Language Selection

English French German Italian Portuguese Spanish

Server

Running Deep Learning Models On Intel Hardware? It's Time To Consider A Different OS

Filed under
GNU
Linux
Server
Hardware

Firstly, Intel has done extensive work to make the Xeon family of processors highly optimized for AI. The Intel Xeon Scalable processors outsmart GPUs in accelerating the training on large datasets.

Intel is telling its customers that they don’t need expensive GPUs until they meet a threshold. Most of the deep learning training can be effectively done on CPUs that cost a fraction of their GPU counterparts.

Beyond the marketing messages and claims, Intel went onto prove that their deep learning stack performs better than NVIDIA GPU-based stack. Recently, Intel published a benchmark to show its leadership in deep learning. Intel Xeon Scalable processers trained a deep learning network with 7878 images per second on ResNet-50 outperforming 7844 images per second on NVIDIA Tesla V100.

Intel’s performance optimization doesn’t come just from its CPUs. It is delivered by a purpose-built software stack that is highly optimized at various levels. From the operating system to the TensorFlow framework, Intel has tweaked multiple layers of software to deliver unmatched performance.

To ease the process of running this end-to-end stack, Intel has turned to one of its open source projects called Clear Linux OS. Clear Linux project was started as a purpose-built, container-optimized, and lightweight operating system. It was started with the premise that the OS running a container doesn’t need to perform all the functions of a traditional OS. Container Linux, the OS developed by CoreOS (now a part of Red Hat) followed the same philosophy.

Within a short span, Clear Linux gained popularity among open source developers. Intel kept improving the OS, making it relevant to run modern workloads such as machine learning training jobs, AI inferencing, analytics and edge computing.

Read more

Also: Intel Core i9 9900KS Allowing 5.0GHz All-Core, Icelake News Coming This Week

Servers: SUSE, Red Hat/IBM and Kubernetes/Containers

Filed under
Red Hat
Server
SUSE
  • SuSE storage spins-up Ceph

    Open source software platform company SuSE has announced SUSE Enterprise Storage 6, a software-defined storage solution powered by Ceph technology.

    Many would argue that storage on its own is snorage (i.e. enough to send you to sleep), but software -defined storage does at least drive us forward into the realm of the software developer.

    By way of a reminder, software -defined storage is a way of managing data storage resources and functionality that is essentially uncoupled from (i.e. has no underlying physical dependencies) the actual hardware resources that offer up the amount of storage being used.

  • IBM Open Sources Razee CD Tool to Support Mega Kubernetes Scaling

    IBM open sourced its Razee continuous delivery (CD) tool that allows developers to manage applications in their Kubernetes-based cluster deployments. The move also continues to bolster IBM’s push into the Kubernetes space.

    Razee consists of two parts: Kaptain, which are components that handle the multi-cluster deployments; and RazeeDash, which is basically the control panel.

    The Kaptain component within Razee provides a pull-based deployment model that supports self-updating clusters. This helps in generating inventory and scripts that describe actions for each cluster or each application running in a Kubernetes environment.

  • Red Hat Open Sources 3scale Code

    Red Hat has completed open sourcing the API management software of 3scale, the company it bought in June 2016 for an undisclosed sum, saying it has been working on the project for the past three years.

    The company’s full code base has been released under the permissive Apache Software License (ASL) 2.0 licence, with the open sourcing process “much more than throwing code over the wall”, Red Hat said.

    In a short post by the company’s David Codelli on Thursday, he noted: “When Red Hat acquires 3scale it was only a matter of time until it would be open sourced in some fashion. “But the process isn’t instantaneous.”

  • Digital Ocean’s Kubernetes service is now generally available

    Like any ambitious cloud infrastructure player, Digital Ocean also recently announced a solution for running Kubernetes  clusters on its platform. At KubeCon + CloudNativeCon Europe in Barcelona, the company today announced that Digital Ocean Kubernetes is now generally available.

    With this release, the company is also bringing the latest Kubernetes release (1.14) to the platform, and developers who use the service will be able to schedule automatic patch version upgrades, too.

