Showing posts with label OND. Show all posts
Showing posts with label OND. Show all posts

Monday, September 17, 2018

Interview: Habana Labs targets AI processors

Habana Labs, a start-up based in Israel with offices in Silicon Valley, emerged from stealth to unveil its first AI processor.

Habana's deep learning inference processor, named Goya, is >2 orders of magnitude better in throughput & power than commonly deployed CPUs, according to the company. The company will offer a PCIe 4.0 card that incorporates a single Goya HL-1000 processor and designed to accelerate various AI inferencing workloads, such as image recognition, neural machine translation, sentiment analysis, recommender systems, etc.  A PCIe card based on its Goya HL-1000 processor delivers 15,000 images/second throughput on the ResNet-50 inference benchmark, with 1.3 milliseconds latency, while consuming only 100 watts of power.

Habana is also developing an inference software toolkit to simplify the development and deployment of deep learning models (topologies) for mass-market use. The idea is to provide an inference network model compilation and runtime that eliminates low-level programming of the processor.

I recently sat down with Eitan Medina, Habana Labs' Chief Business Officer, to discuss the development of this new class of AI processors and what it means for the cloud business.

Jim Carroll:  Who is Habana Labs and how did you guys get started?

Eitan Medina: Habana was founded in 2016 with the goal of building AI processors for inference and training. Currently, we have about 120 people on board, mostly in R&D and based in Israel. We have a business headquarters here in Silicon Valley. In terms of the background of the management team, most of us have deep expertise in processors, DSPs, and communications semiconductors. I previously was the CTO for Galileo Technology (acquired by Marvell), and now I am on the business side. I would say we have a very strong and multidisciplinary team for machine learning. We certainly have the expertise in the processing, software and networking to architect a complete hardware and software solution for deep learning.

In building this company, we identified the AI space as one that deserves its own class of processors. We believe that the existing CPUs and GPUs are not good enough.

The first wave of these AI processors are coming now or being announced now. Habana decided that unlike other semiconductor companies, we would only emerge from stealth once we have an actual product. We have production samples now and that is why we are officially launching the company.

Jim Carroll: Who are the founders and what motivated them to enter this market segment?

Eitan Medina: The two co-founders are David Dahan (CEO) and Ran Halutz (VP R&D), who worked together at Prime Sense, a company that was acquired by Apple. We also have onboard Shlomo Raikin (CTO), who was the Chief SoC Architect at Mellanox and who has 45 patents. We've also been able to recruit top talent from across the R&D ecosystem in Israel. The lead investors are Avigdor Willenz (Chairman), Bessemer, and WALDEN (Lip-Bu Tan).

Jim Carroll: What does the name "Habana" refer to?

Eitan Medina: In Hebrew, Habana means "understanding" -- a good name for an AI company.

Jim Carroll: The market for AI processors, obviously, is in its infancy. How do you see it developing?

Eitan Medina: Well, some analysts are already projecting a market for new class of chipsets for deep learning. Tractica, for instance, divides the emerging market into CPUs, GPUs, FPGAs, ASICs, SoC accelerators, and other devices. We see the need for a different type of processor because of the huge gap between the computational requirements for AI and the incremental improvements that vendors have delivered over the past few years, which right are just small improvements to CPUs and GPUs. Look at the best-in-class, deep learning models and then calculate how much computing power is needed to train them. Look at how these requirements have grown over the past few years. Trying graphing this progression and you will see a log scale graph with a doubling time of three and a half months. That's 10x every year. Initially, people were running machine learning on CPUs, and then they adopted Nvidia's GPUs. What we see in the market today is that training is dominated by GPUs, while influence is dominated by CPUs.

Jim Carroll: So what is Habana's approach?

Eitan Medina: When we looked at the overall deep learning space, we began with the workflows. It is important to understand that there's a training workflow, and there's an inference workflow. What we are introducing today is our "Goya" inference processor. Our "Gaudi" training processor will be introduced in the second quarter of 2019. It will feature a 2Tbps interface per device and its training performance scales linearly to thousands of processors. We intend to sell line cards equipped with these processors, which you can then plug into your existing servers.

The inference processor offloads this workload completely from the CPU. Therefore, you will not need to replace your existing servers with more advanced CPUs. What can this do for you?  This is where our story gets really interesting. We're about more than an order of magnitude improvement.

Look at this graph showing our ResNet-50 inference throughput and latency performance. On the left side is the best performance Intel has shown to date on a dual socket Xeon Platinum.  Latency is not reported, which could be a critical issue. In the middle is Nvidia's V100 Tensor GPU, with shows 6ms of latency -- not bad, but we can do better. Our performance, shown on the right, exceeds 15,000 images per second with just 1.3ms of latency. Our card is just 100 watts, whereas we estimate at least 400 watts for the other guys.

Jim Carroll: Where are you getting these gains? Are you processing the images in a different way?

Eitan Medina: Well, I can say that we are not changing the topology.  If you are an AI researcher with a ResNet-50 topology, we will take you topology and ingest it to our compiler. We're not forcing you to change anything in your model.

Jim Carroll: So, if we try to understand the magic inside a GPU, Nvidia will talk about their ability to process polygons in parallel with large numbers of cores. Where is the magic for Habana?

Eitan Medina: Yeah, Nvidia will say they are really good at figuring out polygons, and may tell you about the massive memory bandwidth they can provide to the many cores. But, at the end of the day, if you are interested in doing image recognition, you only really care about application performance, not the stories of how wonderful the technology is.

Let's assume for a second that there's a guy with a very inefficient image processing architecture, ok? What would this guy do to give you better performance from generation to generation? He would just pack more of the same stuff each time -- more more memory, more bandwidth, and more power. And then he would tell you to "buy more to save more".  Sound familiar? This guy can show you improvements, but if he's carrying that inefficiency throughout the stack, it is just going to be more of the same. If a new guy comes to market, what you want to see is application performance. What's your latency? What's your throughput? What's your accuracy? What's your power? What's your cost? If we can show all of that, then we don't have to have a debate about architecture.

Jim Carroll: So, are you guys using the same "magic" to deliver inference performance?

Eitan Medina: No, but for now, I want to show you what we can do. The lion share of inference processors used by cloud operators today are CPUs -- an estimated 91% of these workloads are running on CPUs. Nvidia so far has not come up with a solution to move this market to GPUs. The market is using their GPUs mainly for training.

Our line card, installed in this server, can ingest and process 15,000 frames per second through the PCI bus.  Because our chip is so efficient, we don't need crazy memory technologies or specialized manufacturing techniques. In fact, this chip is built with 16 nanometer technology, which is quite mature and well-understood. As soon as we got the first device back from TSMC, we had ResNet up and running immediately.

In a cloud data center, three of our line cards could deliver the inference processing equivalent of 169 Intel powered servers or eight of Nvidia's latest Tesla V100 GPUs.

Habana Labs is showcasing a Goya inference processor card in a live server, running multiple neural-network topologies, at the AI Hardware Summit on September 18 – 19, 2018, in Mountain View, CA.

Wednesday, June 6, 2018

FWD: Huawei's Global Connectivity Index 2018

Cisco regularly publishes its Visual Networking Index and Ericsson does likewise with its Mobility Report -- both are well-regarded across the industry for their primary data and forecasts. Huawei is doing the same with its Global Connectivity Index, which is now in its fifth annual edition.

Huawei's Global Connectivity Index provides a big picture look at digital transformation and presents its data on a country-by-country basis.

This year, the GCI report provides a quantitative assessment of the digital health of 79 nations, ranking them by considering their progress in 5 technology enablers: broadband, data centers, cloud, Big Data, and IoT. There are four pillars: supply, demand, experience and potential. The study relies on expertise from inside of Huawei as well as outside the company, including in-depth interviews with top think tanks around the world, including digital strategy advisors, digital economy academics, and policymakers from the World Bank, MIT, Stanford University, and Singapore Infocomm Development Authority. Huawei argues that the long-term return on investment (ROI) for digital technologies is 6.7 times that of non-digital investments.

