Cellular networks were not originally built for IoT

We have to consider that the networks (and most of the standards) originally were designed with consumer (ie mobile phone) usage as the primary use case. This has many implications, of which here are but just a few:

‍

• Radio coverage is planned around population density and spread.
• The network economics have also been based around consumer models for usage - and the standards and technical designs have reflected that.
• The fundamental concept of mobility across national boundaries is based on the idea of a “home” network and “visited” networks in which the user roams for short periods of time and then returns back “home” - effectively reflecting tourist/business travel use cases.

‍

IoT, however, is fundamentally different

IoT has different needs, however:

‍

• Coverage needs for IoT don’t necessarily reflect consumer capacity planning (eg connected agriculture, manufacturing, shipping/logistics) - potentially putting large volumes of devices in areas where network coverage or capacity is low.

• Many IoT devices are fundamentally constrained (low HW capabilities, no UI, battery capacity limitations) and therefore need to interact with the network in a different manner than smartphones.
• Global IoT is often based on distributed deployment across international borders in a permanent manner (for the whole lifecycle of the device) - there is no concept of the device “roaming” and returning “home”.

‍

So how has IoT managed to work up to this point?

The answer is in many cases not very well. Statistics show that around 75% of IoT projects fail.

‍

But more specifically, IoT has basically have piggy-backed on the consumer networks, but with many issues:

‍

• Coverage/network quality issues, and inability to roam between local “home” networks
• Networks not suited to constrained/low capability devices
• The majority cellular IoT has been based on “permanent” roaming, which brings further issues:

• Latency issues with basic traffic scenarios calling for routing traffic back to the home network
• Permanent roaming restrictions (government regulatory or roaming partner operator policy)

‍

What has been done to address these issues? 

Quite a lot - things such as the introduction of low-power radio technologies such as Cat-M and NB-IoT have increased device compatibility, but are still to live up to their promise more than seven years after their introduction.

‍

Other types of “workaround” solutions have also seen significant adoption:

‍

• Proprietary multi-IMSI and other local roaming solutions
• Roaming with local breakout (LBO)

But these issues also point towards the fact that the cellular roaming architecture is fundamentally unsuited to global IoT.

‍

So what do we need?

We need an  alternate mechanism by which we can smoothly transition devices between operator networks, giving:

‍

• The possibility to switch network providers remotely.
• The capability to switch dynamically to the most suitable network/coverage across all available networks.
• A solution that doesn’t siphon traffic (payload or signaling) across international trunks (and the associated costs with this).
• Keeps traffic as local to the device as possible - minimizing latency issues
• Allows you to comply with permanent roaming restrictions.

‍

Introducing…. eSIM/iSIM

eSIM is essentially the virtualization of your network access credentials that you use to get access to a particular cellular network, and the capability to manage them via digital workflows. It’s basically “virtual” SIM-swapping.

‍

We’re all familiar with the “SIM” card - a small smartcard that we receive from our mobile operator that we insert into our mobiles. eSIM basically embeds this chip into the device, along with the capability to digitally (and remotely) change these credentials over the air - there’s no need to physically swap your SIM card to change your network provider (and in most cases, you can’t anyway - the MFF2 eSIM form factor widely used in IoT is actually soldered into the device).

‍

Additionally, a further evolution in the SIM form factor known as the integrated SIM (iSIM) which basically integrates your SIM into the device processor has come to prevalence.

‍

‍

But eSIM has been around for years?

This technology has actually been around for quite some time - there are actually two fully standardized architectures defined by the GSMA - an m2m push-based architecture based on devices with no UI (GSMA spec SGP.02) which came along in 2013 and a pull-based consumer architecture (GSMA spec SGP.22) which became standardized in 2016.

‍

‍

The m2m spec has been the one most relevant for IoT - it specifies remote triggering of credential (profile) download, whereas the consumer spec relies on a user having a device with significant processing capabilities (like a smartphone) which also has a user interface and alternate connectivity method (such as Wifi or Bluetooth).

‍

But this m2m architecture has been lacking in many aspects (it was primarily driven by requirements out of the automotive industry), not the least of which are:

‍

• It is fundamentally unsuitable for many constrained devices as it relies on SMS/HTTPS connections
• Requires complex backend infrastructure integrations in order to build the operator ecosystem

‍

This, along with MNO’s reluctance to adopt a technology which challenges their traditional business model for customer acquisition and churn, means that much of the deployment to date has been for the automotive industry, and it has remained cost-prohibitive to deploy suitable solutions for many other industries.

‍

Consumer eSIM on the other hand started later but has received more uptake than m2m, mainly because of Apple’s introduction of eSIM-only iPhones and the growth of eSIM in smartwatches, and that the network architecture scales much better (no integration required between MNOs)

‍

A new IoT eSIM standard (SGP.32) has just been released

Work has been going on for several years to define a new IoT eSIM standard to address the shortcomings of the m2m architecture, and now the technical specification is ready and was released in May 2023.

‍

‍

To describe what this looks like - it is based on the consumer eSIM standard but with important additional elements - particularly the remote control of profile provisioning, support for constrained device protocols, and removing the necessity of complex infrastructure integrations to build the operator ecosystem.

‍

It’s important to note that whilst the technical spec has been released, the full set of specifications required for compliance (e.g. test plan and compliance process) are yet to be defined, and are not expected to be released until later in 2024, with supported devices and commercial deployments forecast in 2025.

‍

There has been a lot of buzz around the release of SGP.32 and the barriers to adoption (that were present in the m2m architecture) that it is expected to break down. As a result there is high expectations and as it will be riding the wave of momentum produced by the rise of consumer eSIM it is not without reason that many are predicting that we at an inflection point leading into a phase of hyper growth in eSIM.

‍

In the second part of this article we’ll go into how this is set to shake up the market, and what it means for the main players in the ecosystem.