Data Center Knowledge | News and analysis for the data center industry - Industr's Journal
 
[Most Recent Entries] [Calendar View]

Thursday, November 7th, 2013

    Time Event
    12:30p
    Amazon Uses NVIDIA GPU to Power G2 Cloud Instance

    NVIDIA (NVDA) announced that its GRID technology is now available from Amazon Web Services (AWS) through its newly announced Amazon Elastic Compute Cloud (Amazon EC2) G2 instance, delivering GPU acceleration to users running graphics-intensive applications and games in the cloud.

    Available immediately, the new G2 instance is designed for cloud computing, with NVIDIA GRID GPUs making possible a new generation of GPU-accelerated SaaS applications — such as virtual workstations, accelerated virtual desktops and gaming-as-a-service. Based on the NVIDIA Kepler architecture, GRID GPUs include a powerful H.264 encoding engine that lets high-resolution graphics be compressed in real time and streamed to any internet-connected display.

    “Cloud computing has reached an inflection point driven by the growing variety of devices that can take advantage of application streaming,” said Jeff Brown, vice president and general manager of the Professional Visualization business at NVIDIA. “By using the graphics horsepower of NVIDIA GRID on AWS, companies can stream applications to more customers on more screens with a lower barrier to entry.”

    To make it easy for software companies to deploy applications onto G2 instances, OTOY has enabled a Windows- and Linux-based Amazon Machine Image (AMI) with OTOY’s ORBX middleware. SaaS companies can install their applications into OTOY’s AMI and begin streaming to web browsers within minutes.

    “AWS sees a growing benefit for adding GPUs to our cloud,” said Matt Wood, general manager of Data Science at AWS. “The NVIDIA GRID GPUs in our new G2 instances enable graphical applications to be rendered in the AWS cloud and streamed to a world with increasing internet bandwidth and proliferation of device types.”

    1:00p
    DuPont Fabros Buys Land for Expansion in Chicago

    The DuPont Fabros CH1 data center in Elk Grove Village, Ill. The company has bought adjacent land for a second Chicago-area data center. (Photo: DuPont Fabros)

    Data center developer DuPont Fabros Technology has bought 15 acres of land in Elk Grove Village, Illinois to serve as the home for its second data center in the Chicago market. The company says it has nearly filled its existing CH1 facility, and is ready to add more capacity to support customer growth.

    DuPont Fabros (DFT) paid $14.2 million to Prologis to acquire the land, which is adjacent to the CH1 facility. The company says it expects to commence development in the second quarter of 2014 and complete the building in the summer of 2015.

    “We are working on the design of CH2,” said CEO Hossein Fateh in a recent earnings call. “Due to the parcel shape, size and the largest single load the utility can deliver, we expect this building to  total approximately 26 megawatts of critical load.”

    Fateh said that the new facility will feature a new data center design the company has developed. Although the location requires some minor adjustments to DFT’s design, it “provides an optimal growth plan.”

    “Experience shows us that tenants like to take additional space within our campus as their requirements grow,” he said. “We have typically benefited from this embedded organic growth, and we are poised to capture similar growth in CH2.”

    DuPont Fabros operates more than 2.5 million square feet of data center space across four cities, providing 218 megawatts of electricity to power their tenants’ Internet and IT operations. Its Chicago tenants include cloud hosting provider Rackspace Hosting and colocation provider ServerCentral. Having expansion space nearby is attractive to fast-growing companies, which is why Fateh said DuPont Fabros was willing to pay a premium for the property. The price of land typically amounts to about 3 percent of the total cost of a data center.

    “The location for a data center is strategic,” said Fateh. “Being directly across the street, you could throw a tennis ball from this side of the street to the other and hit it. That’s where we really want to be. Did we maybe a little bit pay a premium to get it from ProLogis? Yes, probably. The campus environment of the data center where we have consistent organic growth from one data center to the other is where the leasing becomes easier. It’s definitely worthwhile to pay a little bit more and get the location you need.”

