Cobham – MNO buyers shift focus to 5G, fundamental differences
Next up, we spoke to network equipment provider Cobham, which was hawking its intelligent distributed antenna system (idDAS) to anyone looking to provide massive wireless capacity in buildings or in specific outdoor locations. To this end, we have spoken to Cobham in the past about the potential for IoT applications, but this year, the tone was a little different.
Hebert Sedas, VP of Coverage Sales Americas, noted that despite the hype, the IoT was not moving as quickly as the wireless industry would like, which is also compounded by the carriers not moving quick enough on the 5G front – something that was clear at MWC in general, as the first 5G handsets are arriving, but which don’t yet have sufficient coverage capacity from the MNOs.
In Sedas’ opinion though, IoT is more of a 5G technology anyway. LTE is still just 2G-type voice but now with a data layer fit for media streaming. He argues that 5G provides a very different opportunity, for very dense IoT networks, with many devices bursting small amounts of data – the opposite of the LTE networks that feature fewer devices (usually phones) that stream large amounts of data.
In addition, the spectrum being used is fundamentally different, at much higher frequencies and with far higher bandwidths available. These are required to support such dense networks, but these require different network architectures in return – with more access points ensure that the spectrum can be served where needed, as these higher frequencies suffer far more from physical obstruction. The new antenna technologies are not suitable for the legacy RF bands, lower down the scale, as they don’t have enough bandwidth.
Solutions to that problem could include some out-there designs like reflecting shields, to bounce signals to where they are needed, but in the longer term, the large arrays found in the RAN assembles need to be scaled down to fit in-building deployments, according to Cobham. Small cells are an answer, but they need to be built out first.
In Sedas’ view, Cobham’s carrier customers are not really looking at the IoT now – they are focused on 5G. As for private networks for specific vertical industries, wealthy verticals such as petrochemicals might have the cash to spend on their own private LTE or 5G networks, but verticals like agriculture are not. The carriers are interested in selling to these verticals, but remain focused on the mass-market consumer world. NAME stressed that in these 5G conversations, it was very important to ask his customers ‘what is 5G to you’ in order to avoid confusion and get on the same page. That’s a very similar concern found in our own IoT conversations.
Napatech – networking FPGA equipment, moving to Xilinx & Altera
Napatech sells FPGA-powered Smart NICs, devices that sit inside servers and provide very powerful processing and offload functions for networking workloads. These NICs are very impressive bits of kit, and we sat down with DJ Root, VP Sales Enterprise and Service Providers, and Daniel Proch, VP of Product Management, to talk through the wider shifts in workloads, as ‘the edge’ gains prevalence in operator discussions.
The gist of that discussion is that there has been a significant change in the past year, with Napatech now working on not just being a Xilinx shop. It is working on being more hardware independent, to this end porting its supporting software over to the Intel (Altera) ecosystem – partly in response to the sometimes religious devotion to one ecosystem over the other, that is encountered in some buyers.
In addition, the firm has ensured that it can offer more software and services to buyers, not just the hardware powering their network traffic processing, and to that end, there are announcements regarding cloud services in the works. Napatech wouldn’t say more, but we envision some form of software-only NIC service.
The shift to software and services is, of course they note, driven by a squeeze on hardware revenues, but that shift can be an opportunity. It would let Napatech move from selling just ‘passive’ NIC functions, such as packet capture in security applications, to more ‘active’ features, such as intrusion protection services. To that end, the FPGA needs a lot of extra supporting services, and that’s where Napatech’s strategy comes into play.
The two noted that it was harder to come by this software expertise in the developer market, that hardware engineers are easier to come by in the current climate. In the FPGA world, the current Xilinx-Intel duopoly seems unshakeable, but Napatech is aware that they would like to sell their own Smart NICs, and that would lead them into competition, or potentially acquisition discussions, with Napatech.
We asked if there was a potential shift on the silicon level, that FPGAs could be displaced by ASIC or GPU designs in Smart NICs. In Napatech’s opinion, the actual process of building these silicon designs is very hard, and gets more complex as the manufacturing process gets smaller – shrinking from 22nm, to 14nm, and so on. A complicating factor is the trend of disaggregating silicon functions, moving them to separate chips on a board, but while you might get a more efficient process by using dedicated ASIC chips there, the need to then coordinate those disaggregated functions would become prohibitive. For Napatech, sitting at the heart of these networks, ‘the IoT’ is just data, at the point where it touches Napatech silicon.
Teoco – SLAs and QoS key for MNO IoT, but confusion still reigns
In the business of providing network quality assurance services to telcos, Teoco’s Danny Itzigsohn, Senior Director Technology and Strategy, and Dima Alkin, VP Service Assurance Solutions North America, outlined the shift that these customers are having to make as 5G looms. Scared of becoming dumb pipes, providing the infrastructure that lets other companies profit over the top (OTT) of their networks, the telcos have looked at services as an answer to this problem.
