Nvidia's new TensorRT speeds machine learning predictions

Nvidia's new TensorRT speeds machine learning predictions

Nvidia has released a new version of TensorRT, a runtime system for serving inferences using deep learning models through Nvidia’s own GPUs.

Inferences, or predictions made from a trained model, can be served from either CPUs or GPUs. Serving inferences from GPUs is part of Nvidia’s strategy to get greater adoption of its processors, countering what AMD is doing to break Nvidia’s stranglehold on the machine learning GPU market.

Nvidia claims the GPU-based TensorRT is better across the board for inferencing than CPU-only approaches. One of Nvidia’s proffered benchmarks, the AlexNet image classification test under the Caffe framework, claims TensorRT to be 42 times faster than a CPU-only version of the same test — 16,041 images per second vs. 374—when run on Nvidia’s Tesla P40 processor. (Always take industry benchmarks with a grain of salt.)

Serving predictions from a GPU is also more power-efficient and delivers results with lower latency, Nvidia claims.

TensorRT doesn’t work with anything other than Nvidia’s own GPU lineup, and is a proprietary, closed-source offering. AMD, by contrast, has been promising a more open-ended approach to how its GPUs can be used for machine learning applications, by way of the ROCm open source hardware-independent library for accelerating machine learning.

Source: InfoWorld Big Data

SolarWinds Simplifies Enterprise Network Management With New Products

SolarWinds Simplifies Enterprise Network Management With New Products

SolarWinds has announced two new unified network management solutions — SolarWinds Network Operations Manager and SolarWinds Network Automation Manager — that the company will showcase at Cisco Live® US, June 25-29, 2017, in Las Vegas at the Mandalay Bay® Convention Center.

Ensuring mission-critical enterprise networks remain optimized and running at scale in today’s increasingly complex hybrid IT environments can be a daunting task. Selecting, deploying, and integrating software to help monitor such networks often adds to the challenge. These new SolarWinds offerings build on the comprehensive, single-pane-of-glass visibility of the SolarWinds® Orion® Platform to simplify large-scale network management. These unified products provide everything needed to monitor and manage networks in a seamlessly integrated solution available as a single purchase.

“Leveraging our experience with thousands of large-scale customers, we’ve designed our new unified network management solutions to streamline almost everything about network management software,” said Christoph Pfister, executive vice president of products, SolarWinds. “In essence, we’ve taken our most popular network monitoring and management capabilities and unified them into products specifically designed for large-scale networks. We’ve applied the SolarWinds philosophy of powerful, affordable, and easy-to-use software to enterprise network management.”

SolarWinds Network Operations Manager

Based on the SolarWinds Orion Platform, SolarWinds Network Operations Manager provides:

  • Fault, availability, and performance monitoring
  • Critical path hop-by-hop monitoring both on premises and in the cloud thanks to the SolarWinds NetPath™ feature
  • Bandwidth and traffic monitoring
  • Switch port monitoring with end-user tracking
  • Dynamic drag-and-drop dashboard creation with the SolarWinds PerfStack™ dashboard to better visualize networking, systems, and application relationships

SolarWinds Network Automation Manager

SolarWinds Network Automation Manager includes all the features of SolarWinds Network Operations Manager and adds:

  • Network configuration automation
  • DHCP, DNS, and IP address management
  • High availability for instantaneous failover

 See SolarWinds Unified Network Management Solutions in Action at Cisco Live US, Booth 1111

 Meet SolarWinds Head Geeks™ Patrick Hubbard, Leon Adato, and Destiny Bertucci, as well as other SolarWinds product experts at booth 1111 to participate in live demonstrations of SolarWinds Network Operations Manager and SolarWinds Network Automation Manager, along with the complete portfolio of products from SolarWinds. The company was recently named the global market leader in network management software by industry analyst firm International Data Corporation® (IDC®) in its Worldwide Semi-Annual Software Tracker. 

