Category Archives: Cloud Servers

Nimbix Cloud Adds Nvidia DGX-1 Servers for Deep Learning – HPCwire (blog)

April 11, 2017 Nimbix announced today the immediate availability of high-performance NVIDIA PascalGPUs using the NVIDIA DGX-1AI supercomputer in the Nimbix Cloud. For an on-demand rate, customers gain access to the industry-leading native bandwidth of eight NVIDIA NVLinkinterconnected NVIDIA Tesla P100 GPUs to launch or develop state-of the-art machine learning workflows, accelerated analytics and a host of other GPU-powered applications.

The Nimbix Cloud offers the most diverse set of GPU-powered machines available from a public cloud provider, spanning the NVIDIA portfolio and supporting configurations for both Intel x86 and IBM Power8 processors to deliver the best performance and economics available for both enterprises and developers. Nimbix Cloud machines are interconnected with industry-leading 56Gbps FDR and 100Gbps EDR Infiniband for optimal GPU cluster performance.

Nimbix has tremendous experience in GPU cloud computing, going all the way back to NVIDIAs Fermi architecture, said Steve Hebert, CEO of Nimbix. We are looking forward to accelerating deep learning and analytics applications for customers seeking the latest generation GPU technology available in a public cloud.

Combining the optimized performance of NVIDIA DGX-1 with Nimbixs cloud platform provides customers a flexible option to run their most challenging deep learning and AI workloads in an easy to use cloud system, said Charlie Doyle, senior director for DGX-1, NVIDIA.

In addition to the rich catalog of DGX turn-key workflows for deep learning, developers can use the PushToComputeTM feature of the JARVICE platform to import the latest versions of their custom applications into the Nimbix Cloud and make them available for consumption at scale immediately. Each application, along with its dependencies, executes in JARVICEs container runtime environment, which provides superior performance and scale. This includes sub-second launch times, faster execution, seamless access to supercomputing GPUs, automated heterogeneous data management, and rapid workflow deployment across multiple compute nodes either in a parallel or distributed paradigm. PushToComputeTM also facilitates continuous integration and continuous deployment (CI/CD) for the entire life cycles of containerized applications.

About Nimbix

Nimbix is the leading provider of purpose-built cloud computing for machine learning, AI and HPC applications. Powered by JARVICE, the Nimbix Cloud provides high-performance software as a service, dramatically speeding up data processing for Energy, Life Sciences, Manufacturing, Media and Analytics applications. Nimbix delivers unique accelerated high-performance systems and applications from its world-class data centers as a pay-per-use service.

Source: Nimbix

Originally posted here:
Nimbix Cloud Adds Nvidia DGX-1 Servers for Deep Learning - HPCwire (blog)

IBM Accelerates AI on Cloud – Database Trends and Applications

Apr 10, 2017

IBM has announced it is the first provider to make the NVIDIA Tesla P100 GPU accelerator available on the cloud. The combination of NVIDIAs acceleration technology with IBMs Cloud platform is intended to help organizations more efficiently run compute-heavy workloads, such as artificial intelligence, deep learning and high performance data analytics.

The latest NVIDIA GPU technology delivered on the IBM Cloud is opening the door for enterprises of all sizes to use cognitive and AI to address complex big data challenges, said John Considine, general manager of cloud infrastructure at IBM. IBMs global network of cloud data centers along with its advanced cognitive and GPU capabilities is helping to accelerate the pace of client innovation.

IBM has been working closely with NVIDIA since 2014 to bring GPU technology to the cloud. The company introduced the NVIDIA Tesla K80 GPU accelerator in 2015 and the Tesla M60 in 2016, and launched the Tesla P100 builds on IBMs leadership in bringing the latest NVIDIA GPU technology to the cloud for machine learning, AI and High Performance Computing workloads.

With Tesla P100 GPU accelerator on the cloud, IBM says it will make it more accessible for businesses in industries including healthcare, financial services, energy and manufacturing to extract valuable insight from big data. For example, clients in the financial services industry can use GPUs on the IBM Cloud to quickly run complex risk calculations; healthcare companies can more quickly analyze data or identify possible genetic variations; and energy companies can identify new ways to improve operations.

