Just a decade ago, power usage at data centers was growing at an unsustainable rate, soaring 24% from 2005 to 2010. But a shift to virtualization, cloud computing and improved data center management is reducing energy demand.
According to a new study, data center energy use is expected to increase just 4% from 2014 to 2020, despite growing demand for computing resources.
Total data center electricity usage in the U.S., which includes powering servers, storage, networking and the infrastructure to support it, was at 70 billion kWh (kilowatt hours) in 2014, representing 1.8% of total U.S. electricity consumption.
Based on current trends, data centers are expected to consume approximately 73 billion kWh in 2020, becoming nearly flat over the next four years. “Growth in data center energy consumption has slowed drastically since the previous decade,” according to a study by the U.S. Department of Energy’s Lawrence Berkeley National Laboratory. “However, demand for computations and the amount of productivity performed by data centers continues to rise at substantial rates.”
Improved efficiency is most evident in the growth rate of physical servers.
From 2000 to 2005, server shipments increased 15% each year, resulting in a near doubling of servers in data centers. From 2005 to 2010, the annual shipment increases fell to 5%, but some of this decline was due to the recession. Nonetheless, this server growth rate is now at 3%, a pace that is expected to continue through 2020.
The reduced server growth rate is a result of the increase in server efficiency, better utilization thanks to virtualization, and a shift to cloud computing. This includes concentration of workloads in so-called “hyperscale” data centers, defined as 400,000 square feet in size and above.
Energy use by data centers may also decline if more work is shifted to hyperscale centers, and best practices continue to win adoption.
A San Francisco law slated to take effect next month requires companies like Airbnb to verify that rentals have a valid registration number issued by the city. The ordinance would impose on the company fines of up to $1,000 per day for each offense.
Airbnb’s lawsuit claims that the ordinance violates federal communications laws and asks a judge to block it. The law cannot fix San Francisco’s housing crunch, the company said in a blog post.
“This legislation ignores the reality that the system is not working and this new approach will harm thousands of everyday San Francisco residents who depend on Airbnb,” the company said.
Matt Dorsey, a spokesman for the San Francisco city attorney’s office, said nothing in the ordinance punishes Airbnb for their hosts’ content. Rather, the ordinance is intended to facilitate tax collection, he said.
“In fact, it’s not regulating user content at all – it’s regulating the business activity of the hosting platform itself,” Dorsey said in an email.
The case in U.S. District Court, Northern District of California is Airbnb Inc vs City and County of San Francisco, 16-03615.
Intel has run out of ideas about what it is going to do with it its security business and is apparently planning to flog it off.
Five years ago Intel bought McAfee for $7.7bn acquisition. Two years ago it re-branded it as Intel Security. There was talk about chip based security and how important this would be as the world moved to the Internet of Things.
Now the company has discussed the future of Intel Security with bankers, including potentially the outfit. The semiconductor company has been shifting its focus to higher-growth areas, such as chips for data center machines and Internet-connected devices, as the personal-computer market has declined.
The security sector has seen a lot of interest from private equity buyers. Symantec said earlier this month it was acquiring Web security provider Blue Coat for $4.65 billion in cash, in a deal that will see Silver Lake, an investor in Symantec, enhancing its investment in the merged company, and Bain Capital, majority shareholder in Blue Coat, reinvesting $750 million in the business through convertible notes.
However Intel’s move into the Internet of Things does make it difficult for it to exit the security business completely. In fact some analysts think it will only sell of part of the business and keep some key bits for itself.
Japanese messaging app firm Line Corp has held off on setting a tentative price range for its initial public offering (IPO) by one day, until Tuesday, the company said in a regulatory filing, citing the “market environment”.
The IPO price range was originally scheduled to be announced on Monday. Line still plans to list in New York on July 14 and in Tokyo the following day, the filing showed.
On Friday, the S&P 500 fell 3.6 percent, its biggest one-day drop in 10 months, and Japan’s broad Topix index slid 7 percent after Britain voted to exit the European Union.
The equity market in Japan recovered somewhat on Monday as the Topix closed up 1.8 percent, but the delay will allow the company to assess the market in New York and London on Monday before setting the tentative price range, a Line spokesman told Reuters.
Earlier this month, the company announced plans to sell 35 million new shares in an IPO, which would raise 98 billion yen ($963 million) at its initial reference price of 2,800 yen per share.
Line’s listing will go ahead according to its planned schedule, the company said on Friday.
Companies around the world are wrestling with the aftermath of the Brexit vote, which is likely to delay or disrupt upcoming takeovers and initial public offerings. Companies with direct exposure to the British economy are more likely to see their deals scuppered compared with those who are just caught up in global market volatility.