  • Serverless and containers: Two great technologies that work better together

    Cloud native models using containerized software in a continuous delivery approach could benefit from serverless computing where the cloud vendor generates the exact amount of resources required to run a workload on the fly. While the major cloud vendors have recognized this and are already creating products to abstract away the infrastructure, it may not work for every situation in spite of the benefits.

    Cloud native, put simply, involves using containerized applications and Kubernetes  to deliver software in small packages called microservices. This enables developers to build and deliver software faster and more efficiently in a continuous delivery model. In the cloud native world, you should be able to develop code once and run it anywhere, on prem or any public cloud, or at least that is the ideal.

Databases: Couchbase, Databricks, DataStax and Cloudwashing

Filed under
Server
OSS
  • Couchbase pumps ups ‘Autonomous Operator’ function

    NoSQL database company Couchbase has moved off the sofa (the firm is no couch potato, get it?) and come forward with new features aligned to allow ‘deployers’ to build (and scale) applications.

    The new version of Couchbase Autonomous Operator is enables Kubernetes-based application deployers to bring in a database ready for flexibility and micro-services.

  • Databricks Open Sources Delta Lake to Make Data Lakes More Reliable

    Databricks recently announced open sourcing Delta Lake, their proprietary storage layer, to bring ACID transactions to Apache Spark and big data workloads. Databricks is the company behind the creators of Apache Spark, while Delta Lake is already being used in several companies like McGraw Hill, McAffee, Upwork and Booz Allen Hamilton.

    Delta Lake is addressing the heterogeneous data problem that data lakes often have. Ingesting data from multiple pipelines means that engineers need to enforce data integrity manually, throughout all the data sources. Delta Lake can bring ACID transactions to the data lake, with the strongest level of isolation applied, serializability.

  • DataStax CEO Bosworth : accelerating development on (and in) the cloud

    DataStax CEO Billy Bosworth started out as a database administrator (DBA), so one would hope that he knows how to build, compile, manage and deploy in all senses of those terms, right?

  • DataStax details road to Apache Cassandra future

    DataStax closed out the final day of its ‘Accelerate 2019’ conference by focusing on a selection of platform-level developments including its community development stream.

  • DataStax has stars in its eyes over Constellation, its latest tweak on Apache Cassandra

    DataStax, the business built around the Apache Cassandra open source database, is creating a new system-as-a-cloud service using the platform.

  • Open source is big in databases, but cloud is bigger [Ed: Adobe's Asay calls everything "cloud" (I wonder if he even understands what that means). Had Aasay ever set up a database or written a single line of code (he's a lawyer), he'd know "cloud" just means anything including a database hosted where you have no real control over it. It's outsourcing.]

Why Windows Containers Are Less Attractive Than Linux Containers

Filed under
Linux
Server

The fact that you can run Docker containers on Windows as well as Linux is amazing. Yet, I sometimes struggle to see a clear use case for Windows containers. Compared to Linux containers, there are fewer obvious reasons to run containers on Windows.

I know that’s a somewhat controversial statement, so let me walk through the various reasons why Windows containers are much less attractive than Linux containers.

Read more

Also: Streamlining Software Development and Distribution with Containers [Ed: Paid-for SPAM from EMC. “Buying the news”… the new “biz model”? Companies literally buying not only the narratives but also the space and the staff?]

Servers: Kubernetes, Microservices, Containers and SUSE's Enterprise Storage 6

Filed under
Server
SUSE
  • Is bare Kubernetes still too messy for enterprises?

    Kubernetes is touted as a computing cure-all, fixing up multicloud networking to data mobility. The open-source platform for orchestrating containers (a virtualized method for running distributed applications) may or may not be the panacea it’s hyped up to be. What is certain is that user-ready Kubernetes isn’t as easy as it sounds, so customers should shop carefully for a provider.

    Enterprise users of Kubernetes and containers may not guess just how many moving parts are under the covers. There are a ton of tiny pieces that have to line up just so in order for them to work, according to Mark Shuttleworth (pictured), founder and chief executive officer of Canonical Ltd. He likens these technologies to carefully constructed “fictions.”