Clients will have the option to equip individual IBM Bluemix bare metal cloud servers with two NVIDIA Tesla P100 accelerator cards. The Tesla P100 provides 4.7 teraFLOPS of double-precision performance and 16 gigabytes of GPU memory in a single server to accelerate compute-intensive workloads. The combination of IBMs connectivity and bare metal servers with the Tesla P100 GPUs provides clients with higher throughput than traditional virtualized servers. This high level of performance can allow clients to deploy fewer, more powerful cloud servers to more quickly deliver increasingly complex simulations and big data workloads.

The Tesla P100 joins NVIDIAs portfolio of GPU offerings on the IBM Cloud, including the Tesla M60, Tesla K80 and Tesla K2 GPUs. IBM plans to make the Tesla P100 GPU accelerators available in April 2017.

To learn more, visitwww.ibm.com/cloud-computing/bluemix/gpu-computing.

The rest is here:
IBM Accelerates AI on Cloud - Database Trends and Applications

Amazon, Microsoft, and Google want to own the Cloud – TechSpot

Amazon, Microsoft, and Google (Alphabet Inc.) have been battling for dominance in the arena of cloud computing for almost as long as the term has been around. The growth of software as a service (SaaS) was started and fueled by enterprises seeking ways to deliver tools over large networks. This industry is not only growing, but it is also evolving to include more than just SaaS.

According to the Wall Street Journal, the combined capital expenditures for the big three were $31.54 billion in 2016, which is 22 percent higher than investments from 2015. A substantial portion of these costs has gone into server farms.

Amazon just announced plans to plunk down a massive collection of data centers in Stockholm, reported WSJ.

Amazon did not reveal how much the new region of data centers would cost, but analysts estimate that several hundred million dollars is not an unreasonable guess. An investment of this size would obviously have a proportionately high ROI (return on investment), and according to the Wall Street Journal, it does.

Software as a service is evolving into what is being called, hyperscale computing. It is not just software in the cloud that businesses are looking for anymore. Now they want their infrastructures hosted as well.

Gartner research firm estimates that in 2016 alone, enterprises and startups spent $500 billion in computing, storage, networking, database technology and more.

Most of that spending was in moving from company-owned servers to servers in the cloud. Part of the draw to infrastructure as a service (IaaS) is scalability. With physical, on-premises servers, companies are limited during peak traffic periods.

The WSJ provides an example of how a fantasy football site crashes right before kickoff due to the traffic volume being higher than its servers can handle. With IaaS, resources are virtually unlimited and can be scaled on-demand and in real time.

Oracle is another company that has been seen investing in data-center development. In its most recent fiscal year, Oracle funded $1.7 billion on server farms. While this expenditure is small in comparison to what the big three have budgeted in the last year, Oracle is confident that it can compete.

Steve Daheb, senior vice president of product marketing for the company thinks Oracles infrastructure services are superior to its competitors. For this reason, he feels the company does not need to spend as much on data centers. However, the Journal points out that the numbers do not seem to support this claim.

In the most recent quarter, [Oracles] infrastructure-as-a-service revenue grew 17% to $178 million. Net sales at Amazon Web Services, comprised largely of Amazons infrastructure-as-a-service business, grew 47% to $3.54 billion in its most recent quarter.

It would seem that Oracle is significantly lagging behind and it will be difficult to catch up. Amazon, Microsoft, and Google have already invested so much into computing capacity that they have built a barrier that can only be breached by investments going into the tens of billions of dollars.

Such massive expenditures are easy to see when a region for Amazon Web Services runs between $300-$600 million. Add to that the maintenance and other overhead involved with running the facility, and one can understand why Amazon, Microsoft, and Google are expected to dominate this growing industry.

Read more here:
Amazon, Microsoft, and Google want to own the Cloud - TechSpot

Spending forecast shows early signs of on-premise server demise – ComputerWeekly.com

The strong US dollar will lead to $64bn less growth in IT spending, according to Gartners latest forecast.

Worldwide IT spending is projected to total $3.5tn in 2017, a 1.4% increase from 2016. But the growth rate is down from the previous quarters forecast of 2.7%, due in part to the rising US dollar.

The 10 most important things you need to know about GDPR, and a jargon-buster explanation for some of the key terminology.