Join “Blame the Network, Hybrid IT Edition”

On Tuesday, June 27, 2017 from 5-5:30 p.m. Pacific Time, Hubbard will also present, “Blame the Network, Hybrid IT Edition,” as part of the Cisco Live US 2017 Think Tank at booth 1601. During the talk, Hubbard will address the following:

“Hybrid IT, the conflagration of distributed on-premises, cloud, and Software-as-a-Service (SaaS) applications has a major downside: the opportunity for businesses to blame network engineers for service issues. Troubleshooting and resolution are further complicated by application evolution, causing many of the go-to tools we’ve relied on for years to become less effective, or worse, obsolete. In this session, you’ll learn how both IT organizations and technology vendors are trying to not only catch up, but get ahead in this time of IT transition, and continue to assure visibility of the complete Hybrid IT network.”

Source: CloudStrategyMag

IBM Named A Leader In Gartner Magic Quadrant For DRaaS

IBM Named A Leader In Gartner Magic Quadrant For DRaaS

IBM has announced that Gartner, Inc. has positioned IBM as a Leader in the Gartner Magic Quadrant for Disaster Recovery as a Service (DRaaS) for the third consecutive year. And this year, IBM was once again positioned as one of the highest in ability to execute and the furthest in completeness of vision.

IBM continues to offer one of the most comprehensive sets of related professional and managed services in the industry, and it enhanced its long-standing history in disaster recovery with its 2016 purchase of Sanovi Technologies to complement its IT and applications resiliency orchestration capabilities. IBM also continues to integrate IBM Watson into its resiliency services offerings to enhance disaster avoidance.

“We believe IBM’s recognition in Gartner’s Magic Quadrant for Disaster Recovery as a Service report speaks to the value, consistency and reliability our clients find in our resiliency services across the globe, as we work to innovate and deliver for them every day,” said Laurence Guihard-Joly, general manager, IBM Resiliency Services. “Beyond natural disasters and other outages, the increasing number of cyberattacks and other threats today requires the most advanced and most proactive set of solutions possible so our clients can combat those threats and focus on differentiating in the era of ‘always-on’ expectations.”

IBM Resiliency Services offers an innovative portfolio of resiliency and business continuity solutions and services with expanded public cloud options to enable clients with greater flexibility and agility in managing their backup and disaster recovery workloads. Today, IBM operates over 300 global delivery data centers across 54 countries to help companies worldwide maintain continuous business operations and improve overall resiliency for any size organization.

Source: CloudStrategyMag

Databarracks Recognised In Gartner’s Magic Quadrant For DRaaS For 2017

Databarracks Recognised In Gartner’s Magic Quadrant For DRaaS For 2017

London-based provider Databarracks has been recognised in Gartner’s Magic Quadrant for Disaster Recovery as a Service 2017.

Peter Groucutt, managing director at Databarracks, commented on the announcement, “We are very proud to be included in the Magic Quadrant for DRaaS for a third consecutive year. We feel it is an incredible achievement for us as a specialist provider of disaster recovery and business continuity services and is a testament to the hard work and commitment of our team.

“There has been a great deal of change in the last year. We have seen some commoditisation in the market and this year’s MQ reflects that with many more service providers evaluated. I completely agree that DRaaS has now become a ‘mainstream offering’. Organisations now recognise that DRaaS is a mature service and the benefits extend beyond pure cost savings. The organisations we speak to appreciate the value a specialist DRaaS provider brings. Customers of DRaaS know that in the event of a disaster, they not only have the technology to allow for rapid recovery but the skills of recovery specialists to bring their IT back online, minimising data loss and downtime.

“In the last 12 months we have also seen further change in both the risk landscape and the attitude towards IT downtime. The recent WannaCry attack raised the level of awareness of ransomware beyond the IT team and into the wider business-consciousness both here in the UK and around the world. Major IT outages have also highlighted just how reliant organisations are on technology to deliver their services –— whether that is an airline, a bank or a university.

“Our focus has been to continue to develop the next generation of DRaaS services. That means research and development finding new methods to protect against growing cyber threats. It also means extending our services help our customers build out their Business Continuity Plans and be truly resilient businesses. We have some further exciting announcements to make on that front in the coming months.”

