The upheaval in the wireless infrastructure chip sector continues, with Intel said to be in talks for the biggest deal yet, a takeover of Altera, which had a market cap of $13.3bn on Friday.
If a bid transpires from the negotiations - leaked by sources to The Wall Street Journal - it would be one of Intel's biggest purchase ever, and would reflect the chip giant's urgent need to reduce its reliance on stagnating markets, notably PC processors, and consolidate its strength in servers (still 98% share, but under the first credible attack in years, from ARM).
This is no longer just about enterprises - Intel is pushing its server processor technology into the carrier network and wireless infrastructure spaces, riding on the interest in virtualization, including Cloud-RAN. It has made several acquisitions recently, including those of LSI's Axxia division and Mindspeed's former Picochip unit, aimed at improving its capabilities in base stations, network processors and C-RAN (even at the cost of introducing ARM-based platforms into the fold).
In this market, and in the broader cloud server segment, Intel has a natural advantage, but needs to build on that quickly to fend off ARM's licensees. Since the UK-based core provider moved to 64-bit, partners like Cavium and EZchip have been pushing the Cortex-A architecture into the heartlands of Intel x86 and of proprietary designs, including high end servers, accelerators and network processors.
Altera makes programmable chips (field programmable gate arrays or FPGAs) for base stations and other infrastructure, and leads this segment, ahead of Xilinx and Lattice, especially in the high end sub-sector which comprises over half the market. In 2013, the company became the largest customer for Intel's fledgling foundry business, signing a deal for the US giant to manufacture new top end parts with its cutting edge 14nm FinFet process. At the time, there was also considerable speculation that Intel would need to expand its FPGA expertise, probably via acquisition, in order to take a lead in base stations, and particularly in Cloud-RAN.
This is because FPGAs are increasingly needed, alongside general purpose processors, to satisfy the demands of performance-intensive infrastructure such as cloud servers and carrier networks. FPGAs can run specific tasks very quickly and are increasingly seen in data center platforms used by Microsoft and Baidu alongside Intel chips. Altera has been the leader in penetrating this high growth segment, and if Intel acquired it, there would be the chance to secure those revenues for itself, and also to integrate CPU and FPGA more closely, to gain competitive advantage against rivals.
In particular, that would cut off a valuable source of FPGAs for ARM-based competitors, though that in turn could boost Xilinx, and possibly make it a target for acquisition too (its shares leapt 6% on the reports of the Intel talks with its arch-rival).
As well as servers, FPGAs are an important element of base stations, and of the emerging virtualized version, Cloud-RAN. Intel has been prototyping a base station platform built around its Xeon processor with accelerators optimized for signal processing. Prototypes based on FPGA chips are being used in China Mobile's huge C-RAN market trial, which aims to virtualize network functions on an x86 server in order to control a large number of small and macro cell sites in a flexible, centralized way.
At this year's Mobile World Congress, Altera and Intel were part of China Mobile's demonstrations of C-RAN based on the NFV specifications. Altera signed a strategic collaboration with the China Mobile Research Institute (CMRI) in 2014, focused on the future needs of 5G with regards to virtualization and FPGAs.
"It has been nearly five years since CMRI first introduced the C-RAN concept to the industry, and there is now wide recognition within the industry that this solution is essential for 5G networks," said Chih-Lin I, CMRI's chief scientist, said in Barcelona. "As a key partner helping to realize the industry vision of the C-RAN architecture-based 5G wireless network, Altera provided advanced technologies and great support to our project as we conducted the research and development together."
Intel has also talked about producing a full base station platform, though to date it has mainly worked with partners to get x86 processors into cell site equipment, as seen in its alliances with NSN to create the RACS/Liquid Apps offering, and Cisco/Ubiquisys for the Smart Cell.
The Altera deal, if it materializes, would pull all these strands together, capitalize further on the important Axxia takeover, and give Intel a strong position when the C-RAN market gains scale (which, despite major trials by China Mobile, Telefonica and others, is unlikely to be until the industry gets close to its '5G' stage, around the end of the decade).
Analysts at Citi told Bloomberg that the acquisition could add 4% to Intel's earnings per share and contribute $2bn in annual revenue, as well as being an effective "fab filler" following recent huge investments in new plants and the 14nm process. It could also give Altera the boost it needs in its largest market, wireless and telecoms equipment, where its sales have recently been flat. It has also seen decline in military, automotive and industrial equipment chips, though the networking business has enjoyed growth and it saw a 12% year-on-year rise in sales in 2014, to $1.9bn, with net income pu 7% to $473m.
Founded in 1984, California-based Altera has more than 3,000 employees in 19 countries. Its largest customers are Huawei and Ericsson, which accounted for 10% of its revenue apiece last year and would be attractive targets for Intel, which has made slow progress in penetrating the heart of the mobile infrastructure world, despite its strength in some types of core network processors
by Caroline Gabriel, Research Director
So Mobile World Congress (MWC) is over for another year, amid the usual record-breaking statistics (from 93,000 visitors, up 45% on the first Barcelona event in 2006; to 7.55Gbps wireless transmission speeds demonstrated by SK Telecom and Samsung).