    1:30p
    Four Things Colos Should Consider When Evaluating Modular Construction Solutions

    This is part one of a three-part series by Stephen Madaffari, Principal of Data Centers Delivered, on how various sectors of the business serving the data center industry can effectively partner with modular data center construction companies to achieve success. He previously contributed this article, Four Reasons to Build Data Center Capacity Off-Site.

    Stephen MadaffariSTEPHEN MADAFFARI
    Data Centers Delivered

    Trends in modularizing data centers have varied vastly over the last few years–mostly due to the fact that the term “modular” can be interpreted in many different ways by the data center industry. As a result, some companies, and specifically colos, struggle when trying to define their own level of modularity, and assess if, when, and how to partner with an offsite modular data center builder.

    If you are a colo and looking to partner with a modular data center construction company, there are a few simple steps to consider.

    What Does Modular Mean?

    First, and most important, discuss, debate and determine what going modular means for your business. What business outcome will it help achieve? It’s difficult to implement modularity if data center managers are not sure at what level it’s beneficial to their operations–and there are many depending on the business model of the colo. For example, unit-level modularity (i.e. server or UPS) might work better for retail clients looking to drive a minimal amount of deployment rather than infrastructure modularity (i.e. fully prefabricated electrical or mechanical systems) more suited for wholesale colo clients.

    Assess Your Existing Infrastructure

    Second, assess which parts of your data center infrastructure make sense to modularize. Mechanical systems? Electrical systems? White space? Successful modularization is based on the ability to: 1) customize a repeatable infrastructure design and 2) associate the construction process with these key data center elements and deliver them in smaller pieces to accomplish a common goal.

    Determine Your Requirements

    Third, what does your data center system need to look like? What are its requirements? Engage in design consultation meetings to derive system architecture specifications. Incorporating those findings into modularized deployment plans is a vital part of the pre-construction process when a colo is looking to partner with an off-site modular data center builder such as Data Centers Delivered.

    Thoroughly Vet Your Partner

    Finally, research, evaluate and identify the modular data center construction company to partner with for build. The process of designing and building customized modular infrastructure can be complicated. and your decision should be considered and based upon the builder’s previous experience and record of reliability and safety. Further, it’s mutually beneficial for a colo to partner with a company that designs and builds all of the data center infrastructure needed in one place, including: structural, electrical, piping/mechanical, raised floor and multiple wall systems, among others.

    Keeping all design and construction under one roof allows for a low-risk build process to take place within a predictable time frame that ultimately delivers multiple production benefits and cost efficiencies now and in the future. Also, be sure to confirm the builder’s final product achieves all standards and guidelines for construction certification levels required of factory built data center solutions. For example, a product labeled with the “ETL Listed Mark” gives the client a level of confidence that all components are being built in accordance with nationally recognized standards.

    A Growing Marketplace for Modular

    Modularizing infrastructure to support the data center is growing fast due to the systematic off-site modular construction considerations and advantages discussed above. The future is bright for modular build companies that colos can truly embrace and trust when defining their own modular deployments and implementing them in a fashion that makes sense operationally and cost effectively for their business model.

    Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating. View previously published Industry Perspectives in our Knowledge Library.

    2:00p
    New IBM Cloud Ad Campaign Targets Competitor Amazon

    Brought to you by The WHIR.
    WHIR_logo_100

    IBM wants to remind the public that its cloud hosts 270,000 more websites than Amazon Web Services including “more of the most highly trafficked websites than Amazon and every other cloud provider.” This is part of IBM’s new online and print campaign that singles out AWS as its competition.

    According to a Washington Post report, the campaign will appear in major newspapers and business magazines, as well as online and outdoors this week. Many commentators have noted that it’s peculiar that IBM would mention a competitor in its ads because it has rarely done so in the past.

