Another firm we have a tradition of MWC chinwags with is Altair, but this year, of course, it was not to be. So instead we caught up remotely, with Dima Feldman, Altair’s VP of Product Management and Marketing. In the past year, things have gone quite well for Altair, said Feldman, as the Sony subsidiary looks to capitalize on a growing demand for its IoT-focused cellular chipsets.
The LPWAN market is speeding up, mostly in LTE Cat-M, and Altair customers have launched a number of new products in the past few months, based on designs that are 12 to 24 months old. On this point, Feldman said that time to market was not as short as Altair would like, but that this is getting shorter, which helps customers chase the growing number of device ideas that they have.
Feldman said it was unfortunate that MWC wasn’t going ahead, but that looking back at the decision, it was a good call. Altair had a number of announcements lined up for the show, and wanted to show off a number of new integrated SIM (iSIM) designs – a technology that is now starting to see wholesale carrier acceptance.
However, there is still some considerable division in the iSIM world. Altair is working closely with the GSMA standard initiative, but is also exploring the Deutsche Telekom led nuSIM project, as well as versions from Vodafone, AT&T, and an unnamed Japanese carrier.
We asked if iSIM was helping to reduce the time to market for these IoT designs. Feldman pondered for a moment, before saying that the work needed to accommodate iSIM is essentially a zero-sum game – that this work is about the same level of effort regardless of where it happens in the value chain. However, if it is done at the chipset level, then the work needs only to be done once, so that OEMs don’t have to do it again. Here, Altair wants to do the heavy lifting for its customers, making it much simpler for the device vendors to roll out global products.
Building on this, we asked if iSIM was going to be the solution to the ‘one global SKU’ problem, but Feldman said that while iSIM offers a lot of advantages at the hardware level, being a circuit inside the chipset, there’s still the software ecosystem to contend with. At least with software, you can issue bug fixes, but all the operator profiles are all going to require access to that same bit of integrated circuitry. Consequently, it makes most sense for Altair to ensure its hardware is properly certified, so that all the customers have to worry about is the software.
Feldman explained that the GSMA iSIM process is taking a slightly different direction to nuSIM, although notes that all such standards are compromises of sorts – as the participants have their own agendas. To that end, Feldman is not sure when this standard would be implemented, but believes that the majority of current designs are closer to the GSMA project than nuSIM.
When asked whether SIM cards were a barrier to IoT adoption, Feldman again paused for thought, before saying that they were not. Progress is incremental, he noted, saying that in this game, you win by points rather than a knockout. Here, iSIM isn’t a knockout blow, but it does provide a quicker time to market, and a much better total cost of ownership (TCO). Much better power consumption is also on the table, said Feldman, and in discussions with customers, Altair has yet to find one that explicitly did not want iSIM.
We asked whether this meant that eSIM, from an IoT perspective, was dead on arrival, and likened it to LTE Cat 0 – a technology that was developed and then was emphatically rejected by the market, in favor of the new approach. In this comparison, iSIM would be the LTE Cat-M, to eSIM’s Cat 0, but Feldman thought this was perhaps a little harsh. He said that eSIM was certainly a last-decade technology, and that it is perhaps better to liken it to LTE Cat 1.
Compared to 2019, the 2020 market is much more exciting, according to Feldman, who said that new use cases were finally emerging, enabling new products and services. In particular, smart metering, logistics, and wearable trackers were most interesting.
This brought us to perhaps the most contentious part of the call, from an industry perspective. Feldman believes that LTE networks are going to become major platforms for utility networks, a belief that goes against the traditional wisdom that utilities will not tolerate public networks, because they are uncomfortable with not being able to mandate Quality of Service (QoS) levels, and don’t want to share resources.
Feldman is not alone in this belief, however, as the Utility Broadband Alliance (UBBA) has been making waves. However, the UBBA initiative is focused on private LTE networks, rather than public ones, so there is still a difference here. In Feldman’s view, the major reductions in connectivity prices have led to a step-change in the thinking of some utilities.
Altair has a major utility provider that has to comply with regulation that mandates meter reading services in cities and large buildings, which doesn’t want to rely on unlicensed ISM spectrum because it can’t guarantee that these readings will make it through a noisy and interference-heavy spectrum band. Here, this utility is moving from ISM bands into LTE in order to get those QoS guarantees.
The other trend is a shift in thinking from utilities, who are coming round to the idea of treating these public networks as a sunk cost, which helps drive the cost down. At a global macroeconomic level, Feldman says, the trends are focused on aggregation, and to this end, you would treat cellular networks as existing infrastructure to build on. There is some evidence of this, with new designs for smart meters supporting L-LPWAN protocols, but Feldman stressed that the cost of private networks, from the TCO perspective, can be exorbitant, if you end up having to perform old-school meter drive-by collections with vans, because your private network is struggling with 1% of connections.
Of course, private network supporters would argue that slapping another gateway near these problematic connections is not exactly onerous, but this is a similar argument that crops up in every aspect of the LPWAN market. There are still very strong differences of opinion, and potentially billions of dollars on the line, so it’s not at all surprising that this debate rages on.
Nonetheless, the utility market is a good one for Altair. It has very large volumes, on very stable outlooks, so bidding to replace meters as part of a ten to fifteen year contract can be lucrative. Feldman said that the US, Japan, South Korean, Taiwan, and Australia, are all strong on LTE Cat-M. China is still likely to be completely Cat-NB, with the local networks apparently not even supporting Cat-M, but Europe is a bit more confusing – as MNO choices have not been so uniform.
In terms of rivals, in these markets, Feldman thinks there are going to be more Cat-NB-only rivals, as Cat-M is more complex – meaning that this market would remain the territory of the likes of Altair, Sequans, and Qualcomm. However, most customers actually want dual-mode designs. Huawei and its HiSilicon subsidiary are still dominant in Cat-NB, but Feldman noted that while Huawei now says it has shipped 100mn devices in China, it is about 18-months later than the original timeline.
Given the heavy government support, which include subsidies, this is something to be concerned about. The Cat-NB ecosystem is also hindered by the high-latencies of the end-devices not playing well with public cloud applications that are not used to these network conditions – which leads to a lot of timeouts. Feldman thinks that is a problem that should have been addressed in the standards process.