Source: CloudStrategyMag

Equinix Achieves Microsoft Gold Cloud Platform Status

Equinix Achieves Microsoft Gold Cloud Platform Status

Equinix, Inc. has announced that it has achieved Microsoft Gold Cloud Platform status via its Equinix Professional Services (EPS) organization. This designation reinforces Equinix’s ongoing commitment to delivering Microsoft hybrid cloud to its customers inside a global footprint of International Business Exchange™ (IBX®) data centers. Since Equinix began offering direct access to Microsoft Azure in 2014, it has seen significant growth in the amount of enterprise customers developing hybrid and multi-cloud deployments across five continents including North America, South America, Asia, Europe and Australia. Equinix currently offers direct connections to Microsoft Azure via the Equinix Cloud Exchange™ in 20 metros around the world.

Equinix’s Gold Competency status was awarded in recognition of its expertise in the areas hybrid cloud enablement through EPS. Hybrid cloud continues to be a top priority for enterprise CIOs. In fact, based on results from IDC’s 2017 CloudView Survey, nearly 70% of cloud users expect that within two years the majority of their IT capabilities will be delivered through some form of cloud service — public, private, or hybrid. Yet the path to the hybrid cloud can be technically challenging. Through EPS and Equinix’s ability to provide direct access to Microsoft Azure via the Equinix Cloud Exchange, Equinix helps its enterprise customers advance their cloud strategies and capitalize on these cloud benefits by seamlessly incorporating cloud services into their existing architectures.

“Hybrid cloud is revolutionizing the way enterprises do business, and Equinix is committed to empowering our customers to deploy high-performance, scalable and secure hybrid cloud solutions on a global scale. A critical component of this is interconnectivity to Microsoft Azure and Office 365 platforms inside our vast platform of data centers. We are honored to have reached Gold partner status with Microsoft, and look forward to providing additional ways for our global customers to achieve improved performance of their cloud-based applications,” said Bob Breynaert, global managing director, Microsoft strategic alliance, Equinix.

Microsoft Gold certification, provided to a select group of partners who demonstrate that they meet strict criteria, is the highest level of certification for companies that are part of the Microsoft partnership program. Companies, such as Equinix, that attain this status must provide successful client references and demonstrate that they have developed innovative products and solutions based on Microsoft technologies. Equinix has been aligned with Microsoft for many years to provide high-level interconnection and global access to Microsoft Azure. Equinix will also be expanding its role as a critical partner in Microsoft hybrid cloud solutions by integrating Microsoft Azure Stack into its interconnection offerings.

“We are incredibly proud to have achieved Gold Cloud Platform status. The recipe for our accomplishment ultimately stems from the right mix of a number of symbiotic ingredients: Equinix’s hybrid cloud enabling interconnection platform with unprecedented global reach, Microsoft’s reputable cloud service offerings, the exceptional quality of work delivered by our Professional Services organization, and last but not least, our extremely innovative customers. Teaming with a pioneering organization like Microsoft is instrumental in being able to deliver end-to-end solutions that quickly deliver business results and satisfy the evolving needs of our customers,” said Ryan Mallory, vice president, global solutions enablement, Equinix.

With Equinix, applications and workloads operating in a hybrid cloud are optimized for business value. By directly and securely connecting to Microsoft Azure and Microsoft Azure Stack inside Equinix, customers can take advantage of:

  • Superior application performance and user experience
  • Enhanced compliance by connecting privately to Microsoft Azure and keeping data within region, without going over the public internet
  • Increased cloud security, by bypassing the public internet

EPS offers practical guidance and proven solutions to help enterprise customers optimize and future-proof their data center architecture and cloud journey. EPS provides hands-on experience in assessing, enabling, migrating, optimizing, planning, designing, testing and deploying IT infrastructure, networks and cloud architectures.