  • Data as a microservice: Distributed data-focused integration

    Microservices is the architecture design favored in new software projects; however, getting the most from this type of approach requires overcoming several previous requirements. As the evolution from a monolithic to a distributed system takes place not only in the application space but also at the data store, managing your data becomes one of the hardest challenges. This article examines some of the considerations for implementing data as a service.

  • Container Adoption Shoots Up Among Enterprises In 2019: Survey

    Majority of IT professionals now run container technologies, with 90 percent of those running in production and 7 in 10 running at least 40 percent of their application portfolio in containers — an impressive increase from two years ago, when just 67 percent of teams were running container technologies in production. According to the joint 2019 Annual Container Adoption Survey released by Portworx and Aqua Security, enterprises have started making bigger investments in containers.

    In 2019, nearly one in five organizations is found to be spending over $1 million annually on containers (17%) as compared to just four percent in 2016.

  • SUSE Rolls Out Enterprise Storage 6

    SUSE has announced the latest version of its software-defined storage solution powered by Ceph technology. With SUSE Enterprise Storage 6, IT organizations can adapt to changing business demands. They may also reduce IT operational expense with new features focused on containerized and cloud workload support, improved integration with public cloud, and enhanced data protection capabilities, SUSE said.

PostgreSQL 12 Beta 1 Released!

Filed under
Server
OSS

The PostgreSQL Global Development Group announces that the first beta release of PostgreSQL 12 is now available for download. This release contains previews of all features that will be available in the final release of PostgreSQL 12, though some details of the release could change before then.

In the spirit of the open source PostgreSQL community, we strongly encourage you to test the new features of PostgreSQL 12 in your database systems to help us eliminate any bugs or other issues that may exist. While we do not advise you to run PostgreSQL 12 Beta 1 in your production environments, we encourage you to find ways to run your typical application workloads against this beta release.

Read more

Also: PostgreSQL 12 Beta Released With Performance Improvements

Server: RHEL 8, Docker, Containers and “Natty Narwhal” Still Alive

Filed under
Server
  • What Is New Features In Red Hat Enterprise Linux 8 (RHEL 8)?

    Red Hat Enterprise Linux 8 (RHEL 8) was released on 2019-05-07. The first beta was announced on 14 November 2018.

    There are lot of opportunities in Devops to improve the IT infra to next generation so, Red Hat is more focus on this area.

    To give more space to developers, they were added a lot of tools that developer-friendly capabilities.

    IT infra has been moved from Physical servers to Virtual servers. Now, we are migrating from Virtual servers to Containers & Kubernetes.

  • Distribution Release – Red Hat Enterprise Linux 8
  • The future of Docker containers

    Michael Crosby is one of the most influential developers working on Docker containers today, helping to lead development of containerd as well as serving as the Open Container Initiative (OCI) Technical Oversight Chair. At DockerCon 19, Crosby led a standing-room-only session, outlining the past, present and — more importantly — the future of Docker as a container technology. The early history of Docker is closely tied with Linux and, as it turns out, so too is Docker's future.

    Crosby reminded attendees that Docker started out using LXC as its base back in 2013, but it has moved beyond that over the past six years, first with the docker-led libcontainer effort and more recently with multi-stakeholder OCI effort at the Linux Foundation, which has developed an open specification for a container runtime. The specification includes the runc container runtime which is at the core of the open source containerd project that Crosby helps to lead. Containerd is a hosted project at the Cloud Native Computing Foundation (CNCF) and is one of only a handful of projects that, like Kubernetes, have "graduated", putting it in the top tier of the CNCF hierarchy in terms of project stability and maturity.

  • Why Organizations Want A Kubernetes-based Container Service for Cloud
  • The technology that powers IG

    Instagram run in Ubuntu Linux 11.04 (“Natty Narwhal”). Engineers found past variants of Ubuntu had a wide range of flighty solidifying scenes on EC2 under high traffic, yet Natty has been strong and exceptional

Red Hat, Fedora and SUSE/OpenStack

Filed under
Red Hat
Server
SUSE
  • Rook-Ceph storage Operator now on OperatorHub.io

    We are excited to announce the addition of the Rook-Ceph storage Operator to OperatorHub.io. Operators are design patterns that augment and implement common day one and day two activities with Kubernetes clusters, simplifying application deployments and empowering developers to focus on creation versus remediation. The Rook-Ceph Operator is an upstream effort that Red Hat is leading and is using as part of its work towards Red Hat OpenShift Container Storage 4.

