NVIDIA achieved six new MLPerf AI performance records with the release of the industry’s first broad set of AI benchmarks. Tensor Core GPUs achieved the best performance across every MLPerf benchmark submitted.

MLPerf consortium was formed to build a common set of benchmarks to enable the machine learning (ML) field to measure performance for both training and inference from mobile devices to cloud services. Google, Intel, Baidu, NVIDIA, Facebook, Microsoft and more than dozens of technology leaders support MLPerf.org. It covers AI fields like computer vision, language translation, personalized recommendations, and reinforcement learning tasks.

NVIDIA submitted results for six categories such as image classification, object instance segmentation, object detection, non-recurrent translation, recurrent translation, and recommendation systems. NVIDIA did not submit results for the seventh category, Reinforcement Learning as this test category does not take advantage of GPU acceleration.

NVIDIA, MLPerf Categories

According to NVIDIA announcement, "MLPerf scores are computed by first measuring time to train to the specified target accuracy, then normalizing to the time taken by an unoptimized implementation on a reference platform. The goal of normalization is to have a similar magnitude of scores across different benchmarks since benchmarks take different amounts of time to train. Both the time-to-train and MLPerf scores are published on the MLPerf website, and so we present time to train in minutes."

NVIDIA used NVIDIA DGX systems, including NVIDIA DGX-2 featuring 16 fully connected V100 Tensor Core GPUs and claims that it is the world’s most powerful AI system. Results for NVIDIA time to convergence, both for single-node and at-scale implementations are depicted in the image below.



Vice president and general manager of Accelerated Computing at NVIDIA, Ian Buck, said, “The new MLPerf benchmarks demonstrate the unmatched performance and versatility of NVIDIA’s Tensor Core GPUs. Exceptionally affordable and available in every geography from every cloud service provider and every computer maker, our Tensor Core GPUs are helping developers around the world advance AI at every stage of development.”

The 18.11 release of the NVIDIA GPU Cloud (NGC) deep learning containers with the complete software stack and the top AI frameworks, optimized by NVIDIA that achieved outstanding MLPerf performance are available for download from NVIDIA GPU Cloud (NGC) container registry.
The details of benchmark results are available at https://mlperf.org/results/.

Massachusetts Institute of Technology (MIT) Engineers successfully trained a deep neural network to detect tiny objects in the dark that are normally invisible to human eyes.

Deep learning tries to emulate the behavior of the human brain to process input data sets. It consists of different layers of mathematical models and each accepting input and predicting the output at every stage. The inputs are generally fed in the form of images, text, and sound. Its performance and accuracy often surpass human-level.

The study is published in Physical Review Letters where the researchers reconstructed transparent objects such as Small imperfections in a wine glass or tiny creases in a contact lens, in almost pitch-black conditions. Barbastathis’ co-authors on the paper are lead author Alexandre Goy, Kwabena Arthur, and Shuai Li.

Co-authors on the paper, Alexandre Goy said, “When we look with the naked eye, we don’t see much — they each look like a transparent piece of glass. But there are actually very fine and shallow structures that still have an effect on light.”

From an original transparent etching (far right), engineers produced a photograph in the dark (top left), then attempted to reconstruct the object using first a physics-based algorithm (top right), then a trained neural network (bottom left), before combining both the neural network with the physics-based algorithm to produce the clearest, most accurate reproduction (bottom right) of the original object. Courtesy of the MIT researchers

The 10,000 transparent images of glass-like etchings, based on extremely grainy patterns, were used to train the neural network. The images were taken in very low lighting conditions, with about one photon per pixel. This is far less light than a camera would register in a dark sealed room.

As a use case, transparent features of biological tissues and cells can be illuminated with the help of this technique.

Professor of mechanical engineering at MIT, George Barbastathis, said, “In the lab, if you blast biological cells with light, you burn them, and there is nothing left to the image. When it comes to X-ray imaging, if you expose a patient to X-rays, you increase the danger they may get cancer. What we’re doing here is, you can get the same image quality, but with a lower exposure to the patient. And in biology, you can reduce the damage to biological specimens when you want to sample them.”