    Following IBM’s July acquisition of SoftLayer, IBM more recently announced that it would shut down its own public cloud platform and transition customers to the SoftLayer public cloud platform. The company said this would give clients higher performance and advanced functionality. After the acquisition, SoftLayer CEO Lance Crosby said, “IBM is hyperfocused on cloud.”

    This new marketing effort could be exactly what IBM is looking for to help its SoftLayer division gain even more traction against cloud giant AWS. It may also be significant that longtime SoftLayer competitor Rackspace Hosting was absent from the new advertisements, signalling that AWS and its specific market is IBM’s interest.

    There may be a specific animosity between the companies after a government contract worth roughly $600 million to provide cloud services to the CIA was given to Amazon Web Services after IBM withdrew from a series of legal challenges to the bidding procedure.

    Whether this new advertising campaign escalates into a marketing battle along the lines of soda rivals Coke and Pepsi remains to be seen. But with a cloud-computing market estimated to be worth a total of $40 billion annually and growing, it’s not only a battle over technical capabilities, but also consumer attention and opinion. Any advantage around how services are seen shall not be ignored.

    Original article published at: http://www.thewhir.com/web-hosting-news/new-ibm-cloud-ad-campaign-targets-competitor-amazon

    2:30p
    Equinix Plans Greenfield Data Center in London
    equinix-dc11-hallway

    The interior hallway at an Equinix data center. (Photo: Equinix)

    Data center service provider Equinix will build a new data center to expand its technology campus in Slough, England. The new facility, known as London-6, will be a “greenfield” project in which Equinix builds the facility from the ground up, allowing the colocation market leader to deploy its newest designs and technology.

    The expansion comes as Equinix is preparing to open the final phase of its London-5 data center, which will expand its capacity on the Slough campus by 1,475 cabinets. Equinix is inbvesting $82 million to build the first phases of the London-6 facility, which will be able to house 1,385 cabinets.

    CEO Steve Smith said Equinix is also taking steps to secure the long-term future of its Slough operation.  ”We expect to close shortly on the acquisition of the London-4 and London-5 IBX buildings for approximately $35 million,” Smith said on the company’s recent earnings call. “Concurrent with this acquisition, we will enter into long-term ground leases on the London-4, 5 and future London-6 properties, effectively giving Equinix up to 50 years of control of our Slough campus.”

    Here are some other highlights from the Equinix earnings call:

    More Business Suites: The company said it has experienced solid performance on Business Suites, the “wholesale light” offering that provides customers with dedicated suites for their gear, rather than cabinets or cages. As a result, it will be coming to more locations. “We are expanding offering to New York, a new build that will be tethered into the robust ecosystems in our Secaucus campus,” said Smith.

    Open-IX: Minimal Impact Seen: Analysts wondered how the emergent Open-IX movement and the US market entry of European exchanges LINX, DE-CIX and AMS-IX might impact the interconnection business, where Equinix has been the dominant player in the U.S. Chief Operating Officer Charles Meyers said Equinix is confiodent about its competitive position and the value of the customer ecosystems that have assembled within its data centers.

    “The overall value really of any exchange to a customer really is driven by the scale of the platform,” said Meyers. ” And not just the number of participants directly on the exchange, but the number of peers that can be accessed within the facility for private interconnection. Unsurprisingly, when you have a 15-year head start, as Equinix does, that’s where we really shine. And so we think the depth of our exchange is unmatched and feel very good about that.

    “We’ve got the scale, the commitment and the balance sheet to continue to invest heavily in the platform and we’re delivering industry-leading performance,” Meyers added. “So the dynamics of the markets are such that there typically is room for an alternative exchange in most markets. In some cases, we are actually the secondary exchange, but we’ve always maintained the position that competition is good for the customer. So for Open IX or some of these other ones, there may be an opportunity for them to establish presence, but as long as we price our services competitively and deliver superior value for that price, we believe the impact to our business will be minimal.”