    Developing and deploying cloud-native applications at scale can be complex and challenging. The new Rook-Ceph storage Operator is designed to automate the packaging, deployment, management, upgrading, and scaling of Ceph clusters that provide persistent storage to stateful applications as well as infrastructure services (logging, metrics, registry) in Kubernetes clusters. The release of Rook’s Ceph Operator augments Kubernetes scheduling with a complement of stateful storage services including block, filesystem and object storage.

  • Red Hat Satellite 6.4.3 has been released

    Red Hat Satellite 6.4.3 is generally available. The main drivers for the 6.4.3 release are a Request for Feature Enhancement (RFE) for capsule syncing control as well as general stability fixes.

    The capsule syncing control feature enables the user to have control over when capsule syncs occur. Traditionally the capsule sync occurs automatically after a content view is updated, but some customers may want more granular control over when the synchronization occurs. Satellite 6.4.3 introduces a new setting in Administer —> Settings —> Content —> Sync Capsules after Content View promotion.

  • Contributors are Empowered When They Know the Process

    There is a saying in the legal profession that you should never ask a question you don’t already know the answer to. Despite how this sounds, it is actually a rule most people follow in life. This is the source of that feeling you get when you’re too scared to raise your hand and ask a question. In Open Source we need to make sure that contributors feel like they already “know” the answers, so they will feel confident in making the request.

    As a university lecturer, I always encouraged my students to first think about what they thought the answer was and then ask the question. In some cases, I encouraged them to actually write down what they thought the answer was. In this way, they could judge both their skills and their ability to grow based on what the answer turned out to be. It created an additional feedback loop.

  • Alisha and Shraddha: Positive feedback loops in Fedora

    This post is the second introduction to the Fedora Summer Coding interns Class of Summer 2019. In this interview, we’ll meet Alisha Mohanty and Shraddha Agrawal, who are both working on Fedora Happiness Packets to promote positive feedback loops in the Fedora community.

  • The OpenStack User Survey is now open

    The 2019 OpenStack User Survey is now open and waiting for your input. Whether you’re a user of OpenStack, or an operator utilising it to power your offerings, the OpenStack Foundation (and the rest of the community) want to hear about your usage.

    2018 saw the 11th OpenStack User Survey unveiled at the Berlin OpenStack Summit, giving some fantastic insight into how and where people are using OpenStack across 63 different countries. Usage in Asia surged dramatically in 2018, with 48% of respondents based in that continent, with Europe 2nd at 26% and North America 3rd with 20% of respondents.

Lokomotive: Production-ready Kubernetes distribution with Linux technologies

Filed under
GNU
Linux
Server

Kinvolk.io, a software consultancy specializing in cloud tech for Linux, announced their Kubernetes distribution Lokomotive on May 17, 2019. Under an open source license, Lokomotiv aims towards production-ready performance and a secure, stable Kubernetes distribution.

According to Kinvolk, the fully supported release and commercial support with lokoctl and Lokomotive Components pulls into the station sometime this summer. For now, let us have a look at what’s under the hood and the project’s goals.

Read more

Also: Kubernetes, Cloud Native, and the Future of Software

Databases: NoSQL, EnterpriseDB and RavenDB

Filed under
Server
OSS
  • Top Open source NoSQL database programs

    NoSql, it stands for Not Only SQL, refers to the non-relational database. The next generation database mainly addresses several key points: non-relational, distributed, open source, and horizontally scalable. The non-relational database has developed very rapidly due to its own characteristics. The NoSQL database was created to solve the challenges brought by the multiple data types of large-scale data collection, especially the big data application problem. It also supports easy replication, simple APIs, final consistency (non-ACID), and large data. It is stored by us with the most key-values, and of course other document types, column stores, graph databases, XML databases, and so on. Here are some top available NoSQL database programs in Open source or free category.