By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.

You also agree that your personal information may be transferred and processed in the United States, and that you have read and agree to the Terms of Use and the Privacy Policy.

Cloud computing is having a major impact on datacentre system spending, according to Gartner.

We are seeing a shift in who is buying servers and who they are buying them from, said John-David Lovelock, research vice-president at Gartner.

Enterprises are moving away from buying servers from the traditional suppliers and instead renting server power in the cloud from companies such as Amazon, Google and Microsoft. This has created a reduction in spending on servers, which is impacting the overall datacentre system segment.

Lovelock estimated that although the cost of running a workload on Amazon Web Services (AWS) is probably more expensive than running the same workload on-premise, the difference soon diminishes once all the on-premise support infrastructure and staffing is taken into account.

AWS might be a little more expensive, but you dont need staff or lighting, he said. If all the additional costs to support a workload running on an on-premise server are accounted for, cloud-based IT works out cheaper, said Lovelock.

Lovelock said he expects enterprises to continue to buy two-way and four-way servers, but the decline in server sales will eventually hit a tipping point, where the mainstream server makers can no longer financially justify products. Cloud providers do not generally buy from the major server manufacturers, preferring contract manufacturers. This is exacerbated by enterprises buying fewer on-premise servers, and choosing cloud services for new workloads.

As Computer Weekly reported previously, HPEs first-quarter 2017 results were affected by significantly lower demand for servers from one of its tier-one service provider customers. While CEO Meg Whitman did not confirm whether this was due to the customer choosing to buy from a contract manufacturer, it shows the precarious position the top server makers now face as their enterprise customers opt for cloud computing.

Gartner has forecast that the 2017 worldwide IT services market will grow by 2.3% in 2017 down from 3.6% growth in 2016. The analyst firm attributed the modest changes to the IT services forecast this quarter to adjustments to particular geographies as a result of potential changes in direction for US policy both foreign and domestic.

Gartner expected the business-friendly policies of the new US administration to have a slightly positive impact on the US implementation service market, with the US government thought to be planning a significant increase in infrastructure spending over the next few years.

There are already early signs of companies starting out on the road to a cloud-only world. Lovelock pointed to the $2bn, five-year outsourcing contract Snapchat had signed with Google Cloud.

In a US Securities and Exchange S1 filing, the photo messaging company said its software and computer systems had been built to use computing, storage capabilities, bandwidth and other services provided by Google.

We currently have a capital-light business model because we work with third-party infrastructure partners primarily Google Cloud to run and scale our services rather than building our own infrastructure, which would require significant up-front capital and resources, said the companys filing.

We believe working with these partners will result in lower costs for us in both the short and long term. Large-scale infrastructure providers offer several advantages, including global scale to serve our audience, the ability to handle peak demand more economically, and purchasing power to procure equipment directly from infrastructure equipment suppliers that results in lower net costs to us.

Google is doing IT for Snapchat, added Lovelock.

Driven by strength in mobile phone sales and smaller improvements in sales of printers, PCs and tablets, worldwide spending on devices (PCs, tablets, ultramobiles and mobile phones) is projected to grow by 1.7% in 2017, reaching a potential $645bn.

But in mature markets, spending on laptops, PCs, tablets and smartphones is expected to remain flat in 2017, according to Gartner. There is a shift in spending over to IoT [information of things] devices, said Lovelock.

He predicted that the bring-your-own-device [BYOD] trend would carry on to a point where many organisations will no longer see a need to supply devices to their employees. This would be a great way to get desktop IT off your books, he said.

Organisations that have high security requirements will still need to supply devices to staff, said Lovelock, but he added: If enterprises put critical applications in the cloud, device security is less important. Enterprise containers, software and policy will all help to secure BYOD devices.

Along with BYOD, Lovelock expected users to buy software and apps for their own devices, which they also use at work. Some of these will be free and funded by advertising, such as those on the Google Play store, while some, such as on Apples App Store, may require payment from the consumer.

