The third Telecom Infrastructure Summit was held in London this year, away from its roots in California. This Facebook sponsored initiative has garnered substantial momentum since its launch two years ago, with over 500 member companies. I'd estimate over 1,500 attended this two day bonanza, up from 1,000 last year. TIP offers a route to market for budding start-ups and innovators of cellular radio access equipment, many of which you'd consider as small cell product and component vendors.
I attended only the first of this two day event, focussing on cellular coverage rather than capacity. The second day was oriented more towards Facebook’s Terragraph project that intends to use unlicensed 60GHz Wi-Fi to provide high capacity free wireless internet in urban areas.
The TIP project has been primarily driven and funded by Facebook, strongly supported by BT, DT and SKT. There isn’t much support from any of the US telcos, although they are warmly invited to participate. Initial funding and continued support by Facebook has been critical to the project with a long term view that it will become self-sustainable from telco and member support in the medium term.
From its first summit at Menlo Park (Facebook HQ) two years ago that attracted 300 delegates, last year’s grew to 1,000 and this year seemed to me to expand to around 1,500. There appears to be more interest outside the US than in, which given their initial mission to bring the internet to the unconnected in developing countries seems not entirely surprising.
A rather vague introduction suggested that the scope of the project has expanded to cover almost every aspect of mobile and fixed telco infrastructure. There followed an excellent joint presentation from senior CTOs at Vodafone and Telefonica concentrating on their more immediate intent. They want to shake up their infrastructure supply chain and allow a route to market for innovative new startups, recognising that it is extremely difficult to gain a foothold as a new RAN vendor. Ask any budding small cell vendor from the past 10 years – they should know how difficult it is to do business with a large telco.
They’d like to introduce these potential new vendors in a low risk route, initially by testing and deploying in rural and underserved areas. Parts of LATAM and Africa were mentioned as initial targets. They are happy starting off with 3GPP standard releases, say one or two behind the mainstream, and then work to reduce the delay. It was suggested that 3GPP Release 9 might be more than adequate to provide good voice and data service to an unserved rural community. Earlier trials in India have shown that after testing and debugging, new entrants can match the performance of incumbent vendors. One performance chart showed call drop rates down to 0.5% after a few weeks.
Once proven, these new vendors could then be introduced into more mature markets. As one speaker said, users in areas with no coverage would tolerate occasional downtime or poor performance while users in developed mature markets expect undisrupted and unrestricted service at all times.
RFI winners announced
These two multinationals had previously launched an RFI (Request for Information) for three main trials and announced the winners live on stage. They were looking not just for a 4G solution but sought a platform that could provide 2G, 3G and 4G. Undeveloped markets have many users with older, low cost phones and simpler requirements – battery life being one of the most important factors – where GSM is still a popular and low cost solution.
The winners were Parallel Wireless, Mavenir and Altiostar. Several pilot trials will be launched over the coming months. There was a separate RFI for an LTE relay box, similar to Huawei Libero or Airspan Magic Box. Surprisingly neither of those companies was involved and the winner was a Lime Microsystems led consortium of partners.
Below, I was one of the first to congratulate Rajesh Mishra, CTO of Parallel Wireless, on their success.
It’s unclear how much money or what size of commercial opportunities lie ahead for these companies, but its clear that many of the players originate from the Small Cell ecosystem.
Somewhat unusually for an RFI, Vodafone and Telefonica were quite public about what they were ranking each vendor for and which had done best.
Compliance has substantially improved over the past year as this chart illustrates
The best performing vendors were categorised
The event was also good for networking with others in the industry. I saw many of the same faces I’ve seen at Small Cell conferences over the years, representing everything from chipsets, backhaul, SON, and small cell vendors, other analysts to deployment/installers. As always, this led to many useful and interesting side conversations.
There were also parallel panel sessions, several live demos. 60GHz mmWave technologies featured prominently. 5G was mentioned, with the suggestion that any TIP platform should be software upgradeable from 4G to 5G. That will risk increasing the cost of the hardware substantially – this is far from a generic processor compute function – and frankly just not needed for this application.
So it appears the initial focus for TIP really is for remote and rural areas serving the undeveloped parts of the world. There won’t be too much money in that in the short term, but if these up and coming vendors can impress the big multi-national networks, then there may be a route into a bigger slice of the pie.
Slicing too thinly
One notable aspect was that where initially they had thought of the architecture as simply involving a separate hardware and software component, OpenRAN had now expanded into six:
- Node Integrator
- Network integrator
The number of different project areas has expanded into a huge chart of 20 or more modules. Everything from edge computing to Ai and virtual RAN. One cynic suggested that business will be done primarily through the node or network Integrators rather than with individual companies, leading to a very similar business model used by incumbents.
The Call to Action from our two leading telco presenters is shown below. Of these points, minimising the scope of the project variants where possible is key - fewer 3GPP releases, modes, options and interfaces all make life simpler. At the same time, some serious consideration will be needed about how such new vendors would be integrated into existing networks, highlighting the bullet point about more focus on O&M (Operations and Maintenance) that helps create this new value chain.
The potential problem I foresee is that an expanding scope will lead to a lack of focus. Simply because more delegates attend these events doesn’t necessarily mean that progress will be made. RFIs themselves may provide funding for research and development, but not on the scale that telecom component vendors are used to. A baseband chipset can cost $100 million in R&D, limited sales volumes are the reason why few of the original small cell chipset specialists remain in the business. A fully mature and debugged small cell product can also cost a similar amount.
I do hope that the winners of this first RFI will be nurtured and admitted to the wider telco industry supply chain. I doubt they would see much of the 5G action, but this isn’t necessarily a problem and indeed perhaps best avoided.
Time will tell whether this event next year can demonstrate substantial progress or if it becomes just another talking shop.