Line has little direct exposure to Britain or Europe. Its main markets are Japan, Indonesia, Taiwan and Thailand.
Line delayed its IPO by two years, buying time to fix weaknesses in weak financial reporting controls, bolster staffing and develop its business plan. But in doing so, it left billions of dollars on the table as its valuation shriveled.
The revamped EU-U.S. Privacy Shield was sent to EU member states overnight, according to a report from Reuters. Privacy Shield would govern how multinational companies handle the private data of EU residents.
Member states are expected to vote on the proposal in July, unnamed sources told Reuters. Representatives of the EU and the U.S. Department of Commerce didn’t immediately respond to requests for comments on the reported deal.
Critics of Privacy Shield, including European privacy regulators, have said the deal is too complex and fails to reflect key privacy principles.
The new language sent to member states includes stricter data-handling rules for companies holding Europeans’ information, Reuters reported. The new proposal also has the U.S. government explaining the conditions when they would collect data in bulk, according to the report.
Negotiators on both sides of the Atlantic have been rushing to craft a new trans-Atlantic data transfer agreement since the Court of Justice of the European Union struck down Safe Harbor, the previous transfer pact, last October.
The court ruled that Safe Harbor didn’t adequately protect European citizens’ personal information from massive and indiscriminate surveillance by U.S. authorities. Safe Harbor had been in place since 2000.
The Intel Security business came largely from the company’s acquisition for $7.7 billion of security software company McAfee. Intel announced plans to bake some of the security technology into its chips to ensure higher security for its customers.
With the surge in cyberthreats, providing protection to the variety of Internet-connected devices — such as PCs, mobile devices, medical gear and cars — requires a fundamentally new approach involving software, hardware and services, the company said in February 2011, when announcing the completion of the McAfee acquisition.
Intel has been talking to bankers about the future of its cybersecurity business for a deal that would be one of the largest in the sector, reported The Financial Times, citing people close to the discussions. It said a group of private equity firms may join together to buy the security business if it is sold at the same price or higher than what Intel paid for it.
“I could see them selling a piece of the service, but not all security capabilities,” said Patrick Moorhead, president and principal analyst at Moor Insights & Strategy.
“Intel has a decent security play right now and security is paramount to the future of IoT,” Moorhead said. “Hardware-based security is vital to the future of computing.”
Intel is declining to comment on the report, a company spokeswoman wrote in an email.
The US is clearly embarrassed the Chinese Sunway TiahuLight system is leading the supercomputer arms race. Now the Department of Energy’s (DOE) Oak Ridge National Laboratory has announced that is having a new IBM system, named Summit, delivered in early 2018 that will now be capable of 200 peak petaflops.
That would make it almost twice as fast as TaihuLight. The Summit will be based around IBM Power9 and Nvidia Volta GPUs. Summit use only about 3,400 nodes. Each node will have “over half a terabyte” of coherent memory (HBM + DDR4), plus 800GB of non-volatile RAM that serves as a burst buffer or extended memory.
IBM are not the only ones worried about the Chinese getting ahead on speed. Cray announced this week its Cray XC systems are now available with the latest Intel Xeon Phi (Knights Landing) processors.
The company said the new XC systems, which feature an adaptive design that supports multiple processor and storage technologies in the same architecture, deliver a 100 per cent performance boost over prior generations. Cray also unveiled the Sonexion 3000 Lustre storage system, which can deliver speeds of almost 100GB/sec in a single rack. These should be rather good at number crunching too.
Basically this means that the hardware can be used by the OPNFV collaborative open source community to accelerate the delivery of cloud-enabled networks and applications.
Nokia said the OPNFV Lab will be a testbed for NFV developers and accelerates the introduction of commercial open source NFV products and services. Developers can test carrier-grade NFV applications for performance and availability.
Nokia is making its AirFrame Data Center Solution available as a public OPNFV Lab with the support of Intel, which is providing Intel Xeon processors and solid state drives to give communications service providers the advantage of testing OPNFV projects on the latest and greatest server and storage technologies.
The Nokia AirFrame Data Center Solution is 5G-ready and Nokia said it was the first to combine the benefits of cloud computing technologies to meet the stringent requirements of the telco world. It’s capable of delivering ultra-low latency and supporting the kinds of massive data processing requirements that will be required in 5G.
Morgan Richomme, NFV network architect for Innovative Services at Orange Labs, OPNFV Functest PTL, in a release. “NFV interoperability testing is challenging, so the more labs we have, the better it will be collectively for the industry.”