Google announced the availability of several new features for its Google Cloud Platform (GCP) customers, as part of the beta release of its Cloud Security Command Center (Cloud SCC).

Google Cloud Security Command Center (Cloud SCC) has entered into Beta phase and available to all GCP customers.

This recent announcement was followed after the biggest cloud vendor AWS announced its Security Hub solution and onboarding of major security players. Notably, the alpha release of Cloud Security Command Center was announced in March 2018.

The beta release of Cloud SCC now give insights into the inventory of GCP services such as App Engine, Cloud Datastore, Cloud DNS, Cloud Load Balancing, Cloud Spanner, Cloud Storage, Compute Engine, Container Registry, Kubernetes Engine, and Virtual Private Cloud along with an inventory of your service accounts. In addition, 13 IAM roles added for fine-grained access control.

Other capabilities include the ability to view and search for new, deleted, and total assets over a time period along with new kind of notifications when changes occur.

The notifications can help trigger Cloud Functions from a Cloud SCC query such as policy changes on a network firewall and then restore it back to a secure state, and detect when a Cloud Storage bucket becomes publicly accessible and then revert back to private access. A new set of client libraries including Java, Node, and Go are also made available.

Google Cloud Security Command Center integrates with Google Cloud security services such as Data Loss Prevention (DLP) API, Forseti, Cloud Security Scanner, and anomaly detection from Google as well as with third-party cloud security solutions from vendors such as Cavirin, Chef, and Redlock is also available in beta release.

SCC can uncover security risks and vulnerabilities. According to the blog post by Senior Product Manager, Google Cloud, Andy Chang, "you can view which Cloud Storage buckets are publicly accessible, identify VMs with public addresses, discover overly permissive firewall rules, and be alerted to instances that may have been compromised to perform coin mining. You can also easily see if users outside of your designated domain, or GCP organization, have access to your resources."

Director Information Security, Colgate-Palmolive, Alexander Schuchman, said, "Cloud Security Command Center gives us unprecedented visibility into the security posture of our VM instances and containerized workloads running within GCP. With this security service, we can quickly review and assess risks across all our GCP assets."

Cloud SCC performs regular discovery scans and helps GCP admins to act on policy violations or alerts. Now it is possible to view, monitor and observe an inventory of your cloud assets as well as security anomalies and help protect sensitive data from a single centralized data platform and dashboard

Google Cloud SCC can be accessed here.

The recent sellout of various cryptocurrencies and the ever dipping price of Bitcoin and altcoins have to lead to the first-ever layoff in the crypto industry.  This is a move taken by different firms working to survive the sellout that is going on currently.

ConsenSys, a known venture which is working with the blockchain technology, announced that they would be overhauling their entire business model and cut off about 13% of the total staff. Some other companies which were able to raise money last year have either quietly closed or retrenched sharply.

Steemit, another firm which runs a social network on a blockchain-based network, laid off almost 70% of its staff because of the selloff.

If you take a closer look at 2017, the markets were booming which in turn fueled a hiring spree with hundreds of startups developing new uses of the blockchain. Billions of dollars were raised using ICOs.

The split of Bitcoin has brought down the crypto market to its knees. Bitcoin is more than 80% down from its December 2017 value, with the total value of all cryptocurrencies only about $111 billion, which was floating around $827 just at this beginning of this year.

This is not the end. According to Ernst and Young, more than 86% of ICOs are trading below their listing price and have lost nearly 30% of their value. “As things get real we have to be a lean, business-minded organization,” said ConsenSys founder Joseph Lubin.

The company uses the Ethereum platform and operates as a development house for startups who are looking to build products using the same. They provide services and support in exchange for equity stakes.

Even people who were looking forward to working in the sector are now seemingly losing interest in the industry. Indeed.com, a portal for job searches worldwide, saw a 3% fall in the number of searches through October. Rewind 12 months and you would see a 483% increase in the phenomena. It's not that the number of blockchain related are stagnated, they are still increasing, albeit at a much slower pace than before.