The biggest takeaway from the study is that the global digital economy could nearly double in size to $23 trillion by 2025 from $12.9 trillion in 2017, when it accounted for 17.1% of global GDP.

One outside observation is that a country's connectivity ranking is related to its GDP but not bound by it. Another is that Huawei's own remarkable performance in certain developing countries must certainly have helped to list their ranking but often not enough to push them into front-runner or even adopter status -- and for the U.S. market, which holds the No.1 spot on Huawei's Global Connectivity Index, the company is essentially locked out. 

Some highlights:
  • The U.S. continued to lead in ICT Infrastructure investment and held onto its top position in the rankings.
  • Singapore and Sweden trailed close behind the US, and over the past year closed the gap in rank with the US by two GCI points each. 
  • Qatar, China and Malaysia improved their GCI scores in part due to roll-outs of national ICT initiatives. 
  • The study identifies several countries seeing a drop in GCI rankings, mostly these nations are not yet investing, adopting and capturing the potential of advanced technologies including Cloud, Big Data and IoT.
  • The study finds growing inequality, an ICT version of the “Matthew Effect” – the sociology theory that states: “the rich get richer and the poor get poorer.”

Huawei's Global Connectivity Index can be accessed here:

Friday, June 1, 2018

A look at COBO and the drive for on-board optics

The Consortium for On-Board Optics (COBO) was formed three years ago to develop specifications to permit the use of board-mounted optical modules in the manufacturing of networking equipment, such as switches and even servers. The specifications will cover electrical interfaces, pin-outs, connectors, thermals, etc.  The idea is to drive the development of interchangeable and interoperable optical modules that can be mounted onto motherboards and daughtercards. Microsoft is a founding member of COBO, as are Arista, Broadcom, Cisco, Juniper, and Mellanox. In April, COBO issued its first on-board optics specification targeting 400 Gbps and 800 Gbps data rates and leveraging two electrical interfaces: 8 and 16 lanes of 50-gigabit PAM-4 signals.

I caught up with Brad Booth, president of COBO, at the recent NetEvents 2018 in San Jose, California. In his day job, Brad is also a Network Architect with Microsoft Azure.

Jim Carroll, Optical Networks Daily – Hi Brad, good to see you again. Tell us about your work and the latest developments in the drive toward on-board optics.

Brad Booth, COBO:  I am the president of COBO. That is my volunteer job. My full-time job is a Network Architect for Microsoft Azure. 

A few months ago, COBO, the Consortium for Onboard Optics released our specification 1.0 specification in time for OFC and now we're going around letting people know about it because it is a game changer for how to use embedded optics. Traditionally, embedded optics have been proprietary solutions. We knew that as we progress the technology and move from 10G to 40G to 100G to 400G, and whatever speed comes next, the ability to continue to use faceplate optics is presenting greater complexities. 

First, the thermal environment was getting a little harder to deal with. Signal integrity was becoming more difficult to deal with. The fact that you had people plugging these modules in, which were very sensitive to ASD and sensitive to noise, resulting in damage. And we decided that we needed to start the progression of moving the optics away from the faceplate and closer to the ASIC, if not on to the same footprint of the ASIC eventually. We knew that at 400G it was possible to do faceplate pluggable. We knew in time people would figure out how to do an optics embedded with ASICs, but what was going to be the thing that would allow us that transition in between. What would allow us to have the learnings and the understanding of how this will change our business models. That’s why we came up with COBO.

What is COBO implementation actually going to look like?

Brad Booth, COBO: This is a COBO module. It is a by 16 wide, which means it is a 16-lane interface on one side. This module was designed for the output of four 100G PSM4 signals  This is the medium size version of the module. It comes in a by-8 and a by-16 version. It comes in three lengths: slightly shorter than this, this length, and slightly longer. That allows us to be able to use this module literally from multimode fiber all the way up to long-haul coherent transmission. 
The module at its widest form factor is capable of accepting up to 40 watts of power to drive whatever component tree we need in it. 

It's capable of having a full 1RU heatsink, or greater if we want. You can even do liquid cooling on this if you want. There's the capability of latching either on the high-speed pins or the low-speed pin connector on the bottom. We separated the low-speed power and ground and control away from the high speed to improve the signal integrity. The nice thing is these pins on the bottom are .8mm pitch, which means that we get lots of good contact and power. There is also .6mm pitch, which should be able to handle up to 100 Gbps PAM4 electrical signaling. This means that future iterations of this module will be capable of 1.6 Tbps worth of bandwidth.

What kind of support or feedback are you getting from the industry?

Brad Booth, COBO:  We have over 60 member companies participating in COBO now. Many of them are very well known in the industry. Cisco, Arista, Juniper. Mellanox. We have connector companies involved too., like SEMTECH, TE Connectivity, Luxtera. We have companies that specialize in optical connectors, like SENKO, Huber+Suhner, etc. We have a strong ecosystem. 

We have to figure out how to move those optics from the faceplate further inside the systems to improve the signal integrity, and we have to be able to do it without drawing massive amount of power. And that's one of the advantages of having so many people in the ecosystem that play in this area, we have a lot of good contributions.

Filmed at NetEvents 2018 in San Jose, California

Is COBO targeting 400G and above? Are higher rates going to be supported? 

Brad Booth, COBO: This module is actually a four by 100. So it has 400 gigs worth of bandwidth in it right now. If you were to build it to handle 400G Ethernet, you could put two 400G Ethernets in it, and so that would be 800G worth of bandwidth. When this electrical interface goes to 100G PAM4 in the next version of our specification, this module will be able to do 1.6 terabytes. 

Is that the endpoint? Probably not, because we've actually designed the module to be able to handle the coherent light, or the 400 ZR, which is being worked on by the OIF. That's a 15 watt part and it does 400G per lambda. If the power of that technology continues to drop, you could potentially put more of those in this and actually achieve even greater bandwidth. 

What about the manufacturability of COBO? How is the industry progressing in this regard? 

Brad Booth, COBO: The interesting aspect of building something like this is being able to dissipate the power and the heat. The biggest thing for heat dissipation is actually the maximum height that you can make the heat fin on the heat sink. So, if you get a full 1RU height, you get significantly better thermal performance than just a small one. We've done some general evaluations here. We’ve actually done some thermal modeling. As a next step, we're looking forward to building out true thermal systems and testing it. 

One of the key aspects of this is that you've got a full 1 RU height. Second, you're not blocking the faceplate with optics, and you don't need as much airflow to pull through the system. That means you don't need fans running full out just to be able to pull enough air through to cool it. Another aspect is that COBO will allow you can to change how you implement your system design. You could place these modules in different spots across the motherboard that are potentially cooler or further away from the hotter ASIC. You could implement airflow channels to cool the optics, separate from the actual switch ASIC. 

There is the possibility of using liquid cooling. This module doesn't have a heat sink on it right now, so you could drop a liquid cooling plate on top of it. There's also the expansion capability provided by extra pins on the bottom of the module, which could be used to provide more power and more control signals if necessary for some next-generation technologies. 

What about the challenge of fiber alignment with COBO?

Brad Booth, COBO: Regarding fiber alignments, an interesting aspect is that we expect most people in the first generation to build these modules with pigtail optics. In other words, the optics will already be pre-attached. This module, which is just a mock-up, is shown without them. 

We have work going on to actually do a connectorized interface. We have some people working on how that would actually be implemented. If we can connectorize it here and connectorized it on the faceplate, that means that you wouldn't have to worry about it at manufacturing. You could attach the module in the factory when they build the switch. They could put the optics right in. Then, after the optics and systems are all installed, you could come and hand fiber, trace it out to the pigtails or to the front faceplate and that would actually help eliminate the worry, if you have to replace this, that you have to unstrand the whole pigtail. This provides the option for people to be able to do both. We’ve discussed even short pigtail versions within our optical conductivity study group. 

