Facebook Likes $1K Base Stations

发布时间:2017-11-10 00:00
作者:Ameya360
来源:Rick Merritt
阅读量:1288

  Facebook showed at a gathering of the Telecom Infra Project here its design for an open source cellular base station that could cost less than $1,000. For TIP, the question for the coming year is whether such systems make it into broad deployment.

  To gain an edge on larger rivals such as Amazon and Google, Facebook launched the group of carriers and partners in early 2016to jointly build software-defined networks. SDN is a broad industry move that aims to shave costs by moving functions to high-level software from relatively expensive systems using proprietary ASICs and protocols.

  So far, TIP has laid out a compelling business case, staffed a variety of work groups and produced a few interesting prototypes.

  If all goes well, a handful of carriers will start field trials of the open-source GSM base station before April and an LTE version before June. A module announced here will connect the systems to battery, solar or utility power sources.

  A parallel effort spawned a low-cost, open-source core network switch made by Taiwan’s Accton that carriers will test in Africa and Latin America soon. Cumulus Networks said it will port its Linux-based networking software to the switch.

  At the event, TIP kicked off a new working group to explore open source versions of a radio-access network based on x86 servers a separate effort to apply machine learning to network planning. It already runs a handful of projects on topics such as edge computing and millimeter wave designs for last-mile access, leveraging Facebook’s 60 GHz Terragraph system.

  The focus on gaining new users while lowering costs is critical for carriers.

  “We continue to see exponential traffic growth with no end, our production costs are increasing and the ARPU for traditional triple- or quad-play services is flat or declining,” said Axel Clauberg, TIP’s chair and vice president of technology innovation for Deutsche Telekom AG, calling TIP “the most important transformation project in the industry.”

  But change comes hard. “Many people are in monitoring mode, we need people to move into contributing,” he said, calling for carriers to get more engaged.

  Executives from BT, Telefonica and Vodaphone joined TIP’s board recently and China Unicom became a TIP member. AT&T seems to be playing a lower profile role than in the past, while Verizon and China’s top two carriers were not active at the event.

  “It is still very early, but we want to get to large deployments,” said Jay Parikh, head of infrastructure and engineering at Facebook who helped launch the group.

  To that end, he described the group’s work helping lay a shared fiber optic cable to an area serving three million people in Uganda. A separate effort hopes to deploy a cellular network in a remote area of Peru where 8 million people live.

  In addition, TIP encouraged carriers to launch accelerators in Asia and Europe that have raised more than $200 million and funded ten startups to date.

  One of the startups, Amarisoft of France, claims it has put into a production network software it developed to run LTE baseband functions on a quad-core 2.4 GHz x86 server. The code enables round-trip latency of 1.5 milliseconds for an IQ sample over a 2x2 MIMO FDD network running in a 20 MHz channel, said chief executive Franck Spinelli.

(备注:文章来源于网络,信息仅供参考,不代表本网站观点,如有侵权请联系删除!)

