Category Archives: server

IBM announces Linux mainframe app development cloud

IBM is trying to keep mainframes relevant in the cloud era

IBM is trying to keep mainframes relevant in the cloud era

IBM is open sourcing a large set of Linux mainframe code and launching the LinuxONE Developer Cloud, a cloud-based platform for developers to create applications for a Linux server based on the mainframe.

The LinuxONE Developer Cloud, which will be deployed in select IBM datacentres globally, will provide developers access to a cloud-based development, piloting and testing environment for Linux-based mainframe workloads.

The move coincides with the company’s launch of a portfolio of Linux mainframe services, called LinuxONE, that IBM says are optimised to run cloud-native workloads like Dockerized apps and NoSQL databases.

“Fifteen years ago IBM surprised the industry by putting Linux on the mainframe, and today more than a third of IBM mainframe clients are running Linux,” said Tom Rosamilia, senior vice president, IBM Systems.

“We are deepening our commitment to the open source community by combining the best of the open world with the most advanced system in the world in order to help clients embrace new mobile and hybrid cloud workloads. Building on the success of Linux on the mainframe, we continue to push the limits beyond the capabilities of commodity servers that are not designed for security and performance at extreme scale,” Rosamilia said.

As part of the move the company is contributing tens of thousands of lines of code to the recently created Open Mainframe Project, formed by the Linux Foundation to optimise Linux deployments on mainframes.

“Linux on the mainframe has reached a critical mass such that vendors, users and academia need a neutral forum where they can work together to advance Linux tools and technologies and increase enterprise innovation,” said Jim Zemlin, the Linux Foundation executive director.

“The Open Mainframe Project is a direct response to the demands of Linux users and the supporting open source ecosystem to address unique features and requirements built into mainframes for security, availability and performance,” Zemlin said.

IBM, partners score 7 nm semiconductor breakthrough

IBM, Samsung and Globalfoundries claimed a 7nm semiconductor breakthrough

IBM, Samsung and Globalfoundries claimed a 7nm semiconductor breakthrough this week

Giving Moore’s Law a run for its money, IBM, Globalfoundries and Samsung claimed this week to have produced the industry’s first 7 nanometre node test chip with functioning transistors. The breakthrough suggests a massive jump in low-power computing power may be just on the horizon.

IBM worked with Globalfoundries, the chip division it divested in October last year, and Samsung specialists at the SUNY Polytechnic Institute’s Colleges of Nanoscale Science and Engineering (SUNY Poly CNSE) to test a number of silicon innovations developed by IBM researchers including Silicon Germanium (SiGe) channel transistors and Extreme Ultraviolet (EUV) lithography integration at multiple levels, techniques developed to accommodate the changing nature of the rules of physics that apply at such small scales.

Most microprocessors found in servers, desktops and laptops today are developed with 22nm and 14nm processes, and mobile processors are increasingly being developed with 10nm processors, but IBM claims the 7nm process developed by the semiconductor alliance enjoys 50 per cent area scaling improvements over today’s most advanced chips.

IBM said the move could result in the creation of a chip small and powerful enough to “power everything from smartphones to spacecraft.”

“For business and society to get the most out of tomorrow’s computers and devices, scaling to 7nm and beyond is essential,” said Arvind Krishna, senior vice president and director of IBM Research. “That’s why IBM has remained committed to an aggressive basic research agenda that continually pushes the limits of semiconductor technology. Working with our partners, this milestone builds on decades of research that has set the pace for the microelectronics industry, and positions us to advance our leadership for years to come.”

The companies also said the chips have a 50 per cent power-to-performance improvement over existing server chips, and could be used in future iterations of Power architecture, IBM’s mainframe architecture which it open sourced in a bid to improve its performance for cloud and big data workloads.

IBM has in recent months ramped up silicon-focused efforts. The company is partnering with SiCAD to offer a cloud-based high performance services for electronic design automation (EDA) which the companies said can be used to design silicon for smartphones, wearables and Internet of Things devices. Earlier this month the company also launched another OpenPower design centre in Europe to target development of high performance computing (HPC) apps based on the Power architecture.

Mirantis gets into the OpenStack converged infrastructure game

Mirantis is teaming up with Dell and Juniper to create converged infrastructure solutions

Mirantis is teaming up with Dell and Juniper to create converged infrastructure solutions

Mirantis unveiled plans to work with a range of vendors to deliver OpenStack-based converged infrastructure solutions for enterprises.

The Mirantis Unlocked Appliances are built by Rack Partners and come pre-validated by Mirantis with its own commercial distribution of OpenStack and pre-integrated by Certified Rack Partners.

This is Mirantis’ first big foray into the converged infrastructure area – a segment dominated by Dell and HP, which offers its own HP Helion OpenStack-based converged infrastructure solutions.

“The architecture Redapt designed with Mirantis, with Mirantis OpenStack at the core, solves for the complexity commonly associated with OpenStack,” said Josh Lindenbaum, vice president of business & corporate development, Redapt. “Redapt will work closely with customers to assemble, deliver and install Mirantis Unlocked Appliances that will arrive in the datacentre ready to plug and play.”

Its inaugural offering is tailored specifically for cloud-native applications, though it said it plans to release a broader portfolio of converged infrastructure solutions as it partners with more hardware and software vendors.

The first configuration will be built using server and networking technology provided by Dell (PowerEdge R630 servers for compute and foundation nodes and Dell PowerEdge R730xd for storage) and Juniper Networks (QFX5100s as the data path and a Juniper EX3300 for management) respectively, with configurations ranging from six compute nodes and 12 TB of storage to a full rack comprised of 24 compute nodes and 24 TB.