We invite people to contribute their thoughts and opinions. Come and work with us on this because this is an area that we think is going to become critically important as we start progressing towards actually putting optics embedded with ASICs.

Wednesday, May 9, 2018

FWD: The death of ZTE

Zhongxing Telecommunication Equipment Corporation (ZTE), one of the world's largest suppliers of network infrastructure products, informed the Hong Kong Stock Exchange that "the major operating activities of the Company have ceased". 

If the notice means what we think it means, then ZTE is dead.

It took only 3 weeks from the day that the U.S. Commerce Department' Bureau of Industry and Security (BIS) issued its order prohibiting companies or individuals from participating in any way in an export transaction with ZTE for this multinational giant based in Shenzhen to collapse. I'm not sure there has been any other corporate collapse in the networking sector of this magnitude and in this accelerated time frame. The nearest comparison would be the collapse of Nortel in 2009, but that took years to occur rather than just weeks.

ZTE's website has already started to disappear. Many product, technology and news archive pages are now gone.

The most proximate reason for the death of the company is that without new deliveries of chipsets and optical components from U.S. vendors, the manufacturing lines for ZTE must have already come to a halt, leaving the company unable to ship products. Just-in-time manufacturing probably means that the company has insufficient inventory to sustain operations during a protracted appeal or legal fight with the U.S. Department of Commerce. More importantly, if the market has lost confidence, the sharks smell blood, and normal operations become impossible.

There will be a scramble amongst investors, creditors, employees, competitors, suppliers, and customers to secure whatever value remains in the organisation. There should be plenty.

Salvaging the good bits

First, there is a huge installed base of ZTE equipment worldwide in carrier networks, in data centres, in enterprise IT centres, and in home networks. The value of this equipment could be in the tens of billions if we take a cumulative count of sales over the last four years. These networks, which belong to the customers and their lenders, will need to be supported.  There is ongoing business here for someone.

ZTE holds the No.1 or No. 2 markets share position on many of the core infrastructure projects of the big three carriers in China -- China Mobile, China Telecom, and China Unicom.

All of ZTE's product segments were growing. Here are the 2017 annual growth rates:

  • Carrier networks 8.3%
  • Gov't and corporate 10.4%
  • Consumer 5.2%

Outside of China, ZTE has many current sales contracts and open purchase orders for new equipment, with good prospects of upcoming fibre broadband, 4G, 4G, and core network projects. 

ZTE has a very extensive telecom equipment portfolio, covering every sector of wireless networks, core networks, access & bearer networks, services and terminals. 

ZTE has been listed on the Shenzhen Stock Exchange since 1997 and on the Hong Kong Exchange since 2004. Trading has been suspended since April 16 and so there is no way to know quite yet if the shares are now worthless. The company's balance sheet at the end of 2017 showed RMB 31.647 billion in current assets, and the company's most recent statement said it was conserving cash.

ZTE has abundant in-house and contracted production facilities capable of manufacturing large volume of smartphones, customer premise equipment, and carrier infrastructure products. 

ZTE has many current and next-generation product designs using the latest silicon from U.S., Japanese, Korean, Taiwanese and other international suppliers. The product designs could be sold to other equipment suppliers.

ZTE has a considerable patent portfolio. ZTE claims to amongst the most prolific corporate patent filers in recent years. As of 30-June-2017, ZTE Group 68,000 patents, including 29,000 granted global patents.

As of mid-2017, the company was operating 20 R&D centres in China, the United States, Sweden, France, Japan and Canada, as well as more than 10 joint innovation centres established in association with leading carriers.

There is a talented pool of 74,773 employees (including 58,940 as employees of the parent company), with an average age of 33. Many of these employees have deep subject matter expertise, the vast majority of whom had nothing to do with the business decisions that got ZTE into trouble. 

ZTE was gearing up for a big play in 5G

At this year's  Mobile World Congress in Barcelona, ZTE captured the “Best Technology Innovation for 5G" award for its end-to-end vision encompassing the radio access network, the core network, bearer platforms, custom 5G silicon and CPE terminals. As with other suppliers, many of these are “works in progress” rather than commercially deployable solutions right now.

ZTE's has pushed hard on Massive MIMO, the antenna technology which has been shown to improve spectral efficiency up to 8 times.

It has been pioneering a multi-user shared access (MUSA) technology to effectively increase the number of connections served, and thereby enable support for scenarios involving mass connectivity with low power consumption. This could be extremely useful in very crowded areas, such as subway systems, when everyone is using their smartphone. The MUSA technology works by allowing high overload and eliminating scheduling operations, thereby increasing the number of connections by between 3- and 6-fold. It uses advanced spread spectrum sequence and SIC technology to simplify terminal implementation and help reduce energy consumption.

In the network core, ZTE is ready to commercialize end-to-end 5G network slicing. Its Cloud ServCore platform implements lightweight micro-service components to enable the network slices to operate independently and with easy scalability. This will allow IoT applications, for instance, to scale smoothly and without impacting other network slices.

ZTE is also readying a 5G Flexhaul bearer solution based on next-gen FlexE technology. Part of this vision to achieve a unified bearer network for 3G / 4G / 5G traffic. ZTE says its 5G Flexhaul achieves end-to-end protection switching time of less than 1ms, as well as single node forwarding latency of less than 0.5μs.

ZTE was a $20 billion company on the rise

Prior to receiving the death sentence for sanctions violations and lying to the U.S. government during a probationary period, ZTE was profitable and on a $20 billion per year sales run rate.

For Q1 2018, the company reported revenue of RMB 28.879 billion (US$5.548 billion), up 12% over the same period in 2017. Net profit after extraordinary items attributable to holders of ordinary shares of the listed company amounted to RMB 1.368 billion (US$216 million). For the full year 2017, ZTE reported operating revenue of RMB 108.82 billion, 7.49% higher than a year earlier,  

Net profit for 2017 was reported at RMB 4.55 billion, an increase of 293%. Net cash flow from operating activities for 2017 was approximately RMB 6.78 billion, about 28.88% year-on-year growth. This was a quite a recovery from 2016, when revenues grew just 4% and profits were lower. With booming handset sales in China, India and other developing markets, along with good prospects for 5G, things were looking pretty good for ZTE, until its troubles with the long-running exports violation case came to a head.

Big fine in 2017

ZTE's 2017 results were impacted by troubles with the U.S. government. In March 2017, ZTE made penalty payments of over US$1.19 billion to the U.S. government-- this too for the case involving the shipment of U.S.-origin technology to Iran during the period of economic sanctions. ZTE plead guilty in the case and paid the fine. It also agreed to a number of other conditions, which were not fulfilled, according to the U.S. Commerce Department, or which ZTE subsequently lied about. 

Huawei as the beneficiary? 

ZTE generates about 40% of its revenue abroad. 

We can surmise that many of the large carrier projects that ZTE currently has underway internationally will have been funded by the Bank of China,  the China Development Bank (CDB), or other government-backed, export/import financial institutions. These carrier customers are facing the prospect of suspended or canceled projects. This presents an opportunity for other network vendors to step in and capture the business. 

However, the customer would need to secure another funding source. Huawei is the most likely to be the ZTE replacement, especially if the Bank of China or CDB were to transfer project loans on their behalf. Ericsson, Nokia, Samsung and others also have an opening to entice these ZTE carrier customers with their offerings.

But what if Huawei is next?

However,  it is conceivable that the Trump administration will ratchet up the pressure on Huawei, for instance by extending all of parts of the ZTE export ban to them, or by persuading other governments to block Huawei as has been done in the U.S.. Many analysts expected that the ZTE ban was a bargaining chip in the recent, first round of trade negotiations between the U.S. and China. There was, and perhaps continues to be, hope that the order would be rescinded after the trade talks. This did not happen. Perhaps the trade tensions will get worse, with Huawei coming under pressure next.