    Bullish About the Cloud: Equinix executives see significant upside in the growing popularity of cloud computing. While many startups and small companies are moving directly to the cloud, Equinix executives said many of  these customers are soon hitting a point where they need to move to a hybrid infrastructure that combines public and private clouds. With its direct connectivity to both the Amazon and Azure clouds, Equinix says it’s ideally positioned to benefit from the growth of the hybrid hosting model.

    “We’re actually seeing that new entrants may be moving to that (cloud) model increasingly rather than buying a cab or 2 from us ,” said Meyers. “But what we see later is that as they scale and their requirements become more sophisticated, they very frequently look at how to move to a hybrid model and move certain workloads into a colocation environment. We have several documented case studies, where customers are doing that effectively within Equinix.

    “We believe that’s one of several major vectors within the cloud growth opportunity for us,” he added. “We think public cloud is a major development for enterprise CIOs, and they’re going to make significant use of it. But we’re going to also be net beneficiaries, in that as they sort of balance that with a hybrid cloud architecture for security performance and cost reasons.”

    “We’re very focused at bringing these big public cloud players to have access nodes around the world with us,” said Smith. “And that will also help pull the enterprises, as we’ve discussed, to want to connect private workload with public workload, increase of hybrid cloud environment. We’re starting to see that start to happen at a faster pace. I think that’s going to be an interesting play because at the same time small companies go to public cloud nodes, all those public cloud platform players are going to have nodes at Equinix, and you can access those public clouds at Equinix all day long. And that’s going to be an interesting environment as that starts to scale.”

    3:00p
    Fujitsu Lights up PCI Express with Intel Silicon Photonics
    A PCI Express copper cable vs the MXC optical cable. (Photo: Intel)

    A PCI Express copper cable vs the MXC optical cable used by Intel’s Silicon photonics technology. (Photo: Intel)

    Intel and Fujitsu have demonstrated a new server using Intel silicon photonics technology in an Optical PCI Express (OPCIe) design, which allows the storage and networking to be disaggregated, or moved away from the CPU motherboard. The benefit is that the components are easier to cool and users can choose which components they want to upgrade and are not forced to upgrade everything at the same time. The companies demonstrated the technology Tuesday in Munich, Germany at the annual Fujitsu Forum.

    Intel’s disaggregated rack has been building throughout 2013, beginning with the Open Compute Project announcement in January where a collaboration with Facebook defined next-generation rack technologies. After further demonstrating silicon photonics and a live working Rack Scale Architecture solution, Fujitsu has now reinforced a strategy of delivering a complete solution including cables and connectors that are optimized for photonics, with the first Intel Silicon Photonics link carrying PCI Express protocol.

    Light vs. Heat

    The photonics solution allows components to communicate using fiber optic cabling rather than electrical wiring. The new Fujitsu server relies on a FPGA (field-programmable gate array). Fujitsu took two standard Primergy RX200 servers and added an Intel Silicon Photonics module into each along with an Intel designed FPGA, which handled the necessary signal conditioning to make PCI Express “optical friendly.” Using Intel Silicon Photonics they were able to send PCI Express protocol optically through an MXC connector to an expansion box with several solid state disks and Xeon Phi co-processors.

    The resulting benefits from this architected solution include the ability to  increase the storage capacity of the server, ability to increase the effective CPU capacity of the Xeon E5′s, and cooling density – or how much heat needs to be cooled per cubic centimeter. Photonic signaling has 2 fundamental advantages over copper signaling. One is the noise caused by electromagnetic radiation (EMI) from adjacent wires, and a second is simply lighter cables.

    Intel has more than 10 years of experience in working with silicon photonics (SiPh), and is focusing on a full line of solutions around SiPh modules.

    << Previous Day 2013/11/07
    [Calendar]
    Next Day >>

Data Center Knowledge | News and analysis for the data center industry - Industry News and Analysis About Data Centers   About LJ.Rossia.org