AT&T has officially added Nokia to its list of 5G lab partners working to define 5G features and capabilities. It’s also working with Intel and Ericsson.
Apple announced that is will discontinue its Thunderbolt Display, the high-resolution external display that users of the MacBook and other Macs could use to get a better picture and work with more apps.
The company said Thursday that the 27-inch widescreen display with LED backlight technology will be available on Apple’s online store, in Apple retail stores and from authorized resellers while supplies last.
The Thunderbolt Display currently retails on the Apple online store at $999. It has a 2560 x 1440 resolution.
It isn’t clear whether Apple plans to follow with newer versions that use 5K resolution displays at 5120 by 2880 pixels, which is the display technology Apple uses on its high-end iMac. There was speculation earlier that a new version would be announced at the company’s Worldwide Developers Conference this month.
An Apple spokeswoman declined to comment on whether Apple planned to offer a refresh to the display.
Apple said in an emailed statement that “there are a number of great third-party options available for Mac users.”
Safari 10 was introduced earlier this month as part of macOS Sierra, this year’s operating system upgrade.
Apple typically supports its newest browser on three editions of macOS: The latest version and its two predecessors. The now-current Safari 9, for example, receives updates, including security patches, on last year’s El Capitan, 2014′s Yosemite and 2013′s Mavericks.
Safari 10 will be supported on Sierra, El Capitan and Yosemite. Meanwhile, Mavericks will remain on Safari 9.
The Safari 10 preview is currently available only to registered Apple developers, who pay $99 annually for access to early builds, development tools and documentation.
The general public will get its first look at Safari 10 next month after Apple opens up its broader-based public beta program for Sierra. Those who have signed on to the beta preview will also be able to download preliminary versions of Safari 10 for El Capitan and Yosemite, running the preview browser but sticking with their older, more stable operating systems.
Some of Safari 10′s signature features will be available only within macOS Sierra, including web-based Apple Pay — where payment is authorized with an iPhone or Apple Watch — but others will be supported by older versions of the operating system. Among the most notable are the new ability for developers to distribute and sell Safari add-ons in the Mac App Store, and easy portability of iOS content blockers to macOS.
If Apple replicates last year’s beta schedule, it will release the first public preview of macOS Sierra and Safari 10 around July 14.
Robots that work as assistants in unison with people are set to upend the world of industrial robotics by putting automation within reach of many small and medium-sized companies for the first time, according to industry experts.
Collaborative robots, or “cobots”, tend to be inexpensive, easy to use and safe to be around. They can easily be adapted to new tasks, making them well-suited to small-batch manufacturing and ever-shortening product cycles.
Cobots can typically lift loads of up 10 kilograms (22 lb) and can be small enough to put on top of a workbench. They can help with repetitive tasks like picking and placing, packaging or gluing and welding.
Some can repeat a task after being guided once through the process by a worker and recording it. The price of a cobot can be as little as $10,000, although typically they cost two to three times that.
The global cobot market is set to grow from $116 million last year to $11.5 billion by 2025, capital goods analysts at Barclays estimate. That would be roughly equal to the size of the entire industrial robotics market today.
“By 2020 it will be a game-changer,” said Stefan Lampa, head of robotics of Germany’s Kuka, during a panel discussion organized by the International Federation of Robotics (IFR) at the Automatica trade fair in Munich.
Growth in industrial robot unit sales slowed to 12 percent last year from 29 percent in 2014, the IFR said on Wednesday, weighed by a sharp fall in top buyer China.
The world’s top industrial robot makers – Japan’s Fanuc and Yaskawa, Swiss ABB and Kuka – all have collaborative robots on the market, although sales are not yet significant for them.
But the market leader and pioneer is Denmark’s Universal Robots, a start-up that sold its first cobot in 2009 and was acquired by U.S. automatic test equipment maker Teradyne for $285 million last year.
A few months back Nick wrote about AMD Zen processor found in a Linux Kernel Mailing List confirming that Zeppelin had support for eight bundles of four cores on a single chip, or 32 physical processing cores.
This tied in with a story written in August of 2015 about a MCM Multi Chip module that featured a Zeppelin core, a super-fast 100GB/s interconnection via 4 GMI links and Greenland (Vega) high performance GPU with 4+ TFlops of performance. This APU will still happen, it will just be a bit later – the end of 2017.
Now we have a few more details about Zeppelin cluster and this is proving to be another “Fudzilla told you so” moment. Apparently you can put up to four Zeppelin CPU clusters on a one chip and make a 32 core chip. This will be connected via coherent interconnect (coherent data fabric).