With this possibility at hand, some large carriers in countries such as Japan, Germany or Singapore, may rethink their future plans with Chinese equipment vendors in general on critical projects so as not to face supply disruptions like we now see with ZTE. In Germany, Deutsche Telekom recently announced a 5G pilot deployment in Berlin using Huawei equipment. In the U.S., T-Mobile is prohibited from using Huawei as a supplier. With T-Mobile now seeking to merge with Sprint, U.S. regulators conceivably could require the German parent company to remove all Huawei gear from all of its networks as a condition for approving the merger. 

In other countries, there will be other geopolitical considerations. In Russia, ZTE has just clinched a 70% share of the first stage of  Rostelecom's the access network modernization project. ZTE's Multi-Service Access Network (MSAN) product delivers VDSL. Rostelcom is currently testing G.vectoring and for deployment in a second stage of its upgrade project. Rostelecom, of course, is Russia's leading broadband and pay-TV provider with over 12.7 million fixed-line broadband subscribers and over 9.7 million pay-TV subscribers, over 4.7 million of which are subscribed to its IPTV service. Given the need for Rostelcom to complete this network upgrade successfully, on-time and on budget, they will look for other suppliers.. but probably not from the U.S.

In India, ZTE is now a major supplier of low-cost smartphones and optical transmission gear. In October 2017, ZTE announced a 100G WDM Backbone Network Project and metro area network (MAN) construction contract with Idea Cellular, the third largest mobile operator in India with 189 million subscribers. With this deal, ZTE’s OTN optical transport platform captured a 95% share in the metro optical backbones that carry Idea Cellular’s traffic. ZTE has previously disclosed major contracts with Bharti Airtel as well. This success comes despite some protectionist voices in India warning against Chinese suppliers for critical network infrastructure.

Other recent contract wins include the Ooredoo Group, which serves 164 million customers across the Middle East, and South Africa based MTN. For Ooredoo Group, ZTE was expected to supply end-to-end networks, applications, and terminals in preparation for a 5G launch. MTN was also looking at deploying ZTE’s 5G NR radio access, 5G virtualized network slicing, carrier DevOps and container-based vEPC, and 5G Flexhaul bearer network.

The ZTE effect on suppliers

For those companies who were supplying chipsets, optical components, memories, display technologies, protocol stacks, etc. to ZTE, there will be a waiting game to see who takes up the slack. We can presume that the size and growth of the market will remain the same before and after this incident. If ZTE doesn't supply that core router, someone else will. 

What comes next?

The ZTE statement about ceasing normal activities holds out a glimmer of hope that the U.S. government might hear an appeal and grant a reprieve. Last week, U.S. trade negotiators visited China. Obviously, no deal occurred or ZTE would not have made its statement. 

Whatever comes next, it better happen quickly because sales contracts and talented employees will not stick around to what eventually emerges. The best people and ideas will move on to competitors or new ventures.

The most likely outcome is that ZTE individual business units are sold off, spun out, or otherwise reorganised into new corporate entities. In other words, the same cast of characters with the same products but operating under a new name.

Sunday, April 22, 2018

It may not be too late for cloud giants to enter the mobile market

It was over seven years ago, in December 2010, that NTT DOCOMO launched its 4G LTE network. Japan, of course, was already heavily saturated with 3G coverage and mobile penetration rates were among the highest in the world. DOCOMO’s 4G network was an instant success and very quickly itd subscribers upgraded their phones and moved onto mobile data plans. DOCOMO's network grew and grew, and so did those of its competitors – KDDI and Softbank Mobile. Since then, the Japanese population has not given up their mobile devices. Like everywhere, people are checking their phones all day long, from the moment they awake till late at night.

With their upcoming launch of commercial 5G services over the next 2 years, DOCOMO and KDDI are looking for history to repeat. They will be among the first operators worldwide to deploy the next generation of mobile technology and they hope the market will respond. But, there is a surprise twist. A new entrant, Rakuten, plans to launch a new 4G network by October 2019 – nearly nine years behind the market leaders, and in a market that seems oversaturated and with this little prospect of growth for the “old wave” technology.

This week, Rakuten received approval from Japan’s Minister of Internal Affairs and Communications to launch a new 4G network.based on its own mobile base stations.

Rakuten Mobile Network will operate over the 1.7 GHz frequency band with over 1,825 MHz of spectrum. The company aims to launch service in October 2019. The company plans to raise a maximum of JPY 600 billion to fund the rollout of the network. Of this, Rakuten Inc. (the parent company) plans to provide a maximum of JPY 200 billion - a hedge on its bet.

Rakuten – Japan’s e-commerce giant

Founded in 1999 by Hiroshi Mikitan, Rakuten is Japan’s e-commerce leader – the local equivalent of Amazon or Alibaba – but far from being a me-too follower, the company has consistently innovated and acquired to advance its vision. It now offers online merchandise for consumers and businesses, life insurance, fire insurance, travel insurance, digital content, an advertising network, and a growing list of communications services. Rakuten also operates the country’s biggest Internet bank and third-largest credit card company by transaction value.

Outside of Japan, Rakuten’s  major acquisitions include (now in the US), PriceMinister (France), Ikeda (now Rakuten Brasil), Tradoria (now Rakuten Deutschland), (now in the UK), (now Rakuten TV in Spain), Kobo Inc. (now Rakuten Kobo in Canada), Viber (now Rakuten Viber), Ebates, Viki (now Rakuten Viki), OverDrive, Inc. (now Rakuten OverDrive), Slice (now Rakuten Slice) and The Grommet.
Until now, Rakuten has experimented with being a mobile virtual network operator, claiming 1.5 million users. Going forward, Rakuten reckons that around JPY 600 billion is enough to build a nationwide network of 4G base stations. The company says it has poached key executives from the other three big mobile operators. It is also known to be seeking advice from network equipment suppliers about how to rollout a nationwide network as quickly and efficiently as possible.

A key metric for Rakuten is its global gross transaction value, which is the sum total of everything sold on its platform. For 2017, that figure was up 21%.

Strategic thinking

Given its current size and the deep pull from its consumer base, one might expect that it would be easier and faster for Rakuten to buy out one of the three existing mobile operators compared with the time and trouble of building a whole new 4G network.  Practically speaking, no such option exists in Japan for Rakuten. NTT Docomo obviously is out of reach. KDDI is doing well enough on its own as the main challenger to Docomo, so is unlikely to be interested. And Softbank Japan, with Masayoshi Son at its helm, has big ambitions of its own, making the prospects of a merger or buyout with Rakuten unlikely.

For Rakuten, the value of becoming a mobile operator is not to battle it out with Docomo in hopes of poaching subscribers and earning a thin margin on the sale of monthly 4G data plans. Rakuten’s strategic thoughts must centre on building direct access to its e-shopping consumers.  It might even be willing to accept losses in the first years of operating the forthcoming mobile network, in return for a building a better e-commerce experience for its consumers.

In India, Reliance Jio is another late-comer to the mobile market and it too has been willing to suffer deep losses to build its new nationwide network and to establish its subscriber base. However, in Jio case, there is not a thriving e-commerce business to justify the risk.
There are however other cloud giants who will be watching Rakuten’s entrance into the mobile business. Alibaba might be constrained in doing so in its home market, unless the government wants a fourth competitor. But conceivably Alibaba could try its luck as a mobile operator overseas. The other big player of course is Amazon, who might very well be studying Rakuten’s moves.

Wednesday, April 11, 2018

Xtera in conversation with Verizon - part 1

The evolution of optical transport technology in terrestrial and subsea networks as we approach Shannon's limit is the topic of discussion in this conversation with Glenn Wellbrock, Director, Backbone Network Design, Verizon, Stuart Barnes, Chairman and CSO, Xtera, and Vijay Rudravajjala, VP Engineering, Xtera.

