IBM Resurrects Moore’s Law

Guess Moore’s Law ain’t as dead as we were led to believe. On Jun 5 IBM and Research Alliance partners GLOBALFOUNDRIES and Samsung, along with equipment suppliers announced the development of an industry-first process to build silicon nano sheet transistors that will enable 5nm chips. Previously, IBM announced a 7nm process using a silicon germanium (SiGe) alloy.

As DancingDinosaur wrote in early Oct. 2015, the last z System that conformed to the expectations of Moore’s Law was the zEC12, introduced Aug 2012. IBM could boast then it had the fastest commercial processor available.  The subsequent z13 didn’t match it in processor speed.  The z13 chip runs a 22 nm core at 5 GHz, one-half a GHz slower than the zEC12, which ran its 32nm core at 5.5 GHz. IBM compensated for the slower chip speed by adding more processors throughout the system to boost I/O and other functions and optimizing the box every way possible.

5nm silicon nano-sheet transistors delivers 40% performance gain

By 2015, the z13 delivered about a 10 percent performance bump per core thanks to the latest tweaks in the core design, such as better branch prediction and better pipelining. But even at one-half Ghz slower, the z13 was the first system to process 2.5 billion transactions a day.  Even more importantly for enterprise data centers, z13 transactions are persistent, protected, and auditable from end-to-end, adding assurance as mobile transactions grow to an estimated 40 trillion mobile transactions per day by 2025. The z13 also received and continues to receive praise for its industry leading security ratings as well as its scalability and flexibility.

Just recently Hitachi announced a partnership with IBM to develop a version of the z13 to run its own operating system, VOS3. The resulting z13 will run the next generation of Hitachi’s AP series.

But IBM isn’t back in pursuit of Moore’s Law just to deliver faster traditional mainframe workloads. Rather, the company is being driven by its strategic initiatives, mainly cognitive computing. As IBM explained in the announcement: The resulting increase in performance will help accelerate cognitive computing, the Internet of Things (IoT), and other data-intensive applications delivered in the cloud. The power savings could also mean that the batteries in smartphones and other mobile products could last two to three times longer than today’s devices, before needing to be charged.

Scientists working as part of the IBM-led Research Alliance at the SUNY Polytechnic Institute Colleges of Nanoscale Science and Engineering’s NanoTech Complex in Albany, NY achieved the breakthrough by using stacks of silicon nanosheets as the device structure of the transistor instead of the standard FinFET architecture, which is the blueprint for the semiconductor industry up through 7nm node technology. “For business and society to meet the demands of cognitive and cloud computing in the coming years, advancement in semiconductor technology is essential,” said Arvind Krishna, senior vice president, Hybrid Cloud, and director, IBM Research in the announcement. “That’s why IBM aggressively pursues new and different architectures and materials that push the limits of this industry, and brings them to market in technologies like mainframes and our cognitive systems.”

Compared to the leading edge 10nm technology available in the market, according to IBM, a nanosheet-based 5nm technology can deliver 40 percent performance enhancement at fixed power, or 75 percent power savings at matched performance. This improvement enables a significant boost to meeting the future demands of artificial intelligence (AI) systems, virtual reality, and mobile devices.

These may not sound like the workloads you are running on your mainframe now, but systems with these chips are not going to be shipped in the next mainframe either. So, you have a couple of years. The IBM team expects to make progress toward commercializing 7nm in 2018. By the time they start shipping 5nm systems you might be desperate for a machine to run such workloads and others like them.

DancingDinosaur is Alan Radding, a veteran information technology analyst, writer, and ghost-writer. Please follow DancingDinosaur on Twitter, @mainframeblog. See more of his IT writing at technologywriter.com and here.

 

IBM Puts Open DBaaS on IBM OpenPOWER LC Servers

Sometimes IBM seems to be thrashing around looking for anything hot that’s selling, and the various NoSQL databases definitely are hot. The interest is driven by DevOps, cloud, and demand for apps fast.

A month or so ago the company took its Power LC server platform to the OpenPOWER Developer Conference in San Francisco where they pitched Database-as-a-Service (DBaaS) and a price-performance guarantee: OpenPOWER LC servers designed specifically for Big Data to deliver a 2.0x price-performance advantage over x86 for MongoDB and 1.8x for EDB PostgreSQL 9.5 guaranteed. With organizations seeking any performance advantage, these gains matter.

There are enough caveats that IBM will almost never be called to deliver on the guarantee. So, don’t expect to cash in on this very quickly. As IBM says in the miles of fine print: the company will provide additional performance optimization and tuning services consistent with IBM Best Practices, at no charge.  But the guarantee sounds intriguing. If you try it, please let DancingDinosaur know how it works out.