See original here:
Spending forecast shows early signs of on-premise server demise - ComputerWeekly.com

Private Cloud Server Market Global Industry Analysis, Share, Growth, Trends, and Forecast 2016 2024 – MilTech

Private Cloud Server Market is suitable to businesses with unpredictable or dynamic computing requirements that need direct control over their environment. Companies that have a data-center and developed IT infrastructure will fuel the industry demand over the future. Technological advancements pertaining to firewalls, threat monitoring, and space availability are expected to offer ample opportunities.

Data security and protection is a prime concern restraining the growth of the private cloud server market. Third party attacks, for instance, viruses can threaten the reliability of these solutions. Stringent government rules and regulations tightly protecting the collection of private customer information can provide obstacles to industry growth. Installation, maintenance, and hardware costs of the technology are very high and can discourage businesses from integrating them into their businesses.

Request for an in-depth table of contents for this report @ https://www.gminsights.com/request-toc/upcoming/1286

Private Cloud Server Market size will be fueled by the rising awareness of the benefits of private cloud server demand among enterprises. These solutions maximize resource flexibility and usability and elevate the business performance. The risk of organizational data theft and privacy breach by public cloud usage is extensively reduced as the enterprise is responsible for the management. These solutions allow access to authorized members and business units of the organization by implementation of a virtualized environment that ensures the optimal management of available computer resources. As loss of data can have a major impact on an organization, private cloud server market software allows data recovery, which is vital in times of system failures.

The transition of businesses from profit-focus to customer-focus has given rise to big data. Big data can provide business insights based on historical and real-time data. This requires large storage capabilities and access to information databases across various business units of the organization. Big data platforms hosted by these servers can offer better scalability, lower costs and better resource utilization as compared to traditional big data deployments driving the private cloud server market demand. The growth of hybrid cloud technology is predicted to create lucrative market opportunities. Hybrid platforms combine the private and public clouds to enable effective workload movement between them. It is particularly useful in dynamic or highly changeable work environments and in businesses which require high access to big data.

Private cloud server market segmentation by application includes storage, disaster recovery, development and testing. Storage segment is expected to lead the industry owing to the high availability of customer and organizational data. The data recovery segment is predicted to gain prominence during the forecast timeframe. Critical information requires a back-up in case of system failure, theft or external attack.

Segmentation of the private cloud server market based on organization size comprise of SMBs and large businesses. SMBs are predicted to grow significantly owing to rising awareness about the benefits of the technology.

By industry verticals the private cloud server market consists of BFSI, defense, entertainment and infotainment, retail and consumer goods, aerospace, travel and hospitality, government, healthcare and life sciences, IT and telecommunication, logistics, and others. Managed private services are estimated to experience substantial growth in the retail vertical segment. The requirement of resources at high speed access is essential for the smooth functioning of businesses operating in this sector. Private cloud server market solutions are useful in accumulating and managing of real-time data which can be used in capacity planning of the organization. Although public cloud server market offer economies of scale, higher ROI and cost benefits, private technology offers better reliability and security and ensures data is not lost or misplaced.

The private cloud server market can by integrated by platform types comprising on-premise and external private clouds. On-premise private technology is primarily used by businesses that have already invested in data-center capabilities, infrastructure and have hardware available that can be repurposed for this these solutions. Moreover, on-premise private cloud server market provide higher security as compared to external servers. External service providers offer off-site solutions to data storage to multiple tenants. They provide resources to multiple users, for instance physical hardware. The compromise on security is met with lower costs due to hardware reusability. These benefits are factors promoting the increasing adoption of external solutions segment in the private cloud server market.

Make an inquiry for buying this report @ https://www.gminsights.com/inquiry-before-buying/1286

The U.S. private cloud server market is anticipated to lead the private cloud server market due to escalating demand for secure severs and rapid industrialization. APAC is estimated to witness high growth rate during the forecast period. This is attributed to increasing awareness of the advantages and rising demand for big data storage in emerging economies, for instance China and India.

Players operating in the private cloud server market need to employ business strategies such as collaborations, mergers and acquisitions to sustain growth in the global marketplace. Vendors providing these solutions include Amazon, Microsoft Corp, D-Link Corporation, Cisco Systems, Accenture PLC, Rackspace, BitCloud, Seagate, Google, Just Cloud, MEGA Ltd, ElephantDrive, Hewlett-Packard Development, Hivelocity Ventures, International Business Machine, and Cisco systems. On February 2016, Rackspace announced the release of Rackspace private cloud powered by Red Hat. This strategy will deliver the most easy-to-use and reliable Open stack private and hybrid clouds across the world.