There were plenty of eye-catching devices, with the Galaxy S6 Edge undoubtedly the star of the show in terms of headline power, though otherwise the mobile gadget space is fragmenting rapidly. The days of a line-up of remarkably similar large-screened smartphones are over - those handsets are there, at ever cheaper price points, but they jostle for attention with virtual reality headsets, connected clothing, smart coffee makers and whisky bottles, and of course the connected cars (Fiat 500 seemed to be the most popular model on display). Indeed, wearables and associated IoT (internet of things) apps virtually colonized MWC's second venue (its previous home in the Fira complex at Plaza Espanya).
Other headlines were sparked by the companies which, back in 2006, when the 3GSM show relocated from Cannes and changed its name, scarcely figured. Google's MVNO plans, Facebook's extension of its internet.org initiative, PayPal's endorsement of NFC with its acquisition of Paydiant - these were the talking points, drowning out the traditional keynote addresses by the major mobile operators.
Traditionally, the CEOs of the established cellcos have used their conference platforms to lay down their demands to the industry (remember then-CEO of Vodafone, Arun Sarin, warning the LTE sector in 2007 to speed up its efforts or face the WiMAX threat; or trading insults with his Nokia counterpart over 3G delays in 2004). These days, it is the new breed of service providers which are setting the pace - Google's Sundar Pichai may have announced a fairly cautious MVNO plan, but his speech had far wider implications, including the call for full WiFi/cellular convergence, still a divisive theme at an event dominated by the entrenched interests of 3GPP platforms.
Those interests are particularly threatened in the IoT, which was a huge theme of the show this year. As the news that Freescale and NXP are to merge neatly demonstrated, this is a dangerous world for the traditional wireless operators and vendors. It throws up significant opportunities to extend their businesses into new, high growth markets, bringing companies like Freescale and NXP - which had been squeezed badly in the smartphone segment - back to Barcelona with new connected device platforms. But the margins on those chips are low and the IoT is already sparking consolidation, as this semiconductor mega-merger illustrates, with the old-school suppliers and operators needing to huddle together for warmth in a business of scale.
Of course, the carrier's network - wireless RAN, core and transport, and increasingly virtualized versions of those - remains the heart of the serious conversations and trading at MWC. With that in mind, we selected our key themes of 2015:
The shape of the new cell site:
After several years when the ever-shrinking base station was the central theme in RAN discussions, this year saw most of the major equipment vendors announcing major refreshes of their macro layers. Massive MIMO (or at least, 8x8 arrays), carrier aggregation across three bands and including TDD, Coordinated Multipoint and Cloud-RAN - these were the important features of the new macro. This was not 5G, but technologies that will be deployable this year or in 2016 - indeed, it seems more than likely that, however '5G' turns out, it will be focused on the dense capacity layer, while the macro coverage umbrella will remain 4G for decades to come.
Small cells were out in force too, and in a widening variety of form factors. Traditional homogeneous mini-base stations are part of a very variegated approach to the capacity layer. They may form clusters with their own controller (local or virtualized) to support an enterprise or a rural deployment. For the former, the big news was that Cisco will resell the Spidercloud Enterprise-RAN solution, despite its own 2013 acquisition of small cell pioneer Ubiquisys. For the latter, the Small Cell Forum kicked off its latest Release Program, devoted to easing deployment issues in rural and remote scenarios, from villages to oil rigs to temporary situations such as disaster relief. Quortus, with its virtualized packet core, was one of the first to update its portfolio to target this important area, while Parallel Wireless was showing off its rural solution, implemented by EE in the UK.
The classic small cell is expanding its reach, seeking to provide greater value than basic coverage and capacity. Ip.access, another of the founders of this industry, has gone as far as to position its Presence Cell purely as the enabler of big data and e-commerce services - and not necessarily connected to the main network at all. Its approach has convinced Vodafone, which announced that it would deploy the retail-oriented platform.
Then there were small cells which did not follow the traditional architecture. Stripped-down antenna/radio units for Centralized-RAN; separate antennas optimized to work with urban small base stations, from companies like Kathrein and CommScope; a converged WiFi/cellular unit from Alcatel-Lucent; hosts of carrier WiFi access points and management platforms as well as lower-power DAS solutions. This is a segment where all options are open, and in which operators will pick and choose the solutions which suit their individual spectrum, business model and capacity requirements.
The virtualization of the RAN is a more distant prospect, for most operators, than the lower risk decision to run a packet core or even a CPE as software on off-the-shelf hardware. However, some pioneers were demonstrating their vRANs, notably Telefonica and China Mobile, and Intel was locked in combat with the ARM ecosystem over the market for high performance processors, optimized for C-RAN servers and accelerators, as the industry chases a general purpose chip with the horsepower to run high end network processes as well as customized silicon.