*Source: IDC FutureScape: Worldwide Cloud 2017 Predictions; Doc # US41863916, Nov. 2016

Source: CloudStrategyMag

Rackspace Named A Leader In Hosted Private Cloud Services

Rackspace Named A Leader In Hosted Private Cloud Services

Rackspace has announced that it was named a leader in the Forrester Research, Inc. June 2017 report, The Forrester Wave™: Hosted Private Cloud Services, North America, Q2 2017. The report evaluated nine vendors based on 33 criteria, including current offering, strategy and market presence. Rackspace was ranked above all vendors in strategy and automation capabilities for its Rackspace OpenStack Private Cloud offering.

According to the report, “Rackspace enjoys a highly visible market position by virtue of its lengthy history in the managed hosting segment, OpenStack contributions and close partnerships with the leading hyperscale public cloud providers … Although its current offering is strong, Rackspace stands out by virtue of its customer-centric, ‘Fanatical Support’ strategy as well as its coexistence with hyperscale providers through management and migration services.” Additionally, customer references contacted during the evaluation called out Rackspace for “its superior self-service control options relative to other evaluated solutions, describing it as best-in-class.”

Rackspace scored highest among vendors in the strategy category for its Rackspace® OpenStack® Private Cloud, which Rackspace believes indicates the strength of its go-to-market strategy and future vision for hosted private cloud services. The independent research firm evaluates strategy using eight criteria, and Rackspace received the highest score possible in three of these criteria including: business technology vision, market approach and partner ecosystem.

In addition, Rackspace received the highest score in the automation capabilities criterion within the current offering category, and was among the vendors that scored the highest for self-service portal and user interface, integration strength, service level agreements (SLAs), storage and contract agreements.

“I believe that Rackspace’s position in this report further validates that it is the clear market leader in the OpenStack private cloud space,” said Scott Crenshaw, SVP and GM of OpenStack and Microsoft clouds at Rackspace. “In my view, our positioning is a testament to our strategy to deliver OpenStack private clouds as a service, providing self-service tools and an industry-leading 99.99% API uptime SLA. As a co-founder of OpenStack, Rackspace is well positioned to enable customers of all sizes to succeed through operational expertise gained from more than one billion server hours managing OpenStack clouds at scale. We are proud to be named a leader in this report and will continue to innovate and lead by providing industry-leading SLAs, world-class economics and a revolutionary, utility-based private cloud consumption model.”

Rackspace delivers a broad portfolio of managed private cloud solutions that support multiple platforms including OpenStack, VMware® Private Cloud, and Microsoft® Private Cloud. These private cloud solutions give customers the agility, scalability and efficiency of the public cloud, with the greater levels of control and security of a single-tenant, dedicated environment.

Source: CloudStrategyMag

DataCore Chooses JULABO For Software-Defined Storage

DataCore Chooses JULABO For Software-Defined Storage

DataCore Software has announced that SANsymphony™ was chosen by JULABO GmbH as the software-defined storage platform best suited to support its flexible virtual infrastructure. JULABO, a Germany-based worldwide provider of high-quality temperature control solutions, chose SANsymphony to ensure business continuity and high performance of VMware vSphere virtual machines synchronized across German and U.S. data centers. The VMs support Microsoft Dynamics, Microsoft SQL Server and Matrix42 software and workspace management applications. A JULABO data center in Allentown, Pennsylvania mirrors the German headquarters systems located in the Black Forest using DataCore replication software and Riverbed WAN technology.

“The complete infrastructure project was implemented and supports the network, servers and storage using DataCore SANsymphony cost-effectively. The entire project was carried out at the same cost as what would we would have paid for storage alone if we had chosen a major storage hardware provider,” said Jürgen Jonescheit, CIO JULABO GmbH.

JULABO GmbH is one of the foremost providers of high-quality, reliable and high-performance “Made in Germany” temperature control solutions for high-tech industries. With 10 branch offices on three continents and more than 400 employees, the central IT team provides support for Microsoft Dynamics ERP and other business-critical applications.