Browse Related Reports:

https://www.gminsights.com/industry-analysis/healthcare-cloud-computing-market

About Global Market Insights

Global Market Insights, Inc., headquartered in Delaware, U.S., is a global market research and consulting service provider; offering syndicated and custom research reports along with growth consulting services. Our business intelligence and industry research reports offer clients with penetrative insights and actionable market data specially designed and presented to aid strategic decision making. These exhaustive reports are designed via a proprietary research methodology and are available for key industries such as chemicals, advanced materials, technology, renewable energy and biotechnology.

Contact Us:Arun Hegde Corporate Sales, USA Global Market Insights, Inc. Phone:1-302-846-7766 Toll Free:1-888-689-0688 Email:sales@gminsights.comWeb:https://www.gminsights.comBlog:https://gminsights.wordpress.comConnect with us:Facebook|Google+|LinkedIn|Twitter

Original post:
Private Cloud Server Market Global Industry Analysis, Share, Growth, Trends, and Forecast 2016 2024 - MilTech

How I made my own VPN server in 15 minutes – TechCrunch


TechCrunch
How I made my own VPN server in 15 minutes
TechCrunch
While Algo VPN makes it easier to set up a VPN server on DigitalOcean, AWS, Microsoft Azure and Google Cloud, I also tried using it with Scaleway to see if you could use it on any hosting provider. And it worked perfectly fine on the smallest Ubuntu ...

See original here:
How I made my own VPN server in 15 minutes - TechCrunch

3 Top Dividend Stocks in the Server Industry – Bloomington Pantagraph

Servers are, in many respects, the backbones of technology industry. They serve as critical infrastructure on which a host of other functions rest. Considering their importance, it should come as no surprise that the server market is one of the largest sub-industries in all of tech; global server sales totaled $14.6 billion in Q4 2016, according to research firm IDC.

Cloud computing is undoubtedly changing the industry, but the leading server manufacturers remain several well-known tech companies. In this article, we'll look at three dividend-paying, large-cap stocks that investors interested in the server market should consider.

Image source: Getty Images.

It is the world's largest server provider, and servers sales constitute the single-largest revenue contributor among Hewlett-Packard Enterprises'(NYSE: HPE) six reporting sub-sections, accounting for $14.0 billionof the IT giant's $50.1 billion FY 2016 sales. Like the rest of the market, Hewlett-Packard Enterprise's server sales have remained fairly static, declining just 0.7% from 2015 to 2016.

For those unfamiliar with it, the company was formed from the 2015 split of the iconic tech conglomerate Hewlett-Packard. HP Inc. was created to house the PC and printer division, which remain in secular decline, while Hewlett-Packard Enterprise was created to hold the more promising businesses. However, looking at current analyst estimates, it isn't immediately clear that those businesses are, in fact, more promising.

Wall Street analysts see the company's 2017 estimated sales falling 28% and its FY 2018 revenue declining a further 17%.Interestingly, this kind of sales attrition did not reveal itself in HPE's most recent 10-K filing; FY 2016 total revenues fell a mere 3.8%.As such, investors considering buying HPE for its income-producing abilities should keep in mind that many observers widely expect its business to gradually deteriorate in the coming years. Buyer beware.

In terms of its dividend stats, HPE yields slightly less than the market average of 1.94%. The company raised its dividend in December, so its shares yield 1.28% on a trailing-12-month rate basis, and 1.48% on a forward basis. It also enjoys plenty of room for future rate increases -- HPE's payout ratio sits at just 12.6% -- though its revenue growth outlook profile remains a critical consideration for anyone looking at HPE stock.

Unlike HPE, the dynamics at International Business Machines'(NYSE: IBM) server business better reflect the on-the-ground reality of the server industry. IBM organizes its sales of its three main server products under the umbrella of its Systems reporting segment. For its fiscal 2016, sales of IBM's physical servers plummeted 21.8% to $5.9 billion. This might seem dire, but it's important to remember that cloud computing and IBM's various software-as-a-service offerings still rely on servers to power their systems. In fact, IBM's cloud revenues grew 49% in 2016 to a run rate of $5.8 billion, or roughly the same size as IBM's current server sales base. Said another way, IBM's shifting business model hasn't eroded the server as a key driver of its sales. Rather, the rise of cloud computing means IBM is selling server-powered software to customers, rather than physical servers themselves.