Another blockchain based identity verification firm known as Civic was able to raise $33 million last year from its ICO, but according to the founder and Chief Executive of the firm Vinny Lingham, they have converted 90% of those assets into US dollars.

Token Agency, an advisory firm which consults firms considering ICOs, was forced to cut down its 20 employee team down just four and most of them were paid in cryptocurrencies. Mr. Behnke, the founder of the company, said he didn’t hedge properly against potential declines in Ethereum, which has fallen more than 90% this year. “My biggest issue throughout was not having proper risk management,” he said. “We didn’t anticipate it would be this quick and abrupt.”

“I think what’s going to turn this around is real products actually touching people,” Mr. Behnke said.

According to analysts, this sellout has forced the industry to focus on products which will have real-world implications and are giving organizations a chance to reconsider their strategy regarding the services they are offering.

Cloud Native Computing Foundation, with the aim to enable cloud portability without vendor lock-in, conducts its bi-annual cloud-native survey in Asia. The results were declared last month and we have summarized the readings from this survey.

CNCF has a total of 42 members across China, Japan, and South Korea. Among these members are Alibaba Cloud, Fujitsu, Huawei, and JD.com.

Survey Highlights

Key Takeaways from CNCF cloud-native survey in Asia are:

1. In Asia, the growth of 135% is witnessed in public and private clouds usages since March 2018
2. There is a 48% drop in on-premise cloud usages
3. Home-grown container management solutions usages are up by 690% while commercial solutions up by 58%
4.  Kubernetes usages have grown by11% in Asia.
5. In Asia, the number of Kubernetes clusters in production is increasing. Organizations with 11-50 production clusters have increased by 154% while smaller 1-5 cluster deployments are decreased by 37%.
6. Alibaba Cloud remains at the top position with 38% reported usages among the respondants followed by AWS usages at 24%.
7. New environments that were not previously reported and are taking up additional market share are Huawei Cloud (13%), VMware (6%), Baidu Cloud (21%), Tencent Cloud (7%), IBM Cloud (8%), and Packet (5%).
8. OpenStack usage down to 9% from 26% in March 2018. This downshift looks significant.
9. Asia has 100% growth in serverless technology usages with 29% of respondents using installable software and 21% using a hosted platform.
10. Apache OpenWhisk is the most popular with 11% usages for installable serverless platforms. Other platforms followed are Dispatch (6%), FN (5%) and OpenFaaS, Kubeless, and Fission tied at 4%.
11. AWS Lambda is the stands at 11% with hosted platform usages. Other hosted platforms are Azure Functions (8%), and Alibaba Cloud Compute Functions, Google Cloud Functions, and Cloudflare Functions tied at 7%.

The details of the CNCF cloud-native survey are available here.

Docker, at its European user conference, has announced its enterprise edition for building as well as deploying container-based solutions, which now comes with enterprise-grade security.

According to the firm, it is the easiest, fastest and most secure way for delivery of production-ready containerized apps. Developed for both Windows and Mac, developers can adhere to corporate standards and practices while rapidly and securely delivering apps from development to production.

Docker's solution is designed to enhance productivity features which make it easy to use containerized apps independent of the framework, language used or the development process. The current iterations of available softwares do not allow to sync the production environments and application architectures, thus increasing the time for software deployment.

Chief Product Officer at Docker, Scott Johnston, said, “The success of Docker Desktop among developers has been that it is the best and easiest way to run a Docker Engine and tooling with popular orchestrators like Kubernetes. Customers came to us because they wanted a way to scale Docker Desktop across an enterprise-class environment with the additional security, standardization, and management needed for delivering applications across the dev to ops lifecycle. Docker Desktop Enterprise enables an organization to address corporate standards while providing an environment that fosters developer productivity.”

With almost 1.4 million daily users, Docker's Desktop Enterprise Edition simplifies solution's to various problems by reducing complexity all the while providing powerful functionalities in the form of Docker Engine with Swarm and Kubernetes. The installation is pretty minimal and requires no additional setup.

