Facebook’s Telecom Infra Project (TIP) is the most high profile example of the disruptive trend to introduce open source and commoditized technologies into the heart of the mobile network, opening the way to a whole new cost base. Operators such as BT, Deutsche Telekom, Orange and SK Telecom are supporting the initiative, which will bring together open technologies designed within Facebook and by a series of start-ups which will be incubated by the operators.
Some of the initial chosen few have been announced, giving small firms with an innovative approach to building telecoms networks the chance to enter the mainstream. In the past, MNOs have rarely felt able to take the risk of supporting small suppliers in critical networks, but when these are supported by an open, interoperable WiFi-style ecosystem, the risk is greatly reduced – if one vendor fails or is acquired, another’s equipment can be slotted in, or at least, that is the MNOs’ hope.
One of the chosen start-ups is UK-based Zeetta Networks, maker of the NetOS software to support on-demand virtualization of networks, an important aspect of future visions of dynamic network slicing. Zeetta was chosen to take a place in TIP’s UK Telecom Ecosystem Acceleration Center (TEAC), which is backed by BT and hosted in its Adastral Park R&D facility. The contenders pitched to the UK telco, as well as DT and Orange – each of these carriers will support a TEAC in their home country.
The others chosen for the UK center are London-based Unmanned Life, which develops a drone-based platform; and KETS Quantum Security, like Zeetta from Bristol in south-west England, and also working on network virtualization technologies.
Zeetta was spun out of the University of Bristol in late 2015 and has focused on making the slicing and network-on-demand models far easier to support, for the operator or wholesaler, and for the customer. It has mainly used open source technologies and it has an alliance with Bristol’s smart city initiative, Bristol is Open, which is using the NetOS software to allocate resources to a range of services such as UHD video and citywide WiFi. It has raised $7.8m in funding so far, and could now potentially access TIP funds too, though this is not guaranteed for TEAC participants. BT has allocated up to $165m (£125m) for the TEAC activities.
Zeetta CEO Vassilis Seferidis said: “This is a major achievement for Zeetta and opens up tremendous business opportunities to work with tier one telcos and Facebook to further validate our technology and our capability in tackling challenging networking problems with innovative solutions.”
“We’re supporting the development of technologies that have the potential to tackle some of our society’s biggest challenges around connectivity, around security, and accessibility of digital services,” said Howard Watson, CEO of BT’s technology, service and operations division, in a statement.
The TEACs are the main vehicle by which Facebook’s initiative works through established operators. The operators incubate and sometimes fund the start-ups, which become part of the global network which Facebook hopes TIP will become, mirroring its older Open Compute Project for cloud computer hardware. The aim is to have TEACs around the world, which will, as BT describes it, “create a global, sustainable ecosystem that attracts the brightest entrepreneurial minds and innovative investors to work together to produce breakthrough technologies and products in the telecom infrastructure space.”
The idea is that the operators attract VCs to invest in their TEACs. It has been hard, in recent years, to get VC funds interested in network hardware, since the growth was seen to be on the software side of the industry, while the physical networks were dominated by a small group of large suppliers. However, with the confidence instilled by Facebook and operator backing – and the likely prospect of real trials and deployments by that operator – some of the barriers to investment are broken down.
In BT’s case, VCs have invested about $165m to date but BT and Facebook have not invested directly. The VCs include Touchstone Innovations, Atlantic Bridge, Capital Enterprise, Downing Ventures, Entrepreneur First, Episode1, IP Group and OSI (Oxford Sciences Innovation).
On the other side of the world, SK Telecom will fund and host a TEAC in Seoul and has chosen three start-ups as its first partners. The lucky three are Optella, Kulcloud and Transcelestial, which were selected from over 30 applicants.
Optella, founded in 2015 by a group of senior researchers from the Korean national Electronics and Telecommunications Research Institute (ETRI), is now headquartered in California with an R&D center in Korea. It specializes in silicon optical bench (SiOB) photonics and electronics integrated technology, which it plans to apply to a cost-effective, low power 5G wireless optical solution.