In terms of its dividend payments, IBM shares currently yield 3.2%,well above market averages. The company has done an admirable of job steadily increasing its payouts as well. Though not a Dividend Aristocrat, IBM has increased its annual cash payments for 17 consecutive years. It also bears noting that IBM has been one of the most active companies in corporate America in terms of rewarding its shareholders through stock buybacks as well. So, while IBM doesn't command the same kind of server market share as HPE, the company's continued progress in its ongoing business pivot and above-average yield make it one of the more appealing income-producing stocks to appear on this list.

Another interesting dividend play tied to the server industry could be investing in one of the suppliers that provides hardware components to the industry at large. If that's your flavor, then chip giant Intel (NASDAQ: INTC) is perhaps your best bet.

Intel is widely acknowledged to control an estimated 99% of the server microprocessor market, making it a proxy for the industry as a whole. As such, companies that make and sell physical servers to customers, like HPQ and IBM, use Intel chips to power their products. However, major cloud computing providers -- like Amazon, Alphabet, and others -- also use Intel server chips to power their respective cloud infrastructures. As such, investing in Intel can be seen as a nice way to invest in the industry growth as a whole, while saving the headache of having to determine which specific server manufacturer will perform the best.

Seasoned tech investors will also recognize Intel as one of the industry's best-known dividend stocks. Though fairly mature from a sales growth perspective, Intel handsomely rewards its shareholders through cash distributions and stock buybacks. Intel's current dividend yield sits at a respectable 2.9%. Its 49% payout ratioshould give the company the flexibility to continue to raise payouts in the future as well. And while not a Dividend Aristocrat -- Intel missed raising its dividend as recently as two years ago -- the company has demonstrated a general commitment to growing its dividend. As just one example, Intel's cash distributions grew from just $0.07 in its FY 2000 to $1.04 in its FY 2016. So, while Intel doesn't directly produce servers, its shares can still be an interesting way for investors to generate dividend payments from the server industry all the same.

10 stocks we like better than Intel

When investing geniuses David and Tom Gardner have a stock tip, it can pay to listen. After all, the newsletter they have run for over a decade, Motley Fool Stock Advisor, has tripled the market.*

David and Tom just revealed what they believe are the 10 best stocks for investors to buy right now...and IBM wasn't one of them! That's right -- they think these 10 stocks are even better buys.

*Stock Advisor returns as of April 3, 2017.

Suzanne Frey, an executive at Alphabet, is a member of The Motley Fool's board of directors. Andrew Tonner has no position in any stocks mentioned. The Motley Fool owns shares of and recommends Alphabet (A and C shares) and Amazon. The Motley Fool recommends Intel. The Motley Fool has a disclosure policy.

Read the original post:
3 Top Dividend Stocks in the Server Industry - Bloomington Pantagraph

DigitalOcean launches monitoring service for its cloud servers – BGR India


BGR India
DigitalOcean launches monitoring service for its cloud servers
BGR India
US-based Cloud infrastructure provider DigitalOcean on Wednesday launched a monitoring service that provides insight into the resource utilization and operational health of every Droplet (Cloud server). Developers can collect and visualize metrics in ...

View original post here:
DigitalOcean launches monitoring service for its cloud servers - BGR India

Pebble releases update so that watches will work when cloud servers shut down – Liliputing

Pebble was one of the first major players in the smartwatch space. But when Fitbit acquired the company late last year, pretty much the entire product lineup was scrapped and plans were made to only continue supporting Pebble software and services through the end of 2017.

But that doesnt necessarily mean that your Pebble watch will become useless at the end of the year.

Pebble has just released updates to its Android and iOS apps that will allow you to keep using Pebble watches even after the Pebble cloud servers shut down.

The new apps will let you skip the login process, sideload apps, and install updated firmware even if a Pebble authentication server cannot be reached something that you can be pretty sure will happen in the future.