IBM Power System S822LC for Big Data

BTW, IBM published the price for the S822LC for big data as starting at $6,399.00 USD. Price includes shipping. Linux OS, however, comes for an additional charge.

Surprisingly, IBM is not aiming this primarily to the IBM Cloud. Rather, the company is targeting the private cloud, the on-premises local version. Its Open DBaaS toolkit, according to IBM, provides enterprise clients with a turnkey private cloud solution that pre-integrates an Open Source DB image library, OpenStack-based private cloud, and DBaaS software packages with hardware (servers/storage/network switches/rack) and a single source of support to enable a DBaaS self-service portal for enterprise developers and LOB users to provision MongoDB, Postgres, and others in minutes. But since it is built on OpenStack, it also supports hybrid cloud integration with IBM Cloud offerings via OpenStack APIs.

In terms of cost it seems remarkably reasonable. It comes in four reference configurations. The Starter configuration is ~$80k (US list price) and includes 3 Power 822LC servers, pair of network switches, rack, DBaaS Toolkit software, and IBM Lab Services. Other configurations include Entry, Cloud Scale, and Performance configurations that have been specified for additional compute, storage, and OpenStack control plane nodes along with high-capacity JBOD storage drawers. To make this even easier, each configuration can be customized to meet user requirements. Organizations also can provide their own racks and/or network switches.

Furthermore, the Power 822LC and Power 821LC form the key building blocks for the compute, storage and OpenStack control plane nodes. As a bonus, however, IBM includes the new 11-core Power 822LC, which provides an additional 10-15% performance boost over the 10-core Power 822LC for the same price.

This is a package deal, at least if you want the best price and to deploy it fast. “As the need for new applications to be delivered faster than ever increases in a digital world, developers are turning to modern software development models including DevOps, as-a-Service, and self-service to increase the volume, velocity and variety of business applications,” said Terri Virnig, VP, Power Ecosystem and Strategy at IBM. Open Platform for DBaaS on IBM in the announcement. Power Systems DBaaS package  includes:

• A self-service portal for end users to deploy their choice of the most popular open source community databases including MongoDB, PostgreSQL, MySQL, MariaDB, Redis, Neo4j and Apache Cassandra deployable in minutes
• An elastic cloud infrastructure for a highly scalable, automated, economical, and reliable open platform for on-premises, private cloud delivery of DBaaS
• A disk image builder tool for organizations that want to build and deploy their own custom databases to the database image library

An open source, cloud-oriented operations manager with dashboards and tools will help you visualize, control, monitor, and analyze the physical and virtual resources. A turnkey, engineered solution comprised of compute, block and archive storage servers, JBOD disk drawers, OpenStack control plane nodes, and network switches pre-integrated with the open source DBaaS toolkit is available through GitHub here.

DancingDinosaur is Alan Radding, a veteran information technology analyst, writer, and ghost-writer. Please follow DancingDinosaur on Twitter, @mainframeblog. See more of his IT writing at technologywriter.com and here.

 

Syncsort Drives zSystem and Distributed Data Integration

IBM appears to be so busy pursuing its strategic imperatives—security, blockchain, quantum computing, and cognitive computing—that it seems to have forgotten the daily activities that make up the bread-and-butter of mainframe data centers. Stepping up to fill the gap have been mainframe ISVs like Compuware, Syncsort, Data Kinetics, and a few others.

IBM’s Project DataWorks taps into unstructured data often missed

IBM hasn’t completely ignored this need. For instance, Project DataWorks uses Watson Analytics and natural language processing to analyze and create complex visualizations. Syncsort, on the other hand, latched onto open Apache technologies, starting in the fall of 2015. Back then it introduced a set of tools to facilitate data integration through Apache Kafka and Apache Spark, two of the most active Big Data open source projects for handling real-time, large-scale data processing, feeds, and analytics.

Syncsort’s primary integration vehicle then revolved around the Intelligent Execution capabilities of its DMX data integration product suite with Apache Spark. Intelligent Execution allows users to visually design data transformations once and then run them anywhere – across Hadoop, MapReduce, Spark, Linux, Windows, or Unix, both on premise or in the cloud.

Since then Syncsort, in March, announced another big data integration solution. This time its DMX-h, is now integrated with Cloudera Director, enabling organizations to easily deploy DMX-h along with Cloudera Enterprise on Amazon Web Services, Microsoft Azure, or Google Cloud. By deploying DMX-h with CDH, Syncsort explained, organizations can quickly pull data into new, ready-to-work clusters in the cloud. This accelerates how quickly they can take advantage of big data cloud benefits, including cost savings and Data-as-a-Service (DaaS) delivery.

A month before that, this past February, Syncsort introduced new enhancements in its Big Data integration solution by again deploying DMX-h to deliver integrated workflow capabilities and Spark 2.0 integration, which simplifies Hadoop and Spark application development, effectively enabling mainframe data centers to extract maximum value from their data assets.