See video:

Xtera in conversation with Verizon - part 2

In this segment, the conversation turns to network disaggregation trends. A key question that is raised by network operators is "what value am I going to get out of disaggregation if I am not interested in becoming an equipment manufacturer or system integrator?"  The value may be in standardizing APIs and driving network agility.

On the issue of standards, the topic of how to accurately quantify the capacity of open lines is raised, especially when spectral sharing is employed.

Glenn Wellbrock, Director, Backbone Network Design, Verizon
Stuart Barnes, Chairman and CSO, Xtera
Vijay Rudravajjala, VP Engineering, Xtera

See video:

Tuesday, April 10, 2018

Video: MACOM in conversation with Verizon

MACOM, which has a long history in telecommunications, has moved quickly in recent years to expand its presence in photonics and advanced semiconductors, says Vivek Rajgarthia, SVP, and GM, Lightwave for MACOM.

This conversation with Glenn Wellbrock, Director, Backbone Network Design, Verizon, explores innovation and volume production at the component level, and how this impacts the overall network. 

See video:

Sunday, April 8, 2018

ADTRAN in conversation with Verizon

Unlike FiOS, Verizon's "One Fiber" program is deploying is a higher-count cable deep in the network that can be used for all services, whether enterprise, residential, wireless or internal, says Glenn Wellbrock, Director, Backbone Network Design, Verizon.

In this video, Glenn and Robert Conger, AVP Cloud and Portfolio Strategy for ADTRAN, discuss next-gen network architecture.

Tuesday, March 20, 2018

Telefónica's Telxius infrastructure arm expands its global reach

Telxius, Telefónica's infrastructure arm, was established in February 2016. It owns and operates a portfolio comprising nearly 16,300 telecom towers in five countries and manages an international network with around 65,000 km of submarine optical cable, including around 31,000 km owned by Telxius. The Telxius-owned network includes SAM-1 linking the U.S., Central and South America, PCCS (Pacific Caribbean Cable System) and Unisur, which connects Uruguay and Argentina. It also took over Telefónica's share in older subsea consortium cables, including Columbus III, Atlantis II, and FLAG.

Telxius has two new subsea cables that are expected to enter commercial service shortly: BRUSA, linking Brazil, Puerto Rico and the U.S., and MAREA linking the U.S. and Europe in partnership with Google and Facebook. The two projects bring enormous East-West and North-South capacity to the Atlantic. More about these projects below.

It was just over a year ago that Telefónica announced it has reached an agreement with global investment firm KKR Group for the sale of up to a 40% stake of Telxius Telecom, its global telecommunications infrastructure company, for a total of Euro 1,275 million, or Euro 12.75 per share. The deal with KKR implied an enterprise value of Euro 3,678 million for Telxius, or 11.4 times its 2017 EBITDA.

Telefónica's original plan was to seek a public listing for Telxius as an independent company. An attempted initial public offering was announced in September 2016 but subsequently withdrawn in November.

In terms of valuation, the deal with KKR was fairly close to the earlier aspirations for a public listing. KKR is a private equity firm specializing in infrastructure opportunities. As of the end of 2017, KKR had $168 billion of assets under its management.

From what we gather, the Telxius shared infrastructure business is doing well.

The Massive MAREA project

Construction of the highest-capacity subsea cable to cross the Atlantic was officially completed in September 2017. The 6,600 km MAREA subsea cable, which was jointly funded by Microsoft and Facebook, links Virginia Beach, Virginia to Bilbao, Spain. For these two Internet giants, the collaboration represents a significant change for the subsea cable industry, which previously funded major projects by establishing a consortium of telecom carriers, with bandwidth on the system subsequently divided between eight, ten or even more investing parties. MAREA is not only massive in comparison to the projects from just a decade ago, it also follows a more southerly transatlantic route, landing in Europe on the Iberian peninsula. For the American landing, the cable arrives hundreds of miles to the southern coast of Virginia, in proximity to new hyperscale data centre campuses being built by Microsoft and Facebook.

It is also very interesting that the MAREA cable will be managed by Telxius, providing Telefónica’s new infrastructure company with two very reliable customers whose bandwidth needs are certain to skyrocket in coming years. Telefónica’s European fibre backbone will certainly be an option if either of these Internet giants needs their traffic to be carried onward to other European destinations.

The MAREA cable features eight fibre pairs and an initial, estimated design capacity of 160 Tbps. TE SubCom served as the system supply partner for MAREA.

The BRUSA subsea system

BRUSA is Telxius’ new submarine cable linking Rio de Janeiro and Fortaleza (Brazil) with San Juan (Puerto Rico) and Virginia Beach (USA). The 11,000-km cable consists of 8 fibre pairs, with 135 x 100 Gbps per fibre pair, with 4 direct fiber pairs connecting US-Brazil, and 4 fibre pairs entering Puerto Rico and Fortaleza. Alcatel Submarine Networks is the lead supplier. The system is expected to begin operations in the coming weeks.

The SAM-1 subsea system

Telxius also owns and operates the 25,000 km SAM-1 subsea cable connecting the U.S. with Central and South America. It started operations in 2000, connecting the United States, Puerto Rico, Brazil, Argentina, Chile, Peru and Guatemala, and was extended to reach Ecuador and Colombia in 2007. Infinera’s equipment is deployed in this system, enabling on-demand service capabilities in some locations Infinera solutions are also installed on the terrestrial backhaul routes associated with the cable.

Expanding its tower business

The portfolio of 16,300 cellular towers, most of which were transferred from Telefónica, are largely concentrated in Spain (10,741 towers), where the company of course has its historical roots. Germany is its second largest installed base with 2,350 telecom masts, followed by Brazil (1,665 towers), Peru (849 towers), Chile (327 towers), and Argentina (304 towers).
Considering the business possibilities for in the Telxius portfolio, it is clear that the plan must be to add new mobile operator tenants in as many properties as possible. In addition to conventional mobile operators, a new class of Internet of Things operators is emerging.

One company leading in this sector is Sigfox, a French company founded in 2009 that builds wireless networks to connect low-power objects such as electricity meters, industrial sensors, etc. Sigfox uses a unique technology for extreme energy efficiency in the remote sensor and which remains compatible with Bluetooth, GPS 2G/3G/4G, and WiFi. Over the years, Sigfox has expanded its network to over 45 countries. It now claims to serve around 803 million people, with the ambition of extending the network across 60 countries and regions and reaching 1 billion people in 2018.

At Mobile World Congress, Telxius and Sigfox announced a deal to expand the Sigfox network in Germany to cover more than 80 percent of the country. This entails the deployment of Sigfox equipment on a number of the 2,350 telecom towers that Telxius acquired from Telefónica Germany in early 2016. In addition, Sigfox can utilize further selected antenna locations of Telefónica Germany for further expansion of its network. To reach complete network coverage across Germany requires only about 2,500 Sigfox base stations.

Previously, Sigfox Germany has acquired masts and roofs for its base stations directly. Working with Telxius, enables Sigfox to accelerate the rollout of its network as it will no longer have to negotiate directly with property owners.

If the Sigfox partnership in Germany is successful, Telxius certainly offers telecom masts in all of its other market for supporting a global IoT network.

Tuesday, March 13, 2018

The tower companies will let us know when 5G rollouts gets real - part 2

Preamble: Network densification is a key premise of 5G architecture. Once the rollouts of 5G networks begin in earnest, we should expect to see many more cell sites in urban areas across the world, so one way to measure the progress of 5G rollouts will be to keep an eye on the companies that actually will handle the deployment of physical equipment on the streets.
American Tower is one of those companies.

At this week’s Mobile World Congress, American Tower, in partnership with Philips Lighting, is announcing a street ¬¬¬pole designed to meet the capacity challenges in dense urban centres. This “Smart Fusion Pole” combines LED street lighting with mobile connectivity. Integrated antennas are concealed and can support services from multiple mobile operators on a single pole. American Tower says the design is 5G and IoT ready.