Pebble has also released new developer options, including an offline mode option that will prevent an app from even trying to communicate with Pebble servers. Developers can also sideload a service configuration to change which servers the Pebble app communicates with, which could allow developers to continue offering cloud-based services for Pebble smartwatch users even after the Pebble servers are gone.

Now its just up to developers to decide whether they want to continue investing in updates to their apps for a dead platform.

Other changes in the latest app include the removal of options to suggest new features or contact support, and health data collection and telemetry reporting have been removed.

via The Verge

See the original post here:
Pebble releases update so that watches will work when cloud servers shut down - Liliputing

IBM Bare Metal Cloud Targets AI with New P100 GPUs – HPCwire (blog)

IBM announced today that it will be addingNvidia P100 graphics processors to its Bluemix cloud later this month, becoming the first major global cloud vendor to provide the high-end Pascal GPUs. Big Blue is targeting the new hardware at customers who run compute-heavy workloads, such as artificial intelligence, deep learning, data analytics and high-performance computing.

Unlike Nimbix, the heterogeneous cloud vendor that began offering NVLinkd Nvidia P100 GPUs on the IBM Minsky Power8 platform last October (2016), IBM will be using PCIe form factor cards within an Intel x86 server. This is not really a surprise since IBM operates most of its cloud servers on Intel-based chip sets. Customers will be able to add up to two Nvidia P100 cards to a dual Xeon E5-2690 v3 machine (24-core CPUs running at 2.6 GHz).

The IBM cloud does have some Power server options for specific big data workloads but it does not have an expanded assortment of Power, saysJay Jubran, Global Offering Management for Compute at IBM Cloud. A plan tointegrate Power8 based systems with NVIDIA P100 GPUs into the IBM cloud portfolio is underway. We are are working side by side with the Power Systems team to ensure that IBM Cloud will deliver access to the best of IBM technology to allow customers to run HPC and AI workloads, Jubran told us.

The Power8 Minsky platformenables tight coupling of the Power CPU and P100 GPU overNvidias proprietary NVLink interconnect. Themezzanine form factor P100 also provides nearly 13 percent better raw performance than the PCIe card, 5.3 double-precision teraflops versus 4.7. Both versions provide 16 gigabytes of HBM2stacked memory.Networking on the IBM cloudstands at10 Gigabit Ethernet today with IBM stating thatfuture platforms might go up to 25 Gigabit Ethernet.

IBM will be first to the P100 punch in terms of major cloud providers, but as we have seen, other cloud purveyors are advancing with P100 plays of their own. Heres a rundown:

Nimbix As mentioned above, Nimbix added IBM Power S822LC for HPC systems (codenamed Minsky) to its heterogeneous HPC cloud platform last October. Target markets include high-performance computing, data analytics, in-memory databases, andmachine learning.

Cirrascale On its GPU-driven deep learning infrastructure as a service, San Diego, Calif.-based Cirrascale offers a number of P100-based server configurations, including four-way and eight-way Intel-based GPU servers and IBM Power8 Systemswith two and four GPU options.

Google The Google Cloud platform website states that P100s are coming soon. Google will also be incorporating AMD FirePro S9300 x2 GPUS into its infrastructure. Google began offering K80 GPU-equipped virtual machines (as a beta release) in February of this year.

Microsoft Microsoft last month revealed blueprints for a new open source P100-based accelerator HGX-1 developed under Project Olympus. Its an accelerator box with eight Tesla P100s, connected in the same hypercube mesh as the Nvidia DGX-1 server and also leveraging the NVLink interconnect. The HGX-1 hooks to servers via PCIe interface.Were to assume the boxes, being manufactured by Ingrasys, will show up on Azure but Microsoft hasnt indicated when that will be. The company has had some notable delays in GPU rollouts announcing a planned K80 instance in September 2015, and AWS beating them to general availabilitya year later.

Tencent Two weeks ago, Chinese cloud giant Tencent said it will offer a range of cloud products that will include GPU cloud servers incorporating Nvidia Tesla P100, P40 and M40 GPU accelerators and Nvidia deep learning software. Tencent Cloud launched GPU servers based on Nvidia Tesla M40 GPUs and NVIDIA deep learning software in December; it expects to integrate cloud servers with up to eight Pascal-based GPUs each by mid-year.