In addition, Syncsort brought new integrated workflow capabilities and Spark 2.0 integration to simplify Hadoop and Spark application development. It lets data centers tap value from their enterprise data assets regardless of where it resides, whether on the mainframe, in distributed systems, or in the cloud.

Syncsort’s new integrated workflow capability also gives organizations a simpler, more flexible way to create and manage their data pipelines. This is done through the company’s design-once, deploy-anywhere architecture with support for Apache Spark 2.0, which makes it easy for organizations to take advantage of the benefits of Spark 2.0 and integrated workflow without spending time and resources redeveloping their jobs.

Assembling such an end-to-end data pipeline can be time-consuming and complicated, with various workloads executed on multiple platforms, all of which need to be orchestrated and kept up to date. Delays in such complicated development, however, can prevent organizations from getting the timely insights they need for effective decision-making.

Enter Syncsort’s Integrated Workflow, which helps organizations manage various workloads, such as batch ETL on large repositories of historical data. This can be done by referencing business rules during data ingest in a single workflow, in effect simplifying and speeding development of the entire data pipeline, from accessing critical enterprise data, to transforming that data, and ultimately analyzing it for business insights.

Finally, in October 2016 Syncsort announced new capabilities in its Ironstream software that allows organizations to access and integrate mainframe log data in real-time to Splunk IT Service Intelligence (ITSI). Further, the integration of Ironstream and Compuware’s Application Audit software deliver the audit data to Splunk Enterprise Security (ES) for Security Information and Event Management (SIEM). This integration improves an organization’s ability to detect threats against critical mainframe data, correlate them with related information and events, and satisfy compliance requirements.

DancingDinosaur is Alan Radding, a veteran information technology analyst, writer, and ghost-writer. Please follow DancingDinosaur on Twitter, @mainframeblog. See more of his IT writing at technologywriter.com and here.

 

IBM Introduces Hitachi-Specific z13

Remember when rumors were flying that Hitachi planned to buy the mainframe z Systems business from IBM?  DancingDinosaur didn’t believe it at that time, and now we have an official announcement that IBM is working with Hitachi to deliver mainframe z System hardware for use with Hitachi customers.

Inside the IBM z13

DancingDinosaur couldn’t see Hitachi buying the z. The overhead would be too great. IBM has been sinking hundreds of millions of dollars into the z, adding new capabilities ranging from Hadoop and Spark natively on z to whatever comes out of the Open Mainframe Project.

The new Hitachi deal takes the z in a completely different direction. The plans calls for using Hitachi’s operating system, VOS3, running on the latest IBM z13 hardware to provide Hitachi users with better performance while sustaining their previous investments in business-critical Hitachi data and software, as IBM noted. VOS3 started as a fork of MVS and has been repeatedly modified since.

According to IBM, Hitachi will exclusively adopt the IBM z Systems high-performance mainframe hardware technology as the only hardware for the next generation of Hitachi’s AP series. These systems primarily serve major organizations in Japan. This work expands Hitachi’s cooperation with IBM to make mainframe development more efficient through IBM’s global capabilities in developing and manufacturing mainframe systems. The Open Mainframe Project, BTW, is a Linux initiative.

The collaboration, noted IBM, reinforces its commitment to delivering new innovations in mainframe technology and fostering an open ecosystem for the mainframe to support a broad range of software and applications. IBM recently launched offerings for IBM z Systems that use the platform’s capabilities for speed, scale and security to deliver cloud-based blockchain services for building new transaction systems and machine learning for analyzing large amounts of data.

If you count VOS3, the mainframe now runs a variety of operating systems, including z/OS, z/TPF and z/VM operating systems as well as the Linux. Reportedly, Hitachi plans to integrate its new mainframe with its Lumada Internet of Things (IoT) offerings. With z scalability, security, massive I/O, and performance the z makes an ideal IoT platform, and IoT is a market IBM targets today. Now IBM is seeding a competitor with the z running whatever appealing capabilities Hitachi’s Lumada offers. Hope whatever revenue or royalties IBM gets is worth it.

IBM and Hitachi, as explained in the announcement, have a long history of cooperation and collaboration in enterprise computing technologies. Hitachi decided to expand this cooperation at this time to utilize IBM’s most advanced mainframe technologies. Hitachi will continue to provide its customers with a highly reliable, high-performance mainframe environment built around the Hitachi VOS3 operating system. Hitachi also continues to strengthen mainframe functionality and services which contributes to lower TCO, improved ease of system introduction and operation, and better serviceability.