在线留言询价

相关阅读
Facebook Builds Chip Team, ASIC
A Facebook executive confirmed reports that the social networking giant is hiring chip engineers and designing at least one ASIC. The news came at the @Scale event here, where Facebook announced that five chip companies will support Glow, an open-source, deep-learning compiler that it backs.Facebook “is absolutely bringing up a silicon team focused on working with silicon providers, and we have a chip we’re building, but it’s not our primary focus,” said Jason Taylor, vice president of infrastructure at Facebook. The chip is “not the equivalent of [Google’s] TPU” deep-learning accelerator, he added, declining to provide further details on its focus or time frame.Working with the estimated 50 companies designing AI acceleratorsis one focus for the new Facebook chip group. “There will be a lot of [accelerator] chips in the market,” said Taylor at a press roundtable. “The big question is whether the workloads they are designed for are the important ones at the time.”In a keynote, Taylor described Glow as a generic compiler to let developers target any of the emerging deep-learning accelerators for inference in the cloud or at the edge of the network. It does not target client systems such as smartphones.“We expect that there will be hardware fragmentation [in inference accelerators]. Our work with Glow is to help machine-learning experts design neural nets and not have to do the work required to tune them” to each unique chip.“We know that the fragmentation is coming because no one knows what combination of [hardware] resources [such as on-chip memory blocks and multiply-accumulate arrays] will win, so we’ll let developers focus on the high-level graphs without hand-coding for the specifics of hardware.”Jason Taylor described Glow as a compiler for inference on cloud and edge networks. (Images: Facebook)Glow takes an AI graph produced by a framework such as TensorFlow or Caffe2 and renders it into byte code for hardware accelerators, explained Taylor. The compiler includes several tools including an instruction scheduler, a linear algebra optimizer, a memory allocator to generate efficient code for a chip’s specific memory configuration, and a CPU-based reference implementation for testing the accuracy of the hardware, according to a Facebook blog.Cadence, Esperanto Technologies, Intel, Marvell, and Qualcomm said that they will support Glow on future chips. Taylor said that he expects to add others to the list. “That’s one of the benefits of it being open-source.”One senior chip expert described Glow as a framework for deploying a neural network in production systems. Its input would be a graph created in a framework such as TensorFlow or Caffe2.Some established chipmakers already supply similar software. For example, Nvidia’s Tensor RT takes in a graph from a framework and outputs Cuda code for its GPUs.Traditionally, compilers are tightly optimized for a specific chip. But “what a compiler is these days is quite a bit broader than in the past — the kinds of optimizations in Glow have to do with identifying large portions of a graph that can be rendered to a hardware accelerator,” said Taylor.Glow is the latest example of an effort to plug the gap between software and hardware in the fast-moving world of deep learning. For example, Nvidia’s Tensor RT is now in its fifth version, though it was first released just a year ago. Some accelerator startups express frustration at the level of work needed to support the wide variety of software frameworks and their changes.Facebook, Microsoft, and others are backing ONNX, a standard way to express a graph with its weights. In December, the Khronos Group released NNEF, a hardware abstraction layer for deep-learning accelerators.For its part, Glow is a single component of Pytorch 1.0, a collection of open-source projects that includes merged Caffe2 and Pytorch frameworks. The first developer conference for Pytorch 1.0 is slated for October in San Francisco.In a separate talk, Facebook engineering manager Kim Hazelwood rattled off a list of a dozen different deep-learning workloads that the social network uses, employing at least four different kinds of neural nets. Every day, the AI apps generate more than 200 trillion inferences, translate more than five billion texts, and automatically remove more than a million fake accounts.Some of Facebook’s inference tasks require 100 times more compute than others, she said. Today, Facebook runs the jobs on a handful of CPU and GPU servers that it has designed.Moving from general-purpose to custom hardware would require tailoring chips specific to those still-changing workloads, Hazelwood told EE Times after her talk. She declined to give any insights into Facebook’s thoughts on using any custom AI accelerators.Facebook alone uses at least five kinds of neural networks across at least a dozen deep-learning apps. Click to enlarge.One observer speculated that Glow would be an ideal tool to enable the company to adopt a handful of accelerators suited to its various workloads. Its semiconductor team could help cull through the options to select a few chips and perhaps suggest customizations for some of them.Separately, Facebook posted a blog describing a new software tool that it created that uses deep learning to debug code. SapFix can automatically generate fixes for specific bugs and then propose them to engineers for approval and deployment to production, it said.So far, Facebook has used SapFix to accelerate the process of shipping code updates to millions of devices using the Facebook Android app. Facebook said it will release a version of the tool but did not state when.
2018-09-17 00:00 阅读量:2072
China said to quickly pull approval for new Facebook venture
For Facebook, success in China was brief. Very brief.For several hours, a Chinese government database showed that Facebook had gained approval to open a subsidiary in the eastern province of Zhejiang. Facebook said it would use the company to set up an innovation hub there.Then the registration disappeared and references to the subsidiary were partially censored in Chinese media.Read more from The New York Times:Facebook gains status in China, at least for a momentChinese billionaire is named as party host in Australian sexual assault caseWild about tech, China even loves robot waiters that can’t serveNow the approval has been withdrawn, according to a person familiar with the matter who declined to be named because they were not authorized to speak on the record.While the about-face does not definitively end Facebook’s chances of establishing the company, it makes success very unlikely, the person said. The decision to take down the approval, the person added, came after a disagreement between officials in Zhejiang and the national internet regulator, the Cyberspace Administration of China, which was angry that it had not been consulted more closely.A Facebook spokeswoman declined to respond to a request for comment. The Cyberspace Administration of China, the Zhejiang provincial government and the municipal government in Hangzhou, Zhejiang’s