Not everything can be converted to software of course, though even the physical elements like antennas and radios will be increasingly software-defined and programmable. Pushing that trend to its extreme was Cambridge Consultants, which has developed the IP for the first all-digital radio transmitter, Pizzicato. Unlike conventional software defined radio, it has no analog components, which allows many radios to work together without interference. In the first trial, Cambridge Consultants created 14 simultaneous cellular base station signals at low power, and with the radios "squashed together in a way that analog doesn't tolerate". Such solutions can be programmed to generate manhy combinations of signals at any frequency in an adaptive way. The Pizzicato transmitter consists of an integrated circuit outputting a single stream of bits, and an antenna.
Of course 5G was a massive talking point, though outside the conference halls and the big vendors' glossy demonstrations, there was less hype than expected about the next generation of wireless, with most operators more focused on technology they could deploy in the next 1-2 years, and eager to wait for key decisions at the World Radio Conference in November, and at the 3GPP and other standards bodies, before getting too excited about 5G. Many alliances were formed and roadmaps laid down, but the most tangible aspect of the discussion was the use of millimeter wave spectrum, in which there were many demonstrations for access and backhaul. The high frequency bands are almost certain to play a key role in next generation wireless, and like many supposed elements of 5G, they will start to have a real impact far earlier, as seen in technologies like 60GHz WiGig and some small cell backhaul solutions, notably InterDigital's Peraso baseband system-on-chip for this market.
There was considerable excitement about LTE-LAA (Licensed Assisted Access), which uses 5GHz spectrum for supplemental downlink to a licensed-band 4G network. Although it will not be standardized until next year, supporters like T-Mobile and Qualcomm showed off their plans, along with a companion technology which aggregates a 5GHz WiFi carrier to LTE. Cellular players were trying to dampen down talk of colonizing licence-exempt spectrum, and stressing that LTE and WiFi could coexist peacefully, both in technical terms and in carriers' business models. However, while LAA is clearly a small cell play, given the high frequencies and low power limits involved, some were arguing that the industry would do better to focus on getting 3.5GHz standardized as a specific small cell band, avoiding WiFi showdowns and the quality challenges of unlicensed spectrum.
As noted above, the IoT was an important theme, but given the nature of the event, there was a particular focus on LTE solutions to support IoT applications, and the question of whether these will prove viable as alternatives to WiFi or specialized long range networks such as Sigfox or LoRa. Huawei was demonstrating its contributions to future LTE-M standards, while the LTE-only baseband specialists, such as Sequans and Altair, have a major opportunity to push 4G-only solutions into a mass market. While the 3GPP works on LTE Category 0 as the underpinning of LTE-M, for now the vendors have resurrected Cat-1, whose low data rates made it a Cinderella specification in the broadband world, but whose ultra-low power consumption now makes it a candidate for the cellular IoT. Sequans, Ericsson and Verizon announced that they had run tests on a commercial LTE network, delivering 10Mbps data rates at very low cost and power, and with peaceful coexistence with higher-powered LTE devices.
The new operators:
Facebook and Google both tried to paint pictures in which they had ongoing close alliances with cellular operators, but they managed to visualize a world in which the MNO's role was severely constrained. They are driving new approaches to the network - full WiFi/cellular convergence; harnessing of LTE-Broadcast for social media as well as content; dynamic spectrum allocation on-demand to hundreds of providers; low cost delivery to the 'next billion' world inhabitants. All of these examples see the web giants becoming less over-the-top and actually shaping the network of the future, with the cellcos just providing part of the plumbing, however important that part. The vision will be supported by virtualization and the ability for cloud platforms to support a new generation of network as a service concepts, spanning WiFi, LTE and other connections, and eventually assigning capacity dynamically to large numbers of MVNOs. That is the end game for platforms like XCellAir, which has been spun out of InterDigital. Such services could be run by traditional operators, as AT&T's Domain 2.0 roadmap clearly envisages, but they could equally be controlled by web or IT majors.
The new operating systems:
It isn't all going Google's way though. Android dominated a show in which Apple plays not part (except in everyone's conversations), but the search giant is struggling to control and unify the user experience as large device and service providers create their own user interfaces and developer platforms. Amazon AppStore broke the 400,000 apps mark, for instance, boasting of "huge progress" with its alternative to Google Play. And as smartphones morph into many new types of connected device, many of them driven from the cloud, there may be the chance for different operating systems to break the Android/iOS duopoly. There was considerable interest in the mobile implementations of Windows 10 from Microsoft, while start-up options like Jolla's Sailfish and Mozilla's Firefox Mobile were looking, for the first time, like credible platforms with operator support, not just bright open source ideas.