Of course, the mainframe story is far from over. IBM has been hinting at a new mainframe coming later this year for months.  Since IBM stopped just automatically cranking up core processor speed to boost price/performance it will employ an array of assist processors and software optimizations to boost performance wherever it can, but particularly in the area of its current critical imperatives—security, cognitive computing, blockchain, and cloud. One thing DancingDinosaur doesn’t expect to see in the new z, however, will be cubits embedded, but who knows?

DancingDinosaur is Alan Radding, a veteran information technology analyst, writer, and ghost-writer. Please follow DancingDinosaur on Twitter, @mainframeblog. See more of his IT writing at technologywriter.com and here.

 

IBM Insists Storage is Generating Positive Revenue

At a recent quarterly briefing on the company’s storage business, IBM managers crowed over its success: 2,000 new Spectrum Storage customers, 1,300 new DS8880 systems shipped, 1500 PB of capacity shipped, 7% revenue gain Q1’17. This appeared to contradict yet another consecutive losing quarter in which only IBM’s Cognitive Solutions (includes Solutions Software and Transaction Processing Software) posted positive revenue.

However, Martin Schroeter, Senior Vice President and Chief Financial Officer (1Q’17 financials here), sounded upbeat about IBM storage in the quarterly statement: Storage hardware was up seven percent this quarter, led by double-digit growth in our all-flash array offerings. Flash contributed to our Storage revenue growth in both midrange and high-end. In storage, we continue to see the shift in value towards software-defined environments, where we continue to lead the market. We again had double-digit revenue growth in Software-Defined Storage, which is not reported in our Systems segment. Storage software now represents more than 40 percent of our total storage revenue.

IBM Flash System A9000

Highly parallel all-flash storage for hyperscale and cloud data centers

Schroeter continued: Storage gross margins are down, as hardware continues to be impacted by price pressure. To summarize Systems, our revenue and gross profit performance were driven by expected cycle declines in z Systems and Power, mitigated by Storage revenue growth. We continue to expand our footprint and add new capabilities, which address changing workloads. While we are facing some shifting market dynamics and ongoing product transitions, our portfolio remains uniquely optimized for cognitive and cloud computing.

DancingDinosaur hopes he is right.  IBM has been signaling a new z System coming for months, along with enhancements to Power storage. Just two weeks ago IBM reported achievements with Power and Nvidia, as DancingDinosaur covered at that time.

If there was any doubt, all-flash storage is the way IBM and most other storage providers are heading for the performance and competitive economics. In January IBM announced three all flash DS888* all flash products, which DancingDinosaur covered at the time here. Specifically:

• DS8884 F (the F designates all flash)—described by IBM as performance delivered within a flexible and space-saving package
• DS8886 F—combines performance, capacity, and cost to support a variety of workloads and applications
• DS8888 F—promises performance and capacity designed to address the most demanding business workload requirements

The three products are intended to provide the speed and reliability needed for workloads ranging from enterprise resource planning (ERP) and financial transactions to cognitive applications like machine learning and natural language processing. Doubt that a lot of mainframe data centers are doing much with cognitive systems yet, but that will be coming.

Spectrum Storage also appears to be looming large in IBM’s storage plans. Spectrum Storage is IBM’s software defined storage (SDS) family of products. DancingDinosaur covered the latest refresh of the suite of products this past February.

The highlights of the recent announcement included the addition of Cloud Object Storage and a version of Spectrum Virtualize as software only.  Spectrum Control got a slew of enhancements, including new cloud-based storage analytics for Dell EMC VNX, VNXe, and VMAX; extended capacity planning views for external storage, and transparent cloud tiering for IBM Spectrum Scale.  The on-premises editions added consolidated chargeback/showback and support for Dell EMC VNXe file storage. This should make it clear that Spectrum Storage is not only for underlying IBM storage products.

Along the same lines, Spectrum Storage added VMware 6 support and the certified vSphere Web client. In the area of cloud object storage, IBM added native NFS access, enhance STaaS multi-tenancy, IPV6 support, and preconfigured bundles.

IBM also previewed enhancements coming in 2Q’17.   Of specific interest to DancingDinosaur readers will likely be  the likely updates to the FlashSystem and VeraStack portfolio.

The company is counting on these enhancements and more to help pull IBM out of its tailspin. As Schroeter wrote in the 1Q’17 report: New systems product introductions later in the year will drive improved second half performance as compared to the first. Hope so; already big investors are cashing out. Clients, however, appear to be staying for now.

DancingDinosaur is Alan Radding, a veteran information technology analyst, writer, and ghost-writer. Please follow DancingDinosaur on Twitter, @mainframeblog. See more of his IT writing at technologywriter.com and here.