The City of Huntington Beach, California will be the first to deploy this street pole from American Tower, which has an agreement with the city for exclusive access to 200 smart pole locations. It is a start.

Although we do not know the financial terms of such an agreement, cities will soon realize that the cost of upgrading their street infrastructure in an aesthetically pleasing fashion, in this case street lights, can be paid for by the mobile industry. If the public accepts it, we could soon see thousands of such street poles even in a fairly small and lightly populated (by world standards) city such as Huntington Beach (population 200,000).

So let’s say a tower operator has 5,000 fibre-connected street poles equipped for 5G in Huntington Beach. How much can they charge Sprint or Verizon or AT&T for monthly rent? And what impact will this have on the OPEX budget of mobile operators? Will 5G networks be more costly to operate and maintain than 4G?

Expansion in Mexico

In November, American Tower American Tower acquired a subsidiary of KIO Networks that controls more than 50,000 concrete poles and approximately 2,100 route miles of fiber. The purchase price was $500 million in cash. Most of these facilities are located in dense urban areas. Mexico’s population tends to be more urban than rural.

KIO Networks is one of the major data center operators in Mexico. The company is owned by Maria Asuncion Aramburuzabala, whose family holds investments in many industrial sectors of the country.

Mexico’s Red Compartida project

Mexico’s Red Compartida (shared network) project is a private-public initiative that aimed to break-up the dominance of America Movil by providing a shared infrastructure that would allow competition in mobile services to grow. Red Compartida is sometimes described as the largest construction of a new public access network in the world, although it is eclipsed by the FirstNet emergency response project in the U.S. The primary contractor for the project is ALTÁN Redes.

Key facts of the initiative include:

•   Red Compartida will only offer wholesale mobile services.

•   Red Compartida will build a new national network covering at least 92.2% of the population.

•   Red Compartida's largest investor at 33% is Marapendi Holding BV, an indirect subsidiary of North Haven Infrastructure Partners II, an infrastructure fund with a value-add strategy to invest in OECD countries around the world, managed by Morgan Stanley Infrastructure; Caisse de dépôt et placement du Québec(CDPQ), one of North America's largest pension fund manager holds a 12.68% share; Mr. Miguel S. Escobedo holds a 9.35% share; and Mr. Eugenio Galdón, Chairman of Multitel, holds a 3.34% share.

•   Axtel and Megacable also hold a stake in ALTÁN Redes through a series of non-voting shares and without involvement in management; each has a participation of 4.01%.

•   In March, ALTÁN Redes selected Huawei and Nokia as turn-key technology providers for Red Compartida, a pure IP + LTE network.

•   Huawei technology will be used for central and southern Mexico (telecommunications regions 6 to 9), as well as providing the backbone.

•   Nokia's technology will be rolled out in the northern part of the country (regions 1 to 5). Nokia will also supply the network Core, which includes the Network Operation Center (NOC) and Security Operation Center (SOC).

As of 30-January-2018, ALTÁN Redes confirmed that the project was on target to begin operations with 30 percent coverage of Mexico’s population by 31-March-2018.

Presumably, American Tower’s newly acquired facilities will play some role, as that is the basis of the acquisition, but it remains to be seen how the economics of Red Compartida will ultimately play out. The key question ultimately is if smaller competitors will be able to challenge America Movil, and if AT&T Mexico will be able to participate in Red Compartida, and if so, to what extent it does so.

One interesting observation is the presence of Huawei as a principal vendor for Red Compartida – could that be an obstacle for AT&T Mexico?

Monday, March 12, 2018

The tower companies will let us know when 5G rollouts get real

Network densification is a key premise of 5G architecture. Once the rollouts of 5G networks begin in earnest, we should expect to see many more cell sites in urban areas across the world. How many more? That remains to see, but the count of small cells could be 10x of the number of macro base stations common in the 3G/4G era – or maybe even higher than that. Each one of these small cells will need a physical location, which must be purchased or leased from a property owner. It requires permits and authorization from municipal authorities. It must have electrical power, and it must have the backhaul capacity to support multigigabit services – this implies fibre connectivity, although some schemes for meshed wireless backhaul are under consideration as well.

One way to measure the progress of 5G rollouts will be to keep an eye on the companies that actually will handle the deployment of physical equipment on the streets. Once we see the independent tower operators report increased rents do the deployment of massive MIMO antennas or the provisioning of new fibre contracts, then we will know the game has really begun.

Building and managing the physical infrastructure is a headache, especially when there are tens of thousands of sites dispersed across thousands of municipal jurisdictions. For this reason, mobile operators have moved quickly to exit this tricky business. In some countries, they have formed networking infrastructure sharing alliances with their rivals. In other countries, they have spun-out their cell site infrastructure into new companies. In the U.S. market, they have sold their tower operations either to Crown Castle or American Tower.

American Tower, which operates approximately 149,000 communications sites, including approximately 40,000 towers in the United States and more than 108,000 towers internationally, prides itself on being one of the largest global real estate investment trusts (REITs). In addition to the towers (masts) mentioned above, its portfolio also includes more than 800 Distributed Antenna Systems (DAS) in malls, casinos, and other indoor/outdoor venues. The Boston-based company, which was founded in 1995, has a simple missions statement – “to be the premier wireless infrastructure provider” – even if it never becomes recognized as a household name by the billions of consumers with smartphones in their hands right now.

In the United States, there are a number of requirements for a company to be considered a REIT, including that the majority of income must be generated by rents or leases rather than services. A different tax methodology can make this form of business organization more appealing than ordinary corporate structures, but there are requirements that may restrict the company’s business activities and strategic manoeuvres. For instance, REITs typically must distribute 90 percent of taxable income to shareholders, rather than using it for other purposes. The structure typically is modeled after mutual funds.  Additionally, REITs must invest at least 75 percent of its total assets in assets, which could restrict the capital investments that the company might make in network equipment.

As of press time, we are awaiting the publication of the company’s financial results for Q4 2017, but American Tower’s recent performance has been good --18 consecutive quarters of double-digit growth in terms of its adjusted EBITDA. Revenue per site is on the rise and with the exception of occasional consolidation of its mobile network customers – which is a major and perennial threat to its business, the general trend is upward and forward. In this era of exploding demand for mobile data, it is a sure bet that mobile operators will be needing more towers and more bandwidth at those towers, than ever before.

In the United States, over the past few quarters, American Tower has experienced organic tenant billings growth at a pace of over 6 percent. The company also sees solid business growth across its diverse international footprint.  American Tower is pursuing a capital allocation strategy that includes stock buybacks, dividends, and acquisitions.

Big breakaway growth opportunities include multi-year network deployments with FirstNet, the first responders' network in the U.S., and Red Compartida, the shared network infrastructure project in Mexico, along with the coming 5G upgrade cycle.

The market is waiting to see financial guidance for how 5G will impact future revenue.
American Tower’s property revenue breaks down as follows geographically.

U.S. 55%
Latin America 18%
Asia                 18%
EMEA 10%

As of June 2017, it’s major customers include the big four U.S. mobile operators

AT&T   15%
Verizon 16%
Sprint             9%
T-Mobile         8%

Wednesday, March 7, 2018

A changing of the guard at SES

A changing of the guard is underway at SES, the global satellite communications company, with a new CEO and CFO expected to take up their posts in April. The new blood comes from O3b Networks, the visionary company whose name stood for the "other 3 billion" people on the planet who lacked access to the modern communications grid and who might be reached via a next-gen satellite constellation. The initial plan was for a network of 12 High Throughput Satellites (HTS) in Medium Earth Orbit (MEO) around 8,000 kilometres from the Earth. Spotbeams from each of the satellites are capable of delivering up to 1.6 Gbps of throughput at a low latency of less than 150 milliseconds. Commercial operations began in September 2014. Early investors in O3b included SES, Google, Liberty Global, HSBC Principal Investments, Northbridge Venture Partners, Allen & Company, Development Bank of Southern Africa, Sofina, Satya Capital and Luxempart. Over time, SES increased its equity stake in the venture until eventually acquiring the whole firm in 2016. By this point, O3b's constellation was fully operational and the firm was getting recognized as the fastest growing satellite start-up.

