Google has announced that its cloud tensor processing unit (TPU) chips are available in beta on Google Cloud Platform, aimed at speeding and scaling up specific machine learning workloads.

The company claims that using the new chips will help train business-critical machine learning models in hours rather than days or weeks. The units will utilise TensorFlow, the open source machine learning framework, with up to 180 teraflops of floating point performance – in other words, being capable of handling 180 trillion floating point calculations each second – and 64 GB of high-bandwidth memory onto a single board.

According to a blog post from John Barrus, product manager for cloud TPUs at Google Cloud, and Zak Stone, product manager for TensorFlow and cloud TPUs, the units are available in ‘limited quantities’ today with usage billed by the second at a rate of $6.50 USD per cloud TPU per hour.

As ever, Google has put together various customers exhorting the benefits of the new product. Alfred Spector, chief technology officer at investment provider Two Sigma, said the Google Cloud TPUs were ‘an example of innovative, rapidly evolving technology to support deep learning.’

“We made a decision to focus our deep learning research on the cloud for many reasons, but mostly to gain access to the latest machine learning infrastructure,” said Spector. “Using Cloud TPUs instead of clusters of other accelerators has allowed us to focus on building our models without being distracted by the need to manage the complexity of cluster communication patterns.”

“Here at Google Cloud, we want to provide customers with the best cloud for every ML workload and will offer a variety of high-performance CPUs (including Intel Skylake) and GPUs (including NVIDIA’s Tesla V100) alongside Cloud TPUs,” Barrus and Stone wrote.

You can find out more about the Cloud TPU machine learning accelerators here.

We are given a desktop platform with Java 8 or Java 9 installed and seek to find a way to deploy high-performance Java applications that use Java 3D and/or Jogl without having to run an installer. We are subject to the constraint that the applications be signed and deployed so that they can be run in a trusted environment (i.e., outside of the sandbox). Further, we seek to do this in a way that does not depend on bundling a JRE with our applications, as this makes downloads and installations rather long.

read more

Guided by senior executive goals for digital transformation, more organizations are increasing their use of cloud computing technologies. The typical multi-cloud mix includes a blend of public cloud, private cloud and traditional IT services. Finding the best-fit service mix starts with the business requirements.

The latest Global Cloud Index (GCI) 2016-2021 from Cisco focuses on the worldwide market outlook for enterprise data centre virtualization and cloud computing services. Today's digital business is enabled by Hybrid IT infrastructure that supports the deployment of cloud-based solutions.

Driven by the surging enthusiasm for digital reinvention projects, data centre traffic is growing fast. The market study authors forecast global cloud data centre traffic will reach 19.5 zettabytes (ZB) per year by 2021 — that's up from 6.0 ZB per year in 2016.

Globally, cloud data centre traffic will represent 95 percent of total data centre traffic by 2021, compared to 88 percent in 2016.

Additionally, according to the Cisco assessment, the growth of Internet of Things (IoT) applications requires scalable server and storage solutions to accommodate new and expanding data centre demands.

By 2021, Cisco expects IoT connections to reach 13.7 billion — that's up from 5.8 billion in 2016.

Hyperscale cloud data centre growth

By 2021 there will be 628 hyperscale data centres globally, compared to 338 in 2016 — that's 1.9-fold growth or near doubling over the forecast period. By 2021, hyperscale data centres will support:

• 53 percent of all data centre servers (27 percent in 2016)
• 69 percent of all data centre processing power (41 percent in 2016)
• 65 percent of all data stored in data centres (51 percent in 2016)
• 55 percent of all data centre traffic (39 percent in 2016)

Data centre virtualization and cloud growth

By 2021, 94 percent of workloads and compute instances will be processed by cloud data centres. In contrast, 6 percent will be processed by traditional IT data centres.

Overall data centre workloads and compute instances will more than double (2.3-fold) from 2016 to 2021; however, cloud workloads and compute instances will nearly triple (2.7-fold) over the same period.

The workload and compute instance density for cloud data centres was 8.8 in 2016 and will grow to 13.2 by 2021. Comparatively, for traditional data centres, workload and compute instance density was 2.4 in 2016 and will grow to 3.8 by 2021.

Stored data growth fuelled by big data and IoT

Globally, the data stored in data centres will nearly quintuple by 2021 to reach 1.3 ZB by 2021, up 4.6-fold (a CAGR of 36 percent) from 286 EB in 2016.

Big data will reach 403 exabytes (EB) by 2021, up almost 8-fold from 25 EB in 2016. Big data will represent 30 percent of data stored in data centres by 2021, up from 18 percent in 2016.

The amount of data stored on devices will be 4.5 times higher than data stored in data centres, at 5.9 ZB by 2021.

Driven largely by IoT, the total amount of data created (and not necessarily stored) by any device will reach 847 ZB per year by 2021, up from 218 ZB per year in 2016. Data created is two orders of magnitude higher than data stored.