 

IBM Demonstrates Quantum Computing Advantage

In an announcement last week, IBM reported that scientists from IBM Research and Raytheon BBN have demonstrated one of the first proven examples of a quantum computer’s advantage over a conventional computer. By probing a black box containing an unknown string of bits, they showed that just a few superconducting qubits can discover the hidden string faster and more efficiently than today’s computers. Their research was published in a paper titled, “Demonstration of quantum advantage in machine learning” in nature.com.

With IBM’s current 5 qubit processor, the quantum algorithm consistently identified the sequence in up to 100x fewer computational steps and was more tolerant of noise than the conventional (non-quantum) algorithm. This is much larger than any previous head-to-head comparison between quantum and conventional processors.

Courtesy: IBM Research

The graphic above defines 3 types of quantum computers. At the top is the quantum annealer, described as the least powerful and most restrictive.  In the middle sits analog quantum, 50-100 qubits, a device able to simulate complex quantum interactions. This will probably be IBM’s next quantum machine; currently IBM offers a 5 qubit device. At the bottom sits the universal quantum. IBM suggests this will scale to over 100,000 qubits and be capable of handling machine learning, quantum chemistry, optimization problems, secure computing, and more. It will be exponentially faster than traditional computers and be able to handle just about all the things even the most powerful conventional supercomputers cannot do now.

The most powerful z System, regardless of how many cores or accelerators or memory or bandwidth, remains a traditional, conventional computer. It deals with problems as a series of basic bits, sequences of 0 or 1. That it runs through these sequences astoundingly fast fools us into thinking that there is something beyond the same old digital computing we have known for the last 50 years or more.

Digital computers see the world and the problems you trying to solve as sequences of 0 and 1. That’s it; there is nothing in-between. They store numbers as sequences of 0 and 1 in memory, and they process stored numbers using only the simplest mathematical operations, add and subtract. As a college student DancingDinosaur was given the most powerful TI programmable calculator then available and, with a few buddies, we tried to come up with things it couldn’t do. No matter how many beer-inspired tries, we never found something it couldn’t handle.  The TI was just another digital device.

Quantum computers can digest 0 and 1 but have a broader array of tricks. For example, contradictory things can exist concurrently. Quantum geeks often cite a riddle dubbed Schrödinger’s cat. In this riddle the cat can be alive and dead at the same time because quantum system can handle multiple, contradictory states. If we had known of Schrödinger’s cat my buddies and I might have stumped that TI calculator.

In an article on supercomputing in Explain That Stuff by Chris Woodford he shows the thinking behind Schrödinger’s cat, called superposition.  This is where two waves, representing a live cat and a dead one, combine to make a third that contains both cats or maybe hundreds of cats. The wave inside the pipe contains all these waves simultaneously: they’re added to make a combined wave that includes them all. Qubits use superposition to represent multiple states (multiple numeric values) simultaneously.

In its latest quantum achievement IBM with only a 5 cubit the quantum algorithm consistently identified the sequence in up to a 100x fewer computational steps and was more tolerant of noise than the conventional (non-quantum) algorithm. This is much larger than any previous head-to-head comparison between quantum and conventional processors.

In effect, the IBM-Raytheon team programmed a black box such that, with the push of a button, it produces a string of bits with a hidden a pattern (such as 0010) for both a conventional computation and a quantum computation. The conventional computer examines the bits one by one. Each result gives a little information about the hidden string. By forcing the conventional computer to query the black box many times it can determine the full answer.

The quantum computer employs a quantum algorithm that extracts the information hidden in the quantum phase — information to which a conventional algorithm is completely blind. The bits are then measured as usual and, in about half the time, the hidden string can be fully revealed.

Most z data centers can’t use quantum capabilities for their daily work, at least not yet. As Woodford noted: It’s very early for the whole field—and most researchers agree that we’re unlikely to see practical quantum computers appearing for many years—perhaps even decades. Don’t bet on it; at the rate IBM is driving this, you’ll probably see useful things much sooner. Maybe tomorrow.

DancingDinosaur is Alan Radding, a veteran information technology analyst, writer, and ghost-writer. Please follow DancingDinosaur on Twitter, @mainframeblog. See more of his IT writing at technologywriter.com and here.

 

IBM Shows Off POWER and NVIDIA GPU Setting High Performance Record

The record achievement used 60 Power processors and 120 GPU accelerators to shatter the previous supercomputer record, which used over a 700,000 processors. The results point to how dramatically the capabilities of high performance computing (HPC) has increase while the cost of HPC systems has declined. Or put another way: the effort demonstrates the ability of NVIDIA GPUs to simulate one billion cell models in a fraction of the time, while delivering 10x the performance and efficiency.