Each Zeppelin module has eight Zen cores and each Zen core has 512 KB of L2 cache. Four Zen cores share 8MB or L3 cache making the total amount of L3 cache per Zeppelin cluster 16 MB.
Each Zeppelin cluster will have PCIe Gen 3, SATA 3, and a 10GbE network connection. A server version of the chip has the server controller hub, DDR4 memory controller and AMD secure processors.
AMD will have at least three pin compatible versions of the next generation Opteron using Zeppelin cluster of Zen cores. There will be a 8 core versions with single Zeppelin cluster, dual Zeppelin cluster version and a quad Zeppelin version, that one that we have called Naples which will have 64MB L3 cache. All this sounds rather a lot.
We are expecting to see Zen-based Opterons in eight, sixteen and thirty two core versions for servers in 2017.
This means it can be used for artificial intelligence (AI) and deep brain learning. Intel claims the bootable Xeon Phi scales to any number of processors using Intel’s scalable-system framework (SSF) with integrated on-chip fabric, Omni-Path fiber and silicon photonics.
Speed was bumped to 1.5-GHz for high-performance computing (HPC) orchestration while maintaining its low-power energy budget compared with GPU-based and other competing high-speed multicore arrays tied to PCIe.
Intel is flogging the Xeon Phi is the world’s first CPU with integrated fabric in-package. Because it is now bootable by itself, has integrated in-package memory and can run on systems as small as developer workstations starting at $5k.
Intel’s bootable Xeon Phi fits into its scalable system framework (SSF) which connects with omni-path fiber driven by silicon photonics, to perform high-performance computing (HPC) tasks, especially Deep Learning algorithms.
Intel cliams Xeon Phi arrays of were 1.38 times faster than GPUs compared with a single Intel Xeon Phi processor with 87 percent efficiency compared to 32 Nvidia Tesla K20 GPUs with 62 percent efficiency.
Mobile World Congress, considered by many experts as the most important tech trade show in the world, is coming to the U.S. Trade groups GSMA and CTIA are joining forces to bring a smaller version of the event to the U.S. in 2017.
GSMA Mobile World Congress Americas will debut Sept. 12 to 14, 2017, in San Francisco and will replace U.S. trade group CTIA’s Super Mobility conference. Super Mobility will continue this year in Las Vegas from Sept. 7 to 9.
The new conference will be the “first truly global wireless event” in the Americas, CTIA President and CEO Meredith Attwell Baker said in a statement.
The new trade show, however, will apparently be more focused, spotlighting the leading innovations from the North American mobile industry, John Hofman, CEO of GSMA, said in a statement.
The trade groups expect about 30,000 attendees and 1,000 exhibitors at the 2017 trade show, similar to the numbers from CTIA’s Super Mobility conference.
GSMA’s Mobile World Congress in Barcelona, Spain, earlier this year drew more than 100,000 attendees and 2,200 exhibitors. The 2017 Barcelona event will take place from Feb. 27 to March 2.
The new Mobile World Congress Americas will feature C-level speakers, exhibits featuring the latest mobile technologies, and a regulatory and public policy program.
There have been few details released on the new 10 core Skylake-X which will replace the Intel Core i7-6950X Extreme Edition which has just been released.
Intel has just released its Broadwell-E generation of ten, eight and six cores with Intel Core i7-6950X Extreme Edition being the fastest end the most expensive. But we have managed to get a few details about its replacement – the Skylake-X.
You can expect two SKUs, 140W X versions with 10 cores and the one will less cores called the K version. The new Extreme edition CPUs will have the new R4 socket. This new socket is also called LGA2066 some 55 higher than with the existing socket number.
There will be a Kaby Lake-X 4 core processor with 95+ W TDP using the same LGA 2066 R4 socket . Both Skylake-X and Kaby Lake-X support the new chipset that is known as Kaby Lake-X.
This is 200 series chipset will be the successor of the Skylake 100. The new chipset will come with up to 24 PCIe 3.0 PCI express lanes. In fact this is the only major difference in the chipset. It does support Octane storage techhonlogy, something that 100 series chipset cannot. The Kaby Lake 200 series chipset supports 6 SATA 3.0 ports, up to 10 USB 3.0 ports, DMI 3.0, up to three 4Xports for PCIe 3.0 drives.
If the Zen desktop core gives Intel some serious competition, we bet that Intel won’t charge $1700 for its highest end overclockable desktop CPU. Zen is still at least few months away, we expect it at late 2016 at best.