Steve Collar, who previously was CEO of O3b has been appointed as the next President & CEO of SES, replacing Karim Michel Sabbagh who is stepping down in order to spend time with his family and to pursue new interests. Andrew Browne, who was until recently CFO of O3b Networks and CFO of SES between 2010 and 2013, has been appointed as the next CFO of SES.

Currently, four new O3b satellites have arrived safely at the Guiana Space Centre in Kourou, French Guiana, in preparation for launch by a Soyuz vehicle in March 2018. The new satellites, which will be placed into orbit at a distance of 8,000km, will augment SES’s fleet of 12 O3b satellites.  As with the previous units, the new satellites were built by Thales Alenia Space. SES says a bigger constellation enables it to offer more capacity, enhanced coverage, increased efficiencies and greater reliability while delivering carrier-grade services including MEF Carrier Ethernet 2.0 certified services.

Later this week, SES is expected to release its 4Q 2107 financial results. Most of its revenue comes from long-term contracts in the satellite video distribution business, so major swings from quarter to quarter are unlikely. The company has made a much faster transition to full HD video, and now UHD 4K, than its competitor Eutelsat, although it faces the same long-term challenges as younger consumers turn to on-demand OTT video options rather than broadcast cable TV channels.

About SES

Société Européenne des Satellites (SES), which is based in Betzdorf, Luxembourg and is traded on the Euronext Paris exchange under the ticker symbol SESG, was founded in 1985 with the vision to use geostationary satellites for a pan-European TV broadcasting system.

SES currently has over 50 active GEO satellites, 12 active MEO satellites, and over 60 teleports worldwide that feed into its global terrestrial network. The combined footprint covers virtually the entire global population.

The SES business model is divided into four categories: Video, Fixed Data, Mobility, and Government services. All of these areas are growing, with Mobility being experiencing the strongest growth thanks to the installation of satellite broadband connections by many airlines worldwide.

Video remains key to SES and currently constitutes 68% of its overall sales. This has been the company's cash cow and is projected to remain so in the coming decade. SES says its distribution contracts typically run over 10 years. Current gen GEO satellites typically have 15-year lifespans. For the first nine months of 2017, SES Video generated revenue of EUR 1,031.5 billion, which was up by 1.1% over a year earlier, but down 3.3% on a like-for-like comparative basis. While revenue was sliding, at least usage was up – more households, more channels, more HD channels, more UHD. During the year, SES landed a multi-year renewal with Sky Deutschland covering seven transponders. It also signed a multi-year agreement with Viasat Ukraine to broadcast 40 pay-TV channels, including 13 HDTV channels.

Some highlights:

  • SES estimates that its satellites are delivering video to 325 million households, up 3% YoY.
  • In Europe, SES reaches 156 million households, up 2% YOY, including 30 million IPTV households
  • SES is carrying 7,743 broadcast TV channels globally, up 6% YOY. Of these, 2,601 are HDTV channels, representing an HD penetration rate of 34%, up 7% YOY. 
  • As of November 2017, SES was also carrying its first 24 UHD channels, including Fashion One, Sky and QVC.
  • SES' video feeds 44 million IPTV homes.
  • SES' video feeds over 120 video-on-demand systems
  • Regarding bandwidth intensive UHD, the company calculates that 30 UHD channels are bandwidth equivalent to 220 SD MPEG4 channels. 

Current profile of SES Networks (includes O3b)

SES Networks appears to be the stronger side of the house, at least financially in 2017, growing 12.7%.

The company suffered a scare in June 2017 when its AMC-9 satellite experienced an anomaly on-orbit when contact and control of the spacecraft were lost for multiple days. The AMERICOM-9 (AMC-9) satellite is a hybrid C/Ku-band design based on Alcatel's Spacebus 3000B3 platform. Its C-band payload features 24 - 36 MHz transponders, while the Ku-band payload features 24 - 36 MHz transponders. During the anomaly, SES was forced to transfer all customer traffic to other satellites. AMC-9 was launched in 2003, so its operational life was coming to an end. However, the event brought some economic pain and also reminded us of the vulnerabilities of operating complex electronic systems in the hostile environment of space.

One area to highlight is the strong performance of SES’ mobility business, especially Internet access for airlines and cruise ships.

Part of this is driven by SES-15, which was launched in May 2017 and just entered commercial service in January 2018. SES-15 is the first SES hybrid satellite. It equipped with 16 Ku-band transponders (36MHz equivalent) as well as a 10 GHz high throughput payload. The all-electric satellite operates at 129 degrees West, enabling it to serve North America, Mexico, Central America and the Caribbean. One of its big customers is Gogo, which now has over 400 aircraft equipped with its 2Ku in-flight connectivity technology. Gogo’s 2Ku uses two Ku-band antennas, one for download and the other for upload. The company claims that 2Ku offers peak speeds of 70 Mbps. More impressively, Gogo has a backlog of 1,600 more aircraft awaiting the installation of its antennas. The company also says that 2KU will be compatible with future MEO satellites. Airlines that have contracted for this service include AeroMexico, Air Canada, Air Canada Rouge, Air France, Alaska Airlines, American Airlines, British Airways, Cathay Pacific, Delta, GOL, Iberia, JAL, JTA, KLM, LATAM, United, Virgin Atlantic, and Virgin Australia.

A second SES hybrid satellite is expected to come online shortly, although this one has had a bumpy ride getting there. In January 2018, an Ariane 5 rocket launched from the Guiana Space Center (CSG) suffered an anomaly during the second stage separation process. Ground tracking stations lost telemetry contact with the rocket and its two satellite payloads shortly after the second stage separation. Hours later we learned that all was well, relatively speaking. he mission carried SES-14, the second of SES' hybrid satellite to be launched. SES-14 was equipped with C-band wide beams designed for in-flight connectivity across the Atlantic. This satellite now appears to be in proper orbit, so commercial services should commence after all testing and commissioning activities are complete. SES has other in-flight connectivity partners as well, including Panasonic and Global Eagle.

Saturday, February 24, 2018

Another idea for base stations in balloons

Just in time for Mobile World Congress 2018, Ericsson is introducing “a new tethered-aerial infrastructure product” – basically a base station in a balloon that is tethered to the ground – as a means of delivering mobile broadband in rural locations.

We like this idea! It sounds more reasonable than Google’s Project Loon (update below), which recently has been in service over Puerto Rico, or Facebook’s Aquila drone aircraft, whose current status and future prospects are murkier.

Ericsson’s tethered aerial infrastructure is also being called a “Super Tower”. Essentially, it is a multisector LTE base station placed at a height of 250m (800 feet), perhaps 6X higher than the typical towers you might see alongside a motorway. The base stations, of course, come from Ericsson’s portfolio. The balloon – or aerostat, as it is properly called) was developed by Altaeros Energies, a Boston-based startup with funding from SoftBank, Mitsubishi Heavy Industries, Suhail Bahwan Group, the U.S. Department of Agriculture, the National Science Foundation, and RNT Associates, wholly owned by Mr. Ratan N. Tata, former Chairman of Tata Sons. Altaeros founders come from M.I.T.

From the 250m height, the SuperTower base station can cover the equivalent territory of about 30 base stations. This will bring compelling economics in lightly-populated rural areas.

The helium-based aerostat can be inflated and deployed in 24 hours. The company says that the cost of helium is negligible in an overall operating budget. The floating up employs 3 tethers. If all three were to break, onboard valves would crack open and release helium, causing the balloon to slowly descend. The company says its balloons should be able to survive windstorms of up to 160 kph (100 mph).