Kucloud was set up in 2011, and develops SDN solutions for 5G. In the TEAC, it will focus on developing a technology that satisfies requirements for time sync, QoS and low latency to enable 5G packet fronthaul to be configured on white box switches and using DevOps principles and SDN.
Transcelestial, a Singapore-based start-up, is developing a laser-based technology to link a constellation of nano-satellites using free space optics (FSO) to transfer and relay data for ground, satellite and deep space applications. This aims to achieve 100Gbps speeds, the fastest possible long distance, point-to-point wireless network possible. In the TEAC, Transcelestial will prepare a demonstration of small cell wireless backhaul in a master/slave outdoor urban scenario to prove the reliability and performance of its FSO solution to drive cost-effective and robust 5G infrastructure.
Orange has already made some choices for its own TEAC, while DT has just announced its own plan for a TEAC in Berlin. Orange announced four start-ups earlier this year,
including Amarisoft, whose software can convert a PC to operate as a base station. Others are Athonet, which is developing software-based mobile core platforms; Adipsys, whose WiFi hotspot management system is already in use at Orange; and Horizon Computing, which focuses on cost-efficiency innovations for telco data centers.
These companies, and others in future, will be channelled into TIP via the Facebook alliance – for instance, they will be able to meet other potential operator customers or supporters at the TIP Summit in California in November. They will also be part of a dedicated program for network infrastructure, called Telecom Track, which Orange has set up within its Orange Fab start-up accelerator division and eligible for financial backing from the telco’s investment fund, Orange Digital Ventures, and its VC partners (though the funding is not guaranteed).
The scale of some MNOs’ determination to shake their suppliers out of old complacency was illustrated by Bertrand Rojat, the deputy VP of Orange’s Technocentre R&D unit, who said it was conceivable that start-ups like Amarisoft and Athonet could displace the giants like Ericsson in the future.
Amarisoft’s software virtualizes the base stations and network backbone on standard PC-class hardware. It could also run on open source, low cost appliances designed more specifically for the challenges of the RAN, like the OpenCellular virtualized small cell platform being developed by TIP.
Italy’s Athonet is one of a growing number of companies – including the UK’s Quortus and Germany’s Core Network Dynamics – which implement the evolved packet core on a PC, small cell or other low cost hardware, often to support a localized private sub-network for a vertical industry or for public safety.
“These two start-ups are developing full software core networks and I believe for us it is a very interesting alternative to our regular way of working,” Rojat said. “The challenge is scaling the solution.”
However, while there are clear advantages for a start-up to be exposed to TIP via the support of Orange, there are major dilemmas inherent in that too. Not all innovation in the network is taking place around open source, and many small vendors – as well as many operators – remain concerned by the risks that open source brings in terms of quality, control and patent licensing.
Amarisoft says it does not have an open source model, and has been disappointed in TIP, which has shown “strong resistance” to extending its own approach to include non-open source participants. Spinelli told the Orange event: “We don’t want to let our technology go for nothing. We want money for that … I don’t know what to think about TIP and what your real intention is behind that but I have been disappointed so far with this project.”
There are signs of change on that front. TIP insists it is not entirely open source, despite perceptions to the contrary. While Facebook is open sourcing its own contributions to the project, the option of licensing under RAND (reasonable and non-discriminatory) principles is open to any working group or member.
To emphasize that point, Facebook recently set up a new group under TIP’s auspices, specifically to support RAN licensing. This would avoid deterring would-be contributors, but may also anger members which are committed to open source and believe it is the only way to speed up progress towards a new network cost base – for them, RAND is ill-defined, is mainly associated with old-style standards organizations like ETSI, and can result in time-wasting litigation.
All the TEAC-backed companies will take part in the TIP Summit, to be held on November 8-9 in Silicon Valley and introduce their technologies and solutions to global members of the TIP community. The event will also provide them with a chance to meet and collaborate with start-ups from other TEACs.