“About 20 percent of infrastructure is consumed through the appliance form factor because it is extremely easy to set up and operate,” said Alex Freedland, Mirantis president and co-founder.

“Mirantis Unlocked Appliances combines this ease of use with the openness and flexibility of OpenStack, delivered as a cloud-in-a-box. Our first appliance focuses on the most common OpenStack use case – developing cloud-native applications – and will be built and shipped by Certified Rack Partners across the ecosystem,” Freedland said.

Alibaba throws its weight behind ARM architecture standards

Alibaba is joining the Linaro Group, an organisation which aims to eliminate software fragmentation within ARM-based  environments

Alibaba is joining the Linaro Enterprise Group, an organisation which aims to eliminate software fragmentation within ARM-based environments

Chinese e-commerce and cloud giant Alibaba announced it has joined the Linaro Enterprise Group (LEG), a group of over 200 engineers working on consolidating and optimising open source software for the ARM architecture.

Linaro runs a number of different ARM-based initiatives  aimed at cultivating software standards for ARM chips for networking, mobile platforms, servers and the connected home. It mainly targets upstream development but also aims to coordinate work that helps reduce “costly low level fragmentation.”

More recently the organisation launched a working group focused specifically on developing software standards for ARMv8-A 64-bit silicon, architecture a number of server vendors and ODMs have started adopting in their portfolios in a bid to test the ARM-based cloud server market.

Alibaba, which operates six cloud datacentres – mostly in China – and recently expanded to the US, said it will collaborate with a range of companies within LEG to optimise the ARMv8-A software platforms.

“Alibaba Group’s infrastructure carries the world’s largest e-commerce ecosystem, in addition to China’s leading cloud services,” said Shuanlin Liu, chief architect of Alibaba Infrastructure Service.

“We need the best technical solutions as we step into the DT (data technology) era. Hence, we’re investing heavily in the innovation of a wide range of technologies, including the ARM architecture. We will continue to work closely with partners to accelerate the development and growth of the ecosystem,” Liu said.

Alibaba said the move may help it deliver cloud services that have been workload-optimised right down to the chip, and help lower the TCO; lower energy usage and higher density are two leading characteristics driving interest in ARM for cloud datacentres. But due in part to x86′s dominance in the datacentre there is a conspicuous lack of ARM-based software standards and workloads, which is what LEG

“As one of the world’s largest cloud operators, Alibaba is continually pushing technology boundaries to efficiently deploy new services at a massive scale,” said Lakshmi Mandyam, director, server systems and ecosystems, ARM. “Their collaboration with the ARM ecosystem will accelerate and expand open source software choices for companies wishing to deploy ARMv8-A based servers. We welcome Alibaba’s participation in Linaro and the new dimension it will bring to an already vibrant community.”

The past couple of years have seen a number of large cloud service providers flirt with the prospect of switching to ARM architecture within their datacentres, most notably Amazon. The latest move signals Alibaba is interested in moving in that direction, or at least  signal to vendors it’s willing to do so, but it may be a while before we see the cloud giant roll out the ARM-based servers within its datacentres.

DataCentred adds ARM 64-bit to OpenStack cloud

DataCentred is adding ARM-based OpenStack services to its public cloud portfolio

DataCentred is adding ARM-based OpenStack services to its public cloud portfolio

Manchester-based cloud services provider DataCentred has added ARM AArch64-based servers to its OpenStack-based public cloud platform, a product of its recently announced partnership with Codethink. The company’s head of cloud services told BCN the company is responding to customer demand for putting ARM-based workloads in the cloud.

As part of the move the ARM AArch64 architecture, which allows 32-bit and 64-bit processes to be executed alongside one another, will be added to the company’s OpenStack-based public cloud offering; the company said it will run the platform on HP M400 ARM hardware, and give customers access to Intel and ARM architectures alongside one another within an OpenStack environment.

DataCentrerd said the move will help drive down the cost of data centre operation and of the cost of virtualised instances within a customer’s service framework.

“We are thrilled to be the first OpenStack public cloud operator to feature 64-bit ARM instances. This breakthrough is testament to the considerable skill and expertise of our OpenStack cloud development team.  This is probably the first example of Moonshot AArch64 running in Europe outside of HP’s development labs, and certainly the first example of generally available Moonshot backed AArch64 instances in an OpenStack public cloud anywhere in the world,” said Mike Kelly, chief executive and founder of DataCentred.

“We know that ARM themselves are pleased to hear of this development, as a real world deployment. OpenStack is one of the big success stories for Open Source software, and is likely to be the environment through which enterprise migrates, in a vendor neutral way, to take advantage of elastic cloud compute,” Kelly added.

Matt Jarvis, head of cloud computing at DataCentred told BCN there’s currently a scarcity of ARM in the cloud.

“This deployment is driven by customer demand – we have both new customers who want to access ARM64 on-demand, and existing customers who we’ve been talking to about proof of concept ARM workloads for some time,” Jarvis said.

“There is significant interest from the worldwide community of technology companies currently working with ARM hardware to have access to develop platforms on-demand, along with specific vertical market interest in ARM as part of a longer term technical strategy targeting reduction in operating cost due to power savings,” he added.

ARM for compute seems to be fairly scarce in the cloud world, though it’s clear that OpenStack incumbents are looking to bring the software platform to all kinds of architecture beyond x86. Oracle is looking to marry SPARC and OpenStack while IBM and Rackspace are both working towards getting the open source software platform working on OpenPower.

DataCentred said it plans to move the Moonshot-powered cloud service into production sometime later this year.