You can also auto-generate Docker artifacts or use the Docker CLI with various tools such as Microsoft Visual Studio, Eclipse, IntelliJ IDEA, etc. Docker's Enterprise solution can also reliably and securely create a model using pre-approved templates which can integrate with its access control and image verification tools.

Using MSI and PKG formats, the Enterprise solution. can configure, deploy and manage across distributed developer team with preferred endpoint management tools. With Kubernetes and Docker-engine built in, it can easily swap version packs which instantly replicate production environment configs on the local desktop.

This will enable users to synchronize the desktop development environment with same Kubernetes and Docket API versions used in the production process.

With Docker Desktop Enterprise, organizations can reduce developer friction and compatibility issues and accelerate time-to-production with fully compatible production-ready environments.

In just a matter of 6 months, Google Cloud has expanded the number of GCP regions in the Asia Pacific from three to six.

The addition of each region will lead to a lower latency to each of its customer which will translate to a reduction in the response time from the servers.

It has recently opened its sixth APAC region in Hong Kong and has plans to build another one in Jakarta, Indonesia in the upcoming future.

Specially designed for Hong Kong

Google's Hong Kong business is open as of now. This is the eighteenth overall and is located to provide faster access to applications which are making use of Google's Cloud solution in the Hong Kong and Southeast Asian region.

A recent whitepaper published by Google Hong Kong indicates that around 54% of the Island nations resident is planning to launch products which will ultimately work on a cloud computing platform.

The Hong Kong GCP region is, therefore, designed to support the upcoming wave of customers in the area. It will have three availability zones, each of which helps customers to distribute workloads and storages to enable quick response time.

GCP Customer Opinions

The Group Chief Information Officer, HSBC, Darryl West said, "We’ve made a decision on applications hosting to partner with Google on building some next-gen applications in the full Google stack, top to bottom using Google tools, which is a big departure from our previous architecture. It’s very exciting because I think we’ve got the confidence that the tools they have are going to be great for productivity with developers."

Organizations or people making use of a hybrid cloud can integrate their new and existing deployments using the regional partner ecosystem as well as two dedicated interconnect points of presence.

Another GCP customer, Chief Technology Officer, and Co-Founder at Klook, Bernie Xiong, said, "Google Cloud remains to be a strong partner in Klook’s continuous growth. By using Google big data and analytics, we have been able to draw insights from our data so we can provide better services to travelers worldwide instead of spending time on infrastructure."

The launch definitely brings with itself speedier access to GCP products and services. The latency for end users is said to be improved by up to a staggering 14ms. Also, customers who are using GCP in Vietnam and the Philippines will also see a benefit of up to 25-30% in latency.

Google has made it possible for end users to see how quickly you can access the Hong Kong region. Visit www.gcping.com to check your speed and latency. If you are thinking about using services which are not yet available in the region, don't worry, you can gain access to those using the Google Network.

Google is planning to open it's GCP region in Jakarta, Indonesia and has plans to expand to Osaka, Japan. Jakarta region will be eight regions in Asia Pacific Hong Kong, Mumbai, Sydney, Singapore, Taiwan, and Tokyo.

The availability of various services and regions can be checked by visiting their location page.

StableUSD (USDS), a newly born stable coin, now listed Bittrex, a crypto exchange platform.

Stably is the startup behind StableUSD and backed by venture capital. It already raised seed funding from 500 Startups, Beenext Ventures, and angel investors to kickstart our project.

With this announcement, StableUSD (USDS) has joined the league of other major stablecoins such as the Gemini Dollar (GUSD), MakerDao (DAI), Paxos Standard Token (PAX) and TrueUSD (TUSD).

Digital Exchanges around the world are trying hard to bring stability to crypto market by listing stablecoins. Here is the list of a recent listing of different stablecoins on different crypto exchanges.