Courtesy of IBM: Takes a lot of processing to take you into a tornado

In short, the combined success of IBM and NVIDIA puts the power of cognitive computing within the reach of mainstream enterprise data centers. Specifically the project performed reservoir modeling to predict the flow of oil, water, and natural gas in the subsurface of the earth before they attempt to extract the maximum oil in the most efficient way. The effort, in this case, involved a billion-cell simulation, which took just 92 minutes using 30 for HPC servers equipped with 60 POWER processors and 120 NVIDIA Tesla P100 GPU accelerators.

“This calculation is a very salient demonstration of the computational capability and density of solution that GPUs offer. That speed lets reservoir engineers run more models and ‘what-if’ scenarios than previously,” according to Vincent Natoli, President of Stone Ridge Technology, as quoted in the IBM announcement. “By increasing compute performance and efficiency by more than an order of magnitude, we’re democratizing HPC for the reservoir simulation community,” he added.

“The milestone calculation illuminates the advantages of the IBM POWER architecture for data-intensive and cognitive workloads.” said Sumit Gupta, IBM Vice President, High Performance Computing, AI & Analytics in the IBM announcement. “By running Stone Ridge’s ECHELON on IBM Power Systems, users can achieve faster run-times using a fraction of the hardware.” Gupta continued. The previous record used more than 700,000 processors in a supercomputer installation that occupies nearly half a football field while Stone Ridge did this calculation on two racks of IBM Power Systems that could fit in the space of half a ping-pong table.”

This latest advance challenges perceived misconceptions that GPUs could not be efficient on complex application codes like reservoir simulation and are better suited to simple, more naturally parallel applications such as seismic imaging. The scale, speed, and efficiency of the reported result disprove this misconception. The milestone calculation with a relatively small server infrastructure enables small and medium-size oil and energy companies to take advantage of computer-based reservoir modeling and optimize production from their asset portfolio.

Billion cell simulations in the industry are rare in practice, but the calculation was accomplished to highlight the performance differences between new fully GPU-based codes like the ECHELON reservoir simulator and equivalent legacy CPU codes. ECHELON scales from the cluster to the workstation and while it can simulate a billion cells on 30 servers, it can also run smaller models on a single server or even on a single NVIDIA P100 board in a desktop workstation, the latter two use cases being more in the sweet spot for the energy industry, according to IBM.

As importantly, the company notes, this latest breakthrough showcases the ability of IBM Power Systems with NVIDIA GPUs to achieve similar performance leaps in other fields such as computational fluid dynamics, structural mechanics, climate modeling, and others that are widely used throughout the manufacturing and scientific community. By taking advantage of POWER and GPUs organizations can literally do more with less, which often is an executive’s impossible demand.

DancingDinosaur is Alan Radding, a veteran information technology analyst, writer, and ghost-writer. Please follow DancingDinosaur on Twitter, @mainframeblog. See more of his IT writing at technologywriter.com and here.

 

No letup by IBM on Blockchain

IBM continues to push blockchain. Its latest announcement on the subject, Three Blockchain Adoption Principles Essential for Every CEO, came early this week. The basic pitch: in certain market segments blockchain could potentially help save billions of dollars annually and significantly reduce delays and spoilage. Citing the World Economic Forum, the company adds: “reducing barriers within the international supply chain could increase worldwide GDP by almost five percent and total trade volume by 15 percent.”  That should be sweet music to any C-suite exec.

Blockchain enables transparent food chain

In a related announcement also this week, IBM Japan, Mizuho Financial Group, and Mizuho Bank are building a blockchain-based trade financing platform for trade financing. With the platform, Mizuho is aiming to streamline trading operations and improve supply chain efficiency. The resulting timely and highly secure exchange of trade documents turns out to be essential for stakeholders in the supply chain ecosystem. Digitizing trade information on a blockchain can help alter the way information is shared, infusing greater trust into transactions, making it easier for parties involved in the supply chain, including exporters, importers, shippers, insurance companies, port operators, and port authorities to share critical shipment data in near real-time.

IBM is emerging as a leader in secure open-source blockchain solutions built for the enterprise. An early member of the Linux Foundation’s Hyperledger Project, the company has worked with more than 400 clients across multiple business segments to implement blockchain applications delivered via the IBM Cloud.

DancingDinosaur has its own 3 reasons enterprise data center execs should be excited by blockchain. They are different and more z-centric than IBM’s. First, you probably already have a z System, and the z’s legendary security, availability, and scalability make it a natural for blockchain. Second, the z already comes optimized to handle transactions and most of your transaction data already lives on the z, making it very efficient from a processing standpoint.  Third, until or unless your blockchain grows to some enormous size, it will barely consume any system resources or generate overhead. In that sense, blockchain on your z comes virtually free.