Teoco has seen an increase in OTT IoT players, but more worrying are the growth of networks that bypass the telcos entirely – and therefore don’t provide an opportunity for them to create those services opportunities. Drones as remote sensing devices would negate the need to pay for MNO connectivity, and IoT satellite constellations could see the LPWAN opportunity pulled away from the mobile networks too. Private LTE or 5G networks exacerbate this problem too.
In turn, there are problems with the business case for IoT devices, from the telco perspective. If connectivity is just 10% to 20% of the total value of an application, the question becomes whether there is a desire to build out the IoT networks themselves. If there’s such a small revenue opportunity in the connectivity, then getting to the scale needed to facilitate the larger services (which need massive scale to justify the whole aggregated data approach) looks tricky. Doing all the hard work just to be a dumb pipe isn’t very attractive, and all that hard work would go to waste if there isn’t a guarantee that IoT customers would pay the telcos for their services higher up the stack. Utilities and Smart Cities looks like good candidates, but the scale of devices that need to be connected to make an application useful is a big ask.
As such, Teoco’s view is that the telcos need to look at where they can provide guaranteed Service Level Agreements (SLAs) or Quality of Service (QoS) contracts, which are likely going to be in areas where their customers are going to be liable if those SLAs are not met. Mission-critical applications require network slicing and orchestration, which is good news for Teoco, as those are sales opportunities.
These new technologies are part of the digital transformation that telcos have to begin, if they are planning on not becoming just dumb pipes. Being able to adjust a network under changing conditions, to support those SLAs for instance, is vital. The digital transformation process is what facilitates telcos being able to compete for the whole application value stack, rather than just being stuck at the 20% or so that might be allocated to only being involved in the connectivity. However, there are definite catch-22 vibes here, as the telcos need to enable that connectivity in the first place, otherwise the applications will not come to the fore. It’s a new dance, as DIMA put it, and one that the telcos have to learn – and one that could open the door for new vendors and network operators to emerge.
EMnify – 1NCE goes public, racks up 700+ customers
EMnify’s Marketing Manager Masha Kovaleva was quite pleased to finally have a partner willing to go public with, noting that most wanted to stay in the shadows. This was not the case, she said, with 1NCE, which has announced that it now supports EMnify’s cloud-native mobile core SaaS offering.
This is a nice win for EMnify, which pitches its services to CSPs that want to launch IoT connectivity services. 1NCE says it has sold over a million SIM cards since its launch five months ago, snapping up customers that were drawn in by its flat-fee offering.
With a simpler fee structure than most MNO offerings, 1NCE is now able to offer customers access to EMnify’s management platform – the interface that lets them monitor and control where their devices are connected. For conventional MNOs, that dashboard would also let you switch tariffs according to needs, but this is not the case with 1NCE’s approach.
Kovaleva said that EMnify now has 700 customers, and is growing well. It has just moved into new offices in Berlin. The 1NCE partnership began at last year’s MWC, and was then launched in August 2018. Frank Stoecker, EMnify’s CEO, said that the dramatic growth of 1NCE was proof of the viability of EMnify’s stack, with 1NCE CEO Alexander Sator noting saying that the company’s plan to launch a radically simpler offering was a response to connectivity conventionally being a rather complicated and confusing part of the IoT value chain.
Axonize – Orchestration platform chases channels, end-users seeing light
Axonize outlined its IoT Orchestration Platform, updating us on its success in the past two years, which has seen it deploy 1,332 projects in more than 30 industries. Smart Buildings and Retail are the top use cases for the technology, which Axonize offers in a SaaS model, with per-device fees.
Built on Microsoft Azure IoT, the platform lets service providers and systems integrators (Axonize’s main customers) offer their customers managed IoT services, focused on Smart Enterprises. The SaaS model lets users start small and scale fast, and also lets the SIs get into recurring revenue streams, rather than just one-off project fees.
Axonize offers the service on a fully managed basis, running it in its own cloud, but will also sometimes let customers run the SaaS in their own clouds. An on-premises implementation is in the works. The platform’s analytics and rules engines can be augmented by ‘IoT Hooks,’ which let customers add other services, such as ML-based analytics, to the dashboard. However, around 95% of customers are content with what is provided by Axonize. NAME said this could, of course, change in the future.
The company estimates that it currently orchestrates tens of thousands of devices through its platform. As for risks, NAME said that end-customers often have a hard time seeing the value of IoT, and therefore Axonize’s SI customers have a hard time trying to convince them of this value. The solution to that problem is closer partnerships, as well as louder marketing of the best use cases as a way to relate the success stories to customers.