It should be noted that Altaeros has also been working on a different version of its helium-filled aerostat that floats a wind turbine at the 250m height. Wind speeds at this elevation are typically stronger and more consistent than at ground level. Late last year, Altaeros deployed one of its SuperTowers in rural Maine, but there has been little reporting about the outcome of this pilot run.

Navigational hazard?

Of course, an LTE base station floating at a 250m height does present a navigational hazard to low-flying aircraft. In the U.S., the aerostats need FAA permission, but just like for skyscrapers, much permits are obtainable if due safety considerations are met.

In addition to providing rural broadband, the SuperTowers could also be used for temporarily delivering Internet access following natural disasters or for special events, such ay s outdoor concerts or festivals. The military reportedly is interested in using the aerostats rapidly deployable communication networks. Similar aerostats are already used for border protection and other surveillance applications.  Ericsson plans to offer the SuperTowers to mobile operators in late 2018.

Where are the Loons?

In contrast to Altaeros’s tethered SuperTowers, Google’s Loon balloons can be literally anywhere the wind takes them. The tetherless loons typically fly in the stratosphere at heights of about 18,000m (60,000) – above commercial air traffic.  X (formerly Google X), which is the subsidiary that Alphabet has established to turn this science project into a commercial business, attempts to steer the balloons by adjusting the altitude to catch prevailing winds. This is a difficult proposition for network planning even in good weather because to provide service to users on the ground it is essential that the floating base station not drift too far away.

The Altaeros SuperTower balloons are directly wired into the backhaul network in a predictable structure. The backhaul from a Loon balloon can travel to other Loon balloons first before eventually finding a downlink to a ground station. Like the Altaeros configuration, the Loons carry a lightweight LTE eNodeB enabling them to connect directly to the handsets of users below.

Google initially launched pilot trials of Loon balloons in the southern hemisphere in 2013 from launch sites in New Zealand, Chile, and Argentina.

A real-world use case for the Loon balloons emerged after Hurricane Maria struck the Caribbean in late September 2017 and devastated the communications infrastructure in Puerto Rico and the U.S. Virgin Islands. Two weeks following the disaster, the FCC granted a provisional license for X to launch its balloons in Puerto Rico in hopes of alleviating the dire situation – nearly 90 percent of the islands base stations remained out of service two weeks after the storm.

So how is Loon performing in Puerto Rico? At this point, we only have anecdotal evidence. X has only disclosed limited data on the program. We know, for example, that in the first week, X was launching the loon balloons from Nevada and guiding them all the way to Puerto Rico using its algorithms and machine learning. On November 9, a tweet from the team at X said the balloons were supplying basic Internet connectivity to 100,000 AT&T and T-Mobile users in Puerto Rico – good progress! We also learned that Nokia was supplying base stations for Project Loon. SES Networks and Liberty Cablevision provided ground station connectivity in Puerto Rico.

What is also interesting to see are user reports and balloon sightings shared on social media, especially when the balloons have drifted west to the neighboring island of Hispaniola.

The website even provides tracking maps, from which we surmise that connectivity at any one spot will vary with the wind.  A series of tether balloons could serve the same purpose and maybe at lower cost and greater reliability.

Wednesday, February 21, 2018

Profile of the Telecoms market in Kenya - part 3

Preamble: After nearly a week of blocking broadcasts of Kenya's leading television stations, Kenyan authorities have relented and allowed most of the broadcasters to resume operations. In this series of articles, we profile the vibrant telecommunications market in Kenya.

See part 1part 2part 3part 4part 5part 6\

SECTION 3——Regulatory activities of Communications Authority of Kenya

In mid-August 2015, Kenya’s Business Daily reported that following a dispute which had already gone on for three years Airtel Kenya had gone to the Kenyan High  Court seeking to block another demand from the CA that it should pay $20.025 million spectrum fees before it could renew its operating licence. The dispute arose from the fact that in 2014 Essar’s YuMobile subsidiary in Kenya left the local mobile market and transferred its subscribers and licence to Airtel Kenya for about $6.9 million. According to Airtel, the CAK at that time promised to merge its operating licence with that of yuMobile, with the deal granting Airtel a lease to operate in the country until January 2025. However, subsequently under pressure from the Kenyan Treasury CAK had reneged on that agreement and come back to Airtel to demand KSh2 billion in licence fees.

The Safaricom row

In late September 2017, following accusations by Raila Odinga that Safaricom had deliberately delayed results of the August 8th presidential election, the CAK national communication regulator denied that there had been any such failures of transmission by any of the Kenyan mobile operators.

In late November 2017, DG Francis Wangusi of the CAK announced that following the failure of an earlier agreement between Kenyan operators to establish a common money transfer infrastructure by July 2017 the regulator now planned to make the interoperable wallet system mandatory in December 2017.

In mid-June 2017, Reuters reported that the Kenyan High Court had overturned a legal amendment that would have denied the right of the CA to manage competition in the sector

At the end of November 2017, Kenya’s Standard newspaper reported that Francis Wangusi  the DG of the CAK regulator was questioning the right of Safaricom to make deliveries of goods sold via its Masoko online e-commerce platform on the grounds that the operator’s existing license did not authorise such activities

On December 19th, 2017, Kenya’s Business Daily newspaper reported that the CA had announced that it was soliciting views from the general public on the minimum standards that should be set for internet devices.

As noted elsewhere the CA decided in January 2018 not to breakup Safaricom into separate telecommunications and financial businesses

In mid-January 2018 it was announced that CAK DG Francis Wangusi had been sent on leave for three months with effect from January 12th while an independent audit is being carried out on the Authority’s “structure, promotions and training” following alleged claims of malpractice in those functions.

This surprising and unexplained move has been interpreted by observers in many different ways. On January 23rd 2015 Wangusi was reappointed as CA DG for a second four year term but since then it seems that his relationships with the Kenyan Ministry of Communisations and the Kenyan government have deteriorated for several reasons. One reason may have been the fact that the CA had been unable to extract an additional licence fee of KSh 280 million from Airtel Kenya which was cash the Kenyan Treasury badly needed; secondly it was argued that the CA had awarded a license to Jamii Telecom to use some 700 MHz frequency far too cheaply and at a more questionable level had exercised extremely tight control on his resources refusing to fund some possibly boondoggle trips to symposia for some members of the Ministry.

However, it was announced that following a legal challenge by two students from the University of Nairobi who argued that the action of the CA Board in sending Wanqusi on involuntary leave was unconstitutional Kenya’s Labour Court had ordered a halt to that decision pending a review of the case originally scheduled for January 29th

CAK licences register lists by name
1. International Gateway Operators.=12
2. Submarine Cable Landing Rights Operators=3
3. Network Facilities Providers Tier 1=3
4. Network Facilities Providers Tier 2=23
5. Network Facilities Providers Tier 3=29
6. Application Service Providers=214
7. Content Service Providers=334
8。Dot KE SubDomain Name Registrar Service Providers=62
9. Business Process Outsourcing Services Providers=26
10. Telecommunication Contractors=533
11. Telecommunication Technical Personnel=509
12. Telecommunications Equipment Providers=526
13. Public Communication Access Centres=14

CAK Telecommunications Statistical Report for Q1 2017/18 ended September 30th 2017
The full 32-page report is available to the general public so the following is merely a summary of its main findings

Q4 2016/17
Q1 2017/18
Change %
Mobile subscriptions (mn).                    
Fixed subscriptions
Mobile penetration
Fixed telephone penetration
Number of mobile money subs: (mn).  
Number of MM transactions (mn)
Value of MM transactions (KSh bn)
Data internet subscriptions (mn).          
Of which mobile data (mn)
Of which terrestrial wireless
Of which fixed DSL
Of which fixed fibre optic
Of which fixed cable modem
Of which satellite + other

Available international capacity (Gbit/s)
Used international bandwidth (Gbit/s)
Internet penetration
Broadband penetration
Number of free-to-air TV channels
Number of radio FM stations

to be continued

See also