JULABO was faced with the need to move data between its remote sites and to headquarters, while also supporting rapid growth and new business requirements in its local markets. Growing workloads caused increased demands for high availability and greater flexibility, prompting it to implement a comprehensive virtual infrastructure across the entire organization. Today JULABO maintains mirrored computer centers in two different sites, protected with various fire compartments, as well as a third computer center located to support off-site storage of backups. All of the sites are connected via a high-performance fibre channel network.

Parallel I/O, Fusion-io and High-Speed Caching Drive 5x Performance Increase

DataCore partner and system integrator Leitwerk AG was commissioned to implement VMware vSphere and DataCore SANsymphony on the main site. Today all systems are virtualized, with workloads running across 170 virtual machines on redundantly designed VMware ESX servers on HP Proliant hardware, each with 768 GB RAM and 16 cores.

SANsymphony also runs on an HP Proliant and virtualizes 50TB of storage capacity (SAS, SATA, SSD/flash) while providing synchronized highly available data to applications. DataCore provides auto failover protection and high availability to ensure accessibility to the two HP D2700 disk shelves. In addition to DataCore’s high-speed caching, Fusion-io PCIe flash memory cards were integrated within the DataCore software-defined storage nodes, providing cost-effective performance by using DataCore auto-tiering to manage and optimize performance and resource utilization.

DataCore’s Parallel I/O technology leverages multi-cores to improve I/O performance and response times needed by demanding applications by processing I/O requests in parallel. JULABO estimates that, in comparison with hardware-bound architecture, DataCore has resulted in a five-fold increase in performance. Up to 40,000 IOPS with 64K blocks were achieved during tests under realistic real-world conditions.

“Hardware solutions were not able to achieve the DataCore performance results and value, even with twice the number of spindles,” Jonescheit added.

“We chose DataCore because of its excellent price/performance ratio and its outstanding technical design. Good examples are the automatic failover for high availability and auto-tiering for performance. We have put this functionality into practice successfully. However, the flexibility resulting from hardware independence does not only have technical and functional advantages. It also allows us to flexibly extend our infrastructure cost-effectively to maximize our economic efficiency, as we are now vendor-independent,” Jonescheit concluded.

Benefits at a glance:

• DataCore reduces storage-related costs by 50%.

• Performance has improved 5-fold.

• Storage-related failures were reduced by 90%.

• Time spent on routine administrative tasks was reduced by 90%.

• Planned downtime (data migrations, upgrades, updates) was reduced by over 95%.

• DataCore managed to convert 50% of unused storage space into available free storage capacity.

Source: CloudStrategyMag

Cloud Business Communications Growing By 23% Per Year

Cloud Business Communications Growing By 23% Per Year

New Q1 data from Synergy Research Group shows that the cloud business communications market is growing at an annualized rate of 23%, as strong UCaaS adoption continues to drive the market. Retail UCaaS services account for well over 70% of the total market, though the smaller wholesale UCaaS segment is growing considerably more strongly. The balance of the market comprises cloud communications (virtualized call control and UC applications), which are growing at a similar rate to retail UCaaS. The total cloud communication market is now generating revenues in excess of $2 billion per year. RingCentral is the market leader based on quarterly revenues and has been consistently growing its market share, thanks to leadership of retail UCaaS and a strong position in wholesale UCaaS. It is followed by 8×8, Mitel and ShoreTel.

Retail and wholesale UCaaS business suite services now generate revenue of well over $1.7 billion on an annualized basis. These include private, public and hybrid cloud service offerings. Public UCaaS is the largest of the three sub-segments and is also growing the most rapidly. Overall the cloud business communications market continues to be heavily concentrated in the United States, which accounted for 80% of total revenues in the first quarter of the year.

“UCaaS continues to be a force for change within the business communications market,” said Jeremy Duke, Synergy Research Group’s founder and chief analyst. “Major barriers to cloud adoption are now almost a thing of the past and consequently we are seeing continued erosion of on-premises PBX-based systems. Clearly these trends will continue over the next few years.”

Source: CloudStrategyMag

IBM Launches New Service For Hybrid Cloud Networks

IBM Launches New Service For Hybrid Cloud Networks

IBM has launched a new offering designed to help companies more efficiently manage applications and services across a hybrid cloud environment by gaining real-time visual insights on the performance of supporting infrastructure.