Apps contribute to rise of global data centre traffic

– By 2021, big data will account for 20 percent (2.5 ZB annual, 209 EB monthly) of traffic within data centres, compared to 12 percent (593 EB annual, 49 EB monthly) in 2016.

– By 2021, video streaming will account for 10 percent of traffic within data centres, compared to 9 percent in 2016.

– By 2021, video will account for 85 percent of traffic from data centres to end users, compared to 78 percent in 2016.

– By 2021, search will account for 20 percent of traffic within data centres by 2021, compared to 28 percent in 2016.

– By 2021, social networking will account for 22 percent of traffic within data centres, compared to 20 percent in 2016

SaaS is still the most popular cloud service model

By 2021, 75 percent (402 million) of the total cloud workloads and compute instances will be SaaS workloads and compute instances, up from 71 percent (141 million) in 2016. (23 percent CAGR from 2016 to 2021).

By 2021, 16 percent (85 million) of the total cloud workloads and compute instances will be IaaS workloads and compute instances, down from 21 percent (42 million) in 2016. (15 percent CAGR from 2016 to 2021).

By 2021, 9 percent (46 million) of the total cloud workloads and compute instances will be PaaS workloads and compute instances, up from 8 percent (16 million) in 2016. (23 percent CAGR from 2016 to 2021).

Product connectivity goes hand and hand these days with increased use of personal data. New IoT devices are becoming more personalized than ever before.
In his session at 22nd Cloud Expo | DXWorld Expo, Nicolas Fierro, CEO of MIMIR Blockchain Solutions, will discuss how in order to protect your data and privacy, IoT applications need to embrace Blockchain technology for a new level of product security never before seen – or needed.

read more

Imagine if you will, a retail floor so densely packed with sensors that they can pick up the movements of insects scurrying across a store aisle. Or a component of a piece of factory equipment so well-instrumented that its digital twin provides resolution down to the micrometer.

read more

With continuous delivery (CD) almost always in the spotlight, continuous integration (CI) is often left out in the cold. Indeed, it’s been in use for so long and so widely, we often take the model for granted. So what is CI and how can you make the most of it? This blog is intended to answer those questions.
Before we step into examining CI, we need to look back. Software developers often work in small teams and modularity, and need to integrate their changes with the rest of the project code base. Waiting to integrate code creates merge conflicts, bugs that can be tricky to resolve, diverging code strategies, and duplicated effort. Before CI, developers also had a problem of wasted time: they compiled and built their changes multiple times a day on their own machine and had to sit idle waiting for these processes to complete. Integrating the changes from all teams happened only at night on big dedicated “build” servers and any issues during this build could create further idle time the next morning as developers isolate and resolve last night’s issues.

Before we step into examining CI, we need to look back. Software developers often work in small teams and modularity, and need to integrate their changes with the rest of the project code base. Waiting to integrate code creates merge conflicts, bugs that can be tricky to resolve, diverging code strategies, and duplicated effort. Before CI, developers also had a problem of wasted time: they compiled and built their changes multiple times a day on their own machine and had to sit idle waiting for these processes to complete. Integrating the changes from all teams happened only at night on big dedicated “build” servers and any issues during this build could create further idle time the next morning as developers isolate and resolve last night’s issues.

read more

There is a war a-brewin’, but this war will be fought with wits and not brute strength. Ever since Russian President Vladimir Putin’s declaration that “the nation that leads in AI (Artificial Intelligence) will be the ruler of the world,” the press and analysts have created hysteria regarding the ramifications of artificial intelligence on everything from public education to unemployment to healthcare to Skynet.
Note: artificial intelligence (AI) endows applications with the ability to automatically learn and adapt from experience via interacting with the surroundings / environment. See the blog “Artificial Intelligence is not Fake Intelligence” for a more detailed explanation on artificial intelligence and machine learning.

read more

An interesting series of findings from ISACA in its latest report: almost one in three organisations polled are not calculating return on investment in their cloud computing initiatives.

The findings, which appear in a report titled ‘How Enterprises Are Calculating Cloud ROI – And Why Some Enterprises Are Moving Ahead Without It’, reveal how companies are increasingly moving away from the ROI model. “If ROI is not calculated in advance of implementation, it becomes difficult to validate or refute the expected value,” said Ed Moyle, ISACA director of thought leadership and research.

As a result, enterprises who are not going down the ROI route are instead calculating cloud investments by non-financial criteria, such as greater business agility, and shifting funding from capital expenses to operating expenses.

78% of the more than 100 CIOs polled said their enterprise has implemented some form of cloud computing. Of that figure, 84% said their organisation uses software as a service, with 49% using infrastructure as a service and 36% platform as a service.

Just over a third (35%) of respondents said they calculated ROI on cloud initiatives before and after implementation, compared with 29% who only do it before implementation, and 4% who opt for after. Of those who do calculate ROI, just under half (49%) opt for a hybrid model, compared with a quantitative (45%) and qualitative (23%) model.

When it came to what factors organisations considered when putting together their ROI calculation, operational expense formed 98% of responses, while capital expenses or savings (91%) was also highly considered. Changes in staffing requirements – cited by 68% of those polled – business impact (66%) and transition expenses (57%) were also cited by the majority.