The following blockchain principles are based on IBM’s customer experience:

1. Blockchain has the potential to transform trade, transactions and business processes: The two concepts underpinning blockchain are “business network” and “ledger.” Taken together, these are what make blockchain a smart, tamper-resistant way to conduct trade, transactions and business processes. Network members exchange assets through a ledger that all members can access and share. The ledger syncs across the network with all members needing to confirm a transaction of tangible or intangible assets before it is approved and stored on the blockchain. This shared view helps establish legitimacy and transparency, even when parties are not familiar with one another.
2. The value, it turns out, resides in the ecosystem as the blockchain network grows: This should be no surprise to an exec who saw the growth, first of LANs and WANs, and later the Internet and Web. So too, as a business network blockchain can include several different types of participants. Depending on the number of participants in a blockchain network, the value of assets being exchanged, and the need to authorize members with varying credentials adopters should observe the difference between “permissioned” and “permission-less” blockchain networks. The real value for blockchain is achieved when these business networks grow. With a strong ecosystem, blockchain networks can more easily reach critical mass, allowing the users to build new business models and reinvent and automate transaction processes.
3. Blockchain can significantly improve visibility and trust across business: Block chains can reduce transaction settlement times from days or weeks to seconds by providing immediate visibility to all participants. The technology can also be used to cut excess costs by removing intermediary third-parties, those typically required to verify transactions. Because blockchain is built on the concept of trust, it can help reduce risks of illicit practices carried out over payment networks, helping to mitigate fraud and cybercrimes. Speed, cost efficiency, and transparency are among blockchain’s most significant benefits in the enterprise and within ecosystems of companies conducting trade. IBM, Walmart and Tsinghua University, for example, are exploring the use of blockchain to help address the challenges surrounding food safety [see graphic above]. By allowing members within the supply chain to see the same records, blockchain helps narrow down the source of a contamination

“Critical success factors in blockchain engagements require top-down executive support for innovative use cases and bringing key network participants into the dialogue from the start,” according to Marie Wieck, general manager, IBM Blockchain.

DancingDinosaur is Alan Radding, a veteran information technology analyst, writer, and ghost-writer. Please follow DancingDinosaur on Twitter, @mainframeblog. See more of his IT writing at technologywriter.com and here.

IBM Gets Serious About Open Data Science (ODS) with Anaconda

As IBM rapidly ramps up cognitive systems in various forms, its two remaining platforms, z System and POWER, get more and more interesting. This week IBM announced it was bringing the Anaconda Open Data Science (ODS) platform to its Cognitive Systems and PowerAI.

Anaconda, Courtesy Pinterest

Specifically, Anaconda will integrate with the PowerAI software distribution for machine learning (ML) and deep learning (DL). The goal: make it simple and fast to take advantage of Power performance and GPU optimization for data-intensive cognitive workloads.

“Anaconda on IBM Cognitive Systems empowers developers and data scientists to build and deploy deep learning applications that are ready to scale,” said Bob Picciano, senior vice president of IBM Cognitive Systems. Added Travis Oliphant, co-founder and chief data scientist, Continuum Analytics, which introduced the Anaconda platform: “By optimizing Anaconda on Power, developers will also gain access to the libraries in the PowerAI Platform for exploration and deployment in Anaconda Enterprise.”

With more than 16 million downloads to date, Anaconda has emerged as the Open Data Science platform leader. It is empowering leading businesses across industries worldwide with tools to identify patterns in data, uncover key insights, and transform basic data into the intelligence required to solve the world’s most challenging problems.

As one of the fastest growing fields of AI, DL makes it possible to process enormous datasets with millions or even billions of elements and extract useful predictive models. DL is transforming the businesses of leading consumer Web and mobile application companies, and it is catching on with more traditional business.

IBM developed PowerAI to accelerate enterprise adoption of open-source ML and DL frameworks used to build cognitive applications. PowerAI promises to reduce the complexity and risk of deploying these open source frameworks for enterprises on the Power architecture and is tuned for high performance, according to IBM. With PowerAI, organizations also can realize the benefit of enterprise support on IBM Cognitive Systems HPC platforms used in the most demanding commercial, academic, and hyperscale environments

For POWER shops getting into Anaconda, which is based on Python, is straightforward. You need a Power8 with IBM GPU hardware or a Power8 combined with a Nvidia GPU, in effect a Minsky machine. It’s essentially a developer’s tool although ODS proponents see it more broadly, bridging the gap between traditional IT and lines of business, shifting traditional roles, and creating new roles. In short, they envision scientists, mathematicians, engineers, business people, and more getting involved in ODS.

The technology is designed to run on the user’s desktop but is packaged and priced as a cloud subscription with a base package of 20 users. User licenses range from $500 per year to $30,000 per year depending on which bells and whistles you include. The number of options is pretty extensive.