Agile Service Manager for IBM Netcool Operations Insight will help companies in industries rapidly moving applications to the cloud — such as telecommunications and financial services — better monitor changes in their underlying systems so they can avoid customer service disruption. It’s among the only offerings in the IT industry to provide a detailed topology of infrastructure across a broad set of domains, including different public clouds and on-premises systems, with both real-time updates and historical views.

By having this view of virtual and physical resources, companies can now better manage applications and services deployed across highly distributed and increasingly complex hybrid cloud networks. Agile Service Manager also can help companies prevent and more quickly resolve problems by providing context about the resources that underpin a hybrid cloud network. This is especially important for companies where applications put significant demands on network resources while requiring continuous uptime to meet customer expectations.

For example, a communications service provider could use IBM Netcool Operations Insight with Agile Service Manager to support launching and scaling a new, cloud-based video conferencing service. As the company expands use of the conferencing service, Agile Service Manager could provide real-time insights about the infrastructure supporting the new service, such as response time of specific servers and capacity of storage, letting the company know if these assets are meeting customer needs.

“In today’s fast-changing marketplace, companies are under increasing pressure to roll out new and enhanced applications for customers,” said Denis Kennelly, general manager of IBM Hybrid Cloud. “Agile Service Manager gives valuable new insights about what is going on in the network and how it is impacting service quality and customers in real time.”

Companies can use Agile Service Manager to present a configurable topology view that shows the relationships and states of resources both in real time and within a defined time window. Agile Service Manager extends the capabilities of IBM Netcool Operations Insight, an analytics driven software that helps organizations gain actionable insights from the massive amounts of operational data generated from their hybrid cloud environments.

Source: CloudStrategyMag

Equinix Partners With Alibaba Cloud

Equinix Partners With Alibaba Cloud

Equinix, Inc. has announced a collaboration with Alibaba Cloud to provide enterprises with direct, scalable access to Alibaba Cloud via the Equinix Cloud Exchange™ in the company’s Hong Kong, Silicon Valley, Sydney, and Washington, D.C., International Business Exchange™ (IBX®) data centers, with the addition of Frankfurt and London available soon.

The cloud computing market continues to grow rapidly in Asia, and specifically China. According to the U.S. International Trade Administration, the Chinese cloud market is expected to grow at 40% per year through 2020, reaching $20 billion. Access to Alibaba Cloud is critical for multinational customers looking to expand their cloud-based applications into this growing region in a secure and high-performing manner. By providing multi-national enterprise customers with secure and direct access to Alibaba Cloud, Equinix delivers connectivity to the full suite of Alibaba cloud services to enterprises. Furthermore, Alibaba Cloud Express Connect enables companies to leverage the extensive Alibaba Cloud network in mainland China. 

“As one of the world’s leading and fastest growing cloud computing companies, Alibaba Cloud represents a significant partnership for Equinix as we continue to empower businesses around the globe to build secure, private clouds, without compromising network and application performance. We are thrilled to offer direct access in markets across the U.S., Asia and Europe, and look forward to bringing additional markets online in the near future,” said Greg Adgate, vice president of global technology partners and alliances, Equinix.

With the addition of direct access to Alibaba Cloud on Cloud Exchange in these new markets across Asia Pacific, EMEA and the Americas, Equinix now offers private access to Alibaba Cloud in five markets. The company previously offered access in its Singapore IBX. Alibaba Cloud is also a colocation customer in Dubai with Emirates Integrated Telecommunications Company PJSC (“du”).

“The global reach of Equinix Cloud Exchange makes it simple for Alibaba Cloud to access new markets. We are pleased to provide greater value and bring our services closer to enterprises by leveraging Equinix’s powerful, on-demand cloud connectivity, and in particular to provide greater connectivity to the Chinese market,” said Yeming Wang, deputy general manager of Alibaba Cloud Global.


Source: CloudStrategyMag