  • We need a new type of open source event - here's why

    Open source events tend to focus on developers, this needs to change says EnterpriseDB's Jan Karremans

  • RavenDB Adds Pull Replication and Distributed Online Counters to Its Open Source NoSQL Document Database Offering
Syndicate content

More in Tux Machines

Linux 5.2-rc2

Hey, what's to say? Fairly normal rc2, no real highlights - I think most of the diff is the SPDX updates. Who am I kidding? The highlight of the week was clearly Finland winning the ice hockey world championships. So once you sober up from the celebration, go test, Linus Read more Also: Linux 5.2-rc2 Kernel Released As The "Golden Lions"

Audiocasts/Shows: Linux Action News, Linux Gaming News Punch, Open Source Security Podcast and GNU World Order

Review: Red Hat Enterprise Linux 8.0

My experiment with RHEL 8 got off to a rough start. Going through the on-line registration process produced some errors and ended up with me getting the wrong ISO which, in turn, resulted in some confusion and delays in getting the distribution installed. Things then began to look up as RHEL 8 did a good job of detecting my system's hardware, registered itself without incident and offered good performance on physical hardware. I was particularly pleased that the distribution appears to detect whether our video card will work well with Wayland and either displays or hides Wayland sessions in response. I did have some trouble with the GNOME Classic Wayland session and GNOME Shell on X.Org was a bit sluggish. However, the Classic session on X.Org and GNOME Shell on Wayland both worked very well. In short, it's worthwhile to explore each of the four desktop options to see what works best for the individual. The big issues I ran into with RHEL were with regards to software management. Both GNOME Software and the Cockpit screen for managing applications failed to work at all, whether run as root or a regular user. When using the command line dnf package manager, the utility failed to perform searches unless run with sudo and occasionally crashed. In a similar vein, the Bash feature that checks for matching packages when the user types a command name it doesn't recognize does not work and produces a lengthy error. There were some security features or design choices that I think will mostly appeal to enterprise users, but are less favourable in home or small office environments. Allowing remote root logins by default on the Workstation role rubs me the wrong way, though I realize it is often useful when setting up servers. The enforced complex passwords are similarly better suited to offices than home users. One feature which I think most people will enjoy is SELinux which offers an extra layer of security, thought I wish the Cockpit feature to toggle SELinux had worked to make trouble-shooting easier. I was not surprised that RHEL avoids shipping some media codecs. The company has always been cautious in this regard. I had hoped that trying to find and install the codecs would have provided links to purchase the add-ons or connect us with a Red Hat-supplied repository. Instead we are redirected through a chain of Fedora documentation until we come to a third-party website which currently does not offer the desired packages. Ultimately, while RHEL does some things well, such as hardware support, desktop performance, and providing stable (if conservative) versions of applications, I found my trial highly frustrating. Many features simply do not work, or crash, or use a lot of resources, or need to be worked around to make RHEL function as a workstation distribution. Some people may correctly point out RHEL is mostly targeting servers rather than workstations, but there too there are a number of problems. Performance and stability are provided, but the issues I ran into with Cockpit, permission concerns, and command line package management are all hurdles for me when trying to run RHEL in a server role. I find myself looking forward to the launch of CentOS 8 (which will probably arrive later this year), as CentOS 8 uses the same source code as RHEL, but is not tied to the same subscription model and package repositories. I am curious to see how much of a practical effect this has on the free, community version of the same software. Read more

GNOME 3.34 Revamps the Wallpaper Picker (And Fixes a Longstanding Issue Too)

The upcoming release of GNOME 3.34 will finally solve a long standing deficiency in the desktop’s background wallpaper management. Now, I’ve written about various quirks in GNOME wallpaper handling before, but it’s the lack of option to pick a random wallpaper from a random directory via the Settings > Background panel that is, by far, my biggest bug bear. Ubuntu 19.04 ships with GNOME 3.32. Here, the only wallpapers available to select via the Settings > Background section are those the system ships with and any top-level images placed in ~/Pictures — nothing else is selectable. So, to set a random image as a wallpaper in GNOME 3.32 I tend to ignore the background settings panel altogether and instead use the image viewer’s File > Set as background… option (or the similar Nautilus right-click setting). Thankfully, not for much longer! Read more