Reigning cloud king Amazon does not yet offer Nvidias Pascal-based silicon (the P100 or the P40 inferencing engine). Amazons most recent P2 instance family is backed by Kepler-generation K80 parts, rolled out last September (2016).

IBM emphasized the advantage of its bare metal cloud offering, compared to the multi-tenant environments of AWS and the other mega-cloud providers, especially for HPC workloads. The main reason why people come to IBM cloud, other than the global presence, is the performance and consistency of having access to the bare metal. The bare metal allows us to give better performance than any other virtualized environment with the same specification because we do not have the hypervisor tax which is roughly 10-15 percent of the CPU power, said Jubran.

We find HPC workloads typically find their way to the IBM cloud. If the customer is looking to run HPC on an hourly basis sometimes youll see them go to other clouds, but in terms of monthly consumption we have the best offering in terms of performance and price value, he added.

The bare metal infrastructure is also attractive to the graphics community, for gaming, especially a subset called cognitive gaming, and for engineering, saidJubran. Financial services, healthcare, and retail are all target verticals.

Customers that prioritizehighly elastic resources and pay-by-the-sip pricing typically go to IBMs competitors, Jubran noted, but their core customers are the ones who understand the performance metrics that IBM offers.

We are attracting both digital customers looking for performance, gaming customers and born on the web type customers who are looking for bare metal performance, but scalability of the cloud. And we also get in the higher end of the spectrum in terms of enterprise and that is because of IBM obviously being an enterprise-focused company from day one and they put trust in IBM to bring their workload to our datacenters. So having both aspects of the spectrum keeps us on the innovative side in terms of digital and keeps us on the high-performance secure side for the enterprise, said Jubran.

Aside from the advantage of this enterprise trust factor, IBMs distributed model of 50 datacenters (built up since the Softlayer acquisition in 2013 for a reported $2 billion) gives them the geo-precision to provide local data sovereignty for their customers and is a natural fit foredge computing (important for AI training workflows and for IoT). For many customers, proximity of compute and data are far more important than saving on compute cost offered by the greater elasticity of mega-datacenters. Atypical IBMdatacenter unit consists of roughly 20,000 servers; in the hyperscaler world, thats pretty small.

The Tesla P100 joins Nvidias portfolio of GPU offerings on the IBM Cloud, including the older Tesla K2 GPU, the Tesla M60 for virtualized graphics and the Tesla K80, which IBM added in 2015, about a year ahead of the competition. IBM expects most of its K80 customers will be migrating over to the P100 servers as they begin adding the parts later this month. We also expect newcomers into the AI platform as the P100 is the most powerful GPU in terms of AI workloads that are based on TensorFlow, Caffe, Nvidia SDK or any of the AI SDKs available out today, said Jubran. With so much focus from all the different industries in AI, I think you will see more and more of those workloads coming to IBM cloud and the P100 will enable that. If you look at the Nvidia material for P100 it is the most powerful GPU for both training and inferencing, the two aspects of AI.

With all key deep learning frameworks GPU-accelerated and over 400 HPC applications in a broad range of domains, including the top 10 high performance computing applications, IBM Cloud customers can quickly tap into the power of the our GPU platform to boost performance, accelerate time to results and save money, NvidiasVice President of Accelerated Computing Ian Buck wrote in a blog post.

The cost for the new Pascal-based hardware is $750 per month per P100 GPU card, tacked on to thepriceof the server. This adds a 50 percent premium over the cost of the K80s ($500 per card) but the P100 card offers a 60 percent additional performance improvement over the K80.That should make switching a no-brainer and while IBM wont be forcing customers with active workloads off the K80, they are planning to sunset the older Teslas as inventory depletes.

Editors note April 6, 2017: In an earlier version of this article, we reported (based on information IBM shared with us) that the Power8 Minsky platformwas not on IBMs cloud roadmap. After the article was published, IBM contacted us to let us know that it does have plans toincorporate Power8 based systems with Nvidia P100 GPUs intoits cloud portfolio. We have amended the story to include this updated information.

Original post:
IBM Bare Metal Cloud Targets AI with New P100 GPUs - HPCwire (blog)