• Recently, Liquid cryptocurrency exchange platform listed GUSD and USDC on their platform.
• CobinHood Cryptocurrency Exchange Lists Four Major Stable-Coins namely the Gemini Dollar (GUSD), MakerDao (DAI), Paxos Standard Token (PAX) and TrueUSD (TUSD).
• Binance Cryptocurrency Exchange Platform Lists USDC Stable Coin

Cryptocurrencies such as Bitcoin, Ethereum and other Altcoins can be extremely volatile. Some of them moving up and down by around a factor of 20 to 30 percent within a couple of hours. It is believed that stablecoins can help bring much-needed stability and predictability into crypto markets though this hypothesis remains to be tested to be true. Stablecoins are generally pegged against fiat currency like USD, EUR, CNY or JPY and sometimes against gold or oil.

Brief about USDS Stablecoin

StableUSD Coin (USDS), a regulated, US dollar-backed stable coin governed by Stably and audited by an auditor, Cohen & Co. The claim is that this ratio is guaranteed whenever one purchases or redeems USDS directly with Stably only. Stably holds a one-to-one reserve of USD against all issued USDS in the market. This ratio may not be one-to-one in the open market though.

Etherscan lists the Ethereum transaction history for USDS here. As of now, Bitrex is trading USDC-BTC pair only.

Stably's announcement further said, "The first exchange that USDS will be listed on is Bittrex, which will open trading for USDS/BTC trading pairs at 9:00 AM PST (Thursday, December 6th, 2018). Leading third-party auditor, Cohen & Co., will provide regular attestations on our reserve balance — an attestation on our initial balance will be published next week. Additionally, you can find a link to our smart contract and Ethereum transaction history for USDS here."

More details about USDC stable coin redemptions and purchases are available here.

Microsoft and Docker partnered and announced open source, the cloud-agnostic specification for packaging and running distributed applications called Cloud Native Application Bundle (CNAB). Docker, HashiCorp, and Bitnami will also offer CNAB bundles that customers would be able to host or download from DockerHub, GitHub, etc.

Why CNAB?

As industry moved toward the cloud environment, legacy and new edge modern applications followed the trend to run on cloud computing infrastructure and leverage different cloud capabilities such as elasticity, high availability, and resiliency. However, organizations are moving towards the hybrid cloud strategies that demand on-premise as well as off-premise or public/remote cloud environments.

Tools like Terraform for the infrastructure, Helm charts, and Docker Compose files for the applications, and CloudFormation or ARM templates for the cloud-services needs to be managed separately. And there was a need to define a specification for packaging these multi-service, multi-format distributed applications and this is where Microsoft and Docker came together to release Cloud Native Application Bundle (CNAB).

How CNAB helps?

CNAB specification helps in unifying the management of multi-service, distributed applications across different toolchains into a single all-in-one packaging format. It defines resources that can be deployed to any combination of runtime environments and tooling including Docker Engine, Kubernetes, Helm, automation tools and cloud services.

Docker is the first company to implement CNAB specification and docker-app experimental tool for building, packaging and managing cloud-native applications

According to the Docker announcement, "Docker lets you package CNAB bundles as Docker images, so you can distribute and share through Docker registry tools including Docker Hub and Docker Trusted Registry. Additionally, Docker will enable organizations to deploy and manage CNAB-based applications in Docker Enterprise in the upcoming months."

Microsoft in their CNAB announcement noted below features that CNAB tries to address:

• Manage discrete resources as a single logical unit that comprises an app.
• Use and define operational verbs for lifecycle management of an app (install, upgrade, uninstall).
• Sign and digitally verify a bundle, even when the underlying technology doesn’t natively support it.
• Attest (or attach a signature to any moment in the lifecycle of that bundle) and digitally verify that the bundle has achieved that state to control how the bundle can be used.
• Enable the export of the bundle and all dependencies to reliably reproduce in another environment, including offline environments (IoT edge, air-gapped environments).
• Store bundles in repositories for remote installation.

Snip from duffle.sh

Microsoft has also released a tool called Duffle. It is an open source CNAB client reference implementation. Duffle client can help in installing, upgrading, and uninstalling CNAB bundles. Its other useful capabilities include creating new bundles, cryptographically signing bundles with PGP, and verify their integrity.

CNAB draft specification is available at cnab.io