provincial capital, did not respond to requests for comment.The strange incident underscores how much of a challenge it has become for the globe-spanning social network to get into China — even just to open an innovation center.Its main platform has been blocked in China for almost a decade. Its other services, such as Instagram, were once popular in China, but have since also fallen behind the government’s filters. While Facebook had hoped to dip a toe in the market and work with Chinese developers, its very presence appears to have become a large, and incendiary, political question here.Google, which has also long had its services blocked in China, has recently made progress there, opening an artificial-intelligence center and releasing a game in the market. Last month Google said it would invest $550 million in the Chinese e-commerce company JD.com. Other American tech companies, like LinkedIn, have maintained a presence by censoring content at Beijing’s request.The kerfuffle also illustrates how complicated China’s bureaucracy can be. Foreign companies seeking to expand here must navigate a vast and decentralized government in which provinces, cities and ministries all vie for influence and power. While one part of the government may be happy to support a foreign company like Facebook, that could ruffle feathers elsewhere. Facebook seems to have fallen victim to such a scenario.The government turnaround is only the latest incident in the long, strange history of relations between the world’s largest social network and its largest internet market.Over the years, Facebook’s founder, Mark Zuckerberg, has tried myriad ways to court China. He asked Chinese president Xi Jinping to suggest a Chinese name for his child during a state visit, held a Mandarin-language question-and-answer session with students that was widely shared, and even posted photos from a smog-filled jog through Tiananmen Square.Facebook has experimented with a censorship tool and ran a photo-sharing application in China, called Colorful Balloons, without putting its name to the product. The product was eventually taken down and also did not seem to have any official approval from the Cyberspace Administration.For Facebook, angering officials at China’s internet regulator could make progress particularly difficult, especially as new personnel have just taken over managing the relationship with the Chinese government.At the end of last year, Facebook’s former head of government relations, Wang-Li Moser, left the company. Ms. Moser, who had been brought in from Intel to help Facebook court China’s leadership, had become a sign of the company’s ambitions, though she struggled to make headway.Stepping in to replace her has been Zhang Jingmei, who created the company that ran the anonymous Colorful Balloons app, and William Shuai, who was once a low-level Chinese official. Both were listed on the board of the newly created, and quickly disappeared, subsidiary in Zhejiang.
2018-07-27 00:00 阅读量:2158
Facebook confirms data sharing with Huawei, 3 other Chinese firms
Facebook has admitted sharing users' data with Chinese company Huawei -- facing the heat in the US over data privacy concerns -- along with three other China-based smartphone makers Lenovo, OPPO and TCL.According to a report in Financial Times on Tuesday, Francisco Varela, vice president of Mobile Partnerships at Facebook, said that many technology companies have worked with Huawei."Facebook's integrations with Huawei, Lenovo, OPPO and TCL were controlled from the get-go -- and we approved the Facebook experiences these companies built," Varela said in a statement."Given the interest from Congress, we wanted to make clear that all the information from these integrations with Huawei was stored on the device, not on Huawei's servers," the Facebook executive added.The Facebook confirmation came a day after The New York Times reported that the social media giant provided access to users' data to at least 60 different device makers -- including Apple, Amazon, Microsoft, Samsung and BlackBerry.There are concerns in the US among the FBI, CIA, NSA, the Federal Communications Commission and House Intelligence Committee over Huawei devices.In February, FBI Director Chris Wray said the FBI was "deeply concerned" about the risks posed by Huawei and ZTE. The US military has banned phones made by Huawei and ZTE.According to Senator Mark Warner, the Vice Chairman of the Senate Select Committee on Intelligence, the concerns about Huawei were not new."The news that Facebook provided privileged access to Facebook's API [application programming interface] to Chinese device makers like Huawei and TCL raises legitimate concerns, and I look forward to learning more about how Facebook ensured that information about their users was not sent to Chinese servers," he said in a statement late Tuesday."Concerns about Huawei aren't new - they were widely publicised beginning in 2012, when the House Permanent Select Committee on Intelligence released a well-read report on the close relationships between the Chinese Communist Party and equipment makers like Huawei," Warner added.The Senate Commerce Committee has already sent a letter to Facebook over the reports that it shared data with device makers.Earlier, the social network defended the pacts with device makers, saying that these partnerships do not raise privacy concerns.Facebook said that the partners signed agreements that prevented people's information from being used for any other purpose than to recreate Facebook-like experiences."Partners could not integrate the user's Facebook features with their devices without the user's permission," Ime Archibong, Facebook's Vice President of Product Partnerships, said in a statement.Facebook launched the device-integrated APIs about a decade ago and said that all these partnerships were built on a common interest -- the desire for people to be able to use Facebook whatever their device or operating system."Given that these APIs enabled other companies to recreate the Facebook experience, we controlled them tightly from the get-go," Archibong said.Facebook said that it had already ended 22 of the device partnerships."Now that iOS and Android are so popular, fewer people rely on these APIs to create bespoke Facebook experiences. It's why we announced in April that we're winding down access to them. We've already ended 22 of these partnerships," Archibong noted.
2018-06-07 00:00 阅读量:2159
  • 一周热料
  • 紧缺物料秒杀
型号 品牌 询价
CDZVT2R20B ROHM Semiconductor
TL431ACLPR Texas Instruments
BD71847AMWV-E2 ROHM Semiconductor
MC33074DR2G onsemi
RB751G-40T2R ROHM Semiconductor
型号 品牌 抢购
STM32F429IGT6 STMicroelectronics
BU33JA2MNVX-CTL ROHM Semiconductor
IPZ40N04S5L4R8ATMA1 Infineon Technologies
BP3621 ROHM Semiconductor
TPS63050YFFR Texas Instruments
ESR03EZPJ151 ROHM Semiconductor
热门标签
ROHM
Aavid
Averlogic
开发板
SUSUMU
NXP
PCB
传感器
半导体
相关百科
关于我们
AMEYA360微信服务号 AMEYA360微信服务号
AMEYA360商城(www.ameya360.com)上线于2011年,现 有超过3500家优质供应商,收录600万种产品型号数据,100 多万种元器件库存可供选购,产品覆盖MCU+存储器+电源芯 片+IGBT+MOS管+运放+射频蓝牙+传感器+电阻电容电感+ 连接器等多个领域,平台主营业务涵盖电子元器件现货销售、 BOM配单及提供产品配套资料等,为广大客户提供一站式购 销服务。