all the time. Variable-rate tolling is a clever alternative. Under that approach, people pay high tolls when roads are congested. Tolls are low (or non-existent) when roads are wide open. When managed well, variable-rate tolling can lead to huge improvements in efficiency.

Conventional cellular pricing is inefficient

Most of the time, cell phone networks are not at their max capacities. In these situations, a mobile subscriber can use data without degrading service quality for other users on the network or incurring substantial costs for the network operator. On the other hand, network capacity is a precious resource when networks are congested.

With conventional wireless price structures, a gigabyte of data use costs a subscriber the same amount regardless of how congested a network is. There’s a sense in which it would be way more efficient to vary the cost subscribers pay for a gigabyte based on how congested a network is.

With variable-rate pricing, people with money to burn and a need for high-performance could get great speeds all the time. Budget-sensitive consumers could get super cheap data most of the time, then reduce data use when bandwidth is in high demand.

Network switching

If a small town could have its entire population covered by one cell tower, multiple networks may still build towers. In some sense, this is horribly inefficient. On the other hand, it’s unsurprising given the structure of the wireless industry in the U.S. While roaming agreements allowing subscribers to use other carriers’ towers do a lot to reduce inefficiencies like these, the situation is far from optimal. Mobile phone subscribers are at the whims of whatever roaming agreements are in place between network operators.

Imagine an individual T-Mobile subscriber is out of the range of T-Mobile’s network and near another network’s tower. What if the subscriber could pay for temporary coverage from the tower? It’s not an option today, but there’s no technical obstacle making it impossible.

Google Fi uses a form of dynamic network switching that has huge benefits. While Google Fi typically uses T-Mobile’s network, Fi subscribers are automatically switched to Sprint or U.S. Cellular when those networks can deliver better performance.1 Currently, only a tiny portion of U.S. consumers have access to this kind of network switching.

If more carriers embrace dynamic network switching, consumers will benefit. If dynamic network switching is combined with variable-rate pricing, consumers will benefit enormously.

Mobile X

Yesterday, Peter Adderton, the founder of Boost Mobile, began to tweet teasing a new carrier he’s working on called Mobile X:

While the first tweet was vague, it seemed to hint at some of the unconventional features I’d like to see. Today, Adderton shared a more promising tweet:

The image is the part I find most interesting. While I don’t know what Adderton is building, the mockup interface sure looks like it fits with a service that involves both dynamic network switching and user-selected levels of service quality.

SIM Cards

TracFone Teases SmartSIM

TracFone, the company behind several large MVNOs, appears to be working on a new product called SmartSIM. Apparently, some TracFone customers recently received a marketing email that mentioned SmartSIM.1 I was briefly able to access NoDeadZone.com, a website that shared some basic information about SmartSIM. Oddly, the website now automatically redirects to locations.totalwireless.com (Total Wireless is a brand owned by TracFone).

While NoDeadZone.com was accessible, it offered a short video explaining SmartSIM. Apparently, the technology will allow subscribers to switch rapidly between multiple networks based on which network offers the best signal. There’s been some speculation about how up-and-coming eSim technology may enable more people to take advantage of dynamic network switching of this sort. However, the video I gave me the impression that SmartSIM would involve a conventional, removable SIM card rather than an eSIM. At the moment, I’m unsure if TracFone is licensing switching technology Google Fi built, introducing new technology, or something else.

NoDeadZones.com allowed visitors to enter their zip codes to see if SmartSIM was available where they lived. I tried several zip codes, and all were ineligible. It seems that other people had the same experience. I’m not sure whether any zip codes were really eligible for the service.

So far, I haven’t heard of TracFone responding to any requests for more information about SmartSIM. I’m curious about the lack of communication along with the decision to redirect NoDeadZone.com to the main Total Wireless site. It’s enough to make me wonder whether a mistake was made that led SmartSIM to become public knowledge before TracFone intended.

I’ll be keeping close tabs on how the story develops. Dynamic network switching has the potential to improve wireless service and change how it’s priced. With switching technology, it may be possible to charge different rates to different subscribers based on factors like a subscriber’s location, the extent of network congestion, or the quality of service a subscriber receives. Dynamic pricing could potentially lead to far more efficient network usage than conventional pricing—which might ultimately lead to a decrease in how much consumers pay for wireless service.