According to IBM, this started with PowerAI to accelerate enterprise adoption of open-source ML/DL learning frameworks used to build cognitive applications. Overall, the open Anaconda platform brings capabilities for large-scale data processing, predictive analytics, and scientific computing to simplify package management and deployment. Developers using open source ML/DL components can use Power as the deployment platform and take advantage of Power optimization and GPU differentiation for NVIDIA.

Not to be left out, IBM noted growing support for the OpenPOWER Foundation, which recently announced the OpenPOWER Machine Learning Work Group (OPMLWG). The new OPMLWG includes members like Google, NVIDIA and Mellanox to provide a forum for collaboration that will help define frameworks for the productive development and deployment of ML solutions using OpenPOWER ecosystem technology. The foundation has also surpassed 300-members, with new participants such as Kinetica, Red Hat, and Toshiba. For traditional enterprise data centers, the future increasingly is pointing toward cognitive in one form or another.

DancingDinosaur is Alan Radding, a veteran information technology analyst, writer, and ghost-writer. Please follow DancingDinosaur on Twitter, @mainframeblog. See more of his IT writing at technologywriter.com and here.

 

IBM Nvidia GPU Speed Workloads in Face of Zettabyte Volumes

Last week IBM became the  first major global cloud provider to make the NVIDIA Tesla P100 GPU accelerator available on the cloud. By combining NVIDIA’s acceleration technology with IBM’s Cloud platform, businesses can expect to more quickly and efficiently run compute-heavy workloads.

A few compute-heavy workloads, however, will barely be noticed in the aggregate volume of data that analyst firms like IDC see coming by 2025. IDC analysts David Reinsel, John Gantz and John Rydning, in their latest Digital Universe study conclude: The 163 zettabyte (ZB) global datasphere projected in Data Age 2025 is only the beginning.

Courtesy: VMware

Today, IBM and Nvidia see the big zettabytes driver as cognitive analytics, but that’s just one part and not even the largest. As IBM explains: The combination of IBM’s connectivity and bare metal servers with the Tesla P100 GPUs enables higher throughput than traditional virtualized servers. This high level of performance can allow organizations to deploy fewer, more powerful cloud servers to more quickly deliver increasingly complex simulations and big data workloads. Even before 2025, you can’t deploy enough NVIDIA technology to matter.

The IDC analysts take a more visionary view: Imagine being awoken and tended to by a virtual personal assistant that advises you on what clothing from your wardrobe is best suited to the weather report and your schedule for the day or being transported by your self-driving car. Or maybe you won’t need to commute to an office at all as technology will allow you to conjure workspaces out of thin air using interactive surfaces, and holographic teleconferencing becomes the norm for communicating virtually with colleagues.

Behind this vision are data and processing. As the IDC analysts put it: In earlier periods, data growth stemmed largely from the rise of the personal computer and the consumption of digital entertainment. The world today, however, contains more consumer devices (PCs, phones, game consoles, and music players) than human beings, and all these devices require data.

The embedding of computing power in a large number of endpoint devices also contributes to data growth. Today, the number of embedded system devices feeding into datacenters is less than one per person globally, but over the next 10 years, that number will increase to more than four per person.  And the number of files generated will be unimaginable, measured in quintillions per year. BTW, it would take Niagara Falls 210,000 years to move one quintillion gallons of water.

Don’t worry about storing all these zettabytes. As the analysts note: The vast majority of the global datasphere is used and discarded rather than stored. This is primarily a reflection of the fact that most data is fundamentally disposable. Here is where metadata comes into play. From the huge amount of data created, you can prioritize which data has sufficient value to be stored through metadata, classification, tagging, and such.

The key: apply cognition to the selection of data to store. Let cognition determine the smart criteria to identify which data to retain, in what form, and for how long. That way you can hang onto critical information without the need to store all the data produced.

Even still, the quantity of data will outpace any reasonable expectation to store all the data. For example, it would take roughly 16 billion of today’s largest 12TB enterprise HDDs to store the 163ZB data expected to be created in 2025. But over the past 20 years, the entire disk drive industry shipped only 8 billion HDDs amounting to nearly 4ZB of storage.

Obviously, you can’t store 163 ZB of data or even a miniscule piece of it. The solution points to intelligence embedded throughout the system. Start with the application of smart criteria described above. And don’t ignore intelligent automation; people can’t possibly handle this without assistance.

And the 163ZB global datasphere projected in Data Age 2025 is only just the beginning. Another decade in technology years will likely bring about unforeseen advancements, use cases, businesses, and life- changing services that rely on yet more data. As IDC notes, the quantity of data generated will continue to outpace any notion of storing all of the data. Start now incorporating cognition to determine metadata, file tagging, data classification, and expiration dating.

DancingDinosaur is Alan Radding, a veteran information technology analyst, writer, and ghost-writer. Please follow DancingDinosaur on Twitter, @mainframeblog. See more of his IT writing at technologywriter.com and here.