The wide range of responses and lack of consensus leads ISACA to one conclusion; the need for an industry-wide tool. “Although the majority of enterprises do continue to calculate cloud ROI at some point in the cloud implementation process, the challenges associated with the mechanisms available for them to do so are clear,” the report concludes. “Without an industry-wide consensus for cloud ROI calculation, adoption of a formalised model is understandably stunted.”

The issue of calculating cloud ROI has naturally been around since the very first initiatives. ISACA itself proposed such a universal model in its previous ‘Calculating Cloud ROI: From the Customer Perspective’ whitepaper. Back in 2011, this publication explored some of the models around at the time, including Azure, Amazon Web Services, and VMware.

You can read the full ISACA paper here (registration required).

• Enterprise investments in machine learning will nearly double over the next three years, reaching 64% adoption by 2020.
• International Data Corporation (IDC) is forecasting spending on artificial intelligence (AI) and machine learning will grow from $8B in 2016 to $47B by 2020.
• 89% of CIOs are either planning to use or are using machine learning in their organizations today.
• 53% of CIOs say machine learning is one of their core priorities as their role expands from traditional IT operations management to business strategists.
• CIOs are struggling to find the skills they need to build their machine learning models today, especially in financial services.

These and many other insights are from the recently published study, Global CIO Point of View. The entire report is downloadable here (PDF, 24 pp., no opt-in). ServiceNow and Oxford Economics collaborated on this survey of 500 CIOs in 11 countries on three continents, spanning 25 industries. In addition to the CIO interviews, leading experts in machine learning and its impact on enterprise performance contributed to the study. For additional details on the methodology, please see page 4 of the study and an online description of the CIO Survey Methodology here.

Digital transformation is a cornerstone of machine learning adoption. 72% of CIOs have responsibility for digital transformation initiatives that drive machine learning adoption. The survey found that the greater the level of digital transformation success, the more likely machine learning-based programs and strategies would succeed. IDC predicts that 40% of digital transformation initiatives will be supported by machine learning and artificial intelligence by 2019.

Key takeaways from the study include the following:

90% of CIOs championing machine learning in their organizations today expect improved decision support that drives greater topline revenue growth

CIOs who are early adopters are most likely to pilot, evaluate and integrate machine learning into their enterprises when there is a clear connection to driving business results. Many CIO compensation plans now include business growth and revenue goals, making the revenue potential of new technologies a high priority.

89% of CIOs are either planning to use or using machine learning in their organizations today

The majority, 40%, are in the research and planning phases of deployment, with an additional 26% piloting machine learning. 20% are using machine learning in some areas of their business, and 3% have successfully deployed enterprise-wide. The following graphic shows the percentage of respondents by stage of their machine learning journey.

Machine learning is a key supporting technology leading the majority finance, sales and marketing, and operations management decisions today

Human intervention is still required across the spectrum of decision-making areas including Security Operations, Customer Management, Call Center Management, Operations Management, Finance and Sales & Marketing. The study predicts that by 2020, machine learning apps will have automated 70% of Security Operations queries and 30% of Customer Management ones.

Automation of repetitive tasks (68%), making complex decisions (54%) and recognizing data patterns (40%) are the top three most important capabilities CIOs of machine learning CIOs are most interested in

Establishing links between events and supervised learning (both 32%), making predictions (31%) and assisting in making basic decisions (18%) are additional capabilities CIOs are looking for machine learning to accelerate. In financial services, machine learning apps are reviewing loan documents, sorting applications to broad parameters, and approving loans faster than had been possible before.

Machine learning adoption and confidence by CIOs varies by region, with North America in the lead (72%) followed by Asia-Pacific (61%)

Just over half of European CIOs (58%) expect value from machine learning and decision automation to their company’s overall strategy. North American CIOs are more likely than others to expect value from machine learning and decision automation across a range of business areas, including overall strategy (72%, vs. 61% in Asia Pacific and 58% in Europe). North American CIOs also expect greater results from sales and marketing (63%, vs. 47% Asia-Pacific and 38% in Europe); procurement (50%, vs. 34% in Asia-Pacific and 34% in Europe); and product development (48%, vs. 29% in Asia-Pacific and 29% in Europe).

CIOs challenging the status quo of their organization’s analytics direction are more likely to rely on roadmaps for defining and selling their vision of machine learning’s revenue contributions

More than 70% of early adopter CIOs have developed a roadmap for future business process changes compared with just 33% of average CIOs. Of the CIOs and senior management teams in financial services, the majority are looking at how machine learning can increase customer satisfaction, lifetime customer value, improving revenue growth. 53% of CIOs from our survey say machine learning is one of their core priorities as their role expands from traditional IT operations to business-wide strategy.

Sources: CIOs Cutting Through the Hype and Delivering Real Value from Machine Learning, Survey Shows

The need in agility and scalibility neccecisited the evolution on software development from monolithic to microservices to serverless architecture.

read more