Industry Economy TECH GAME
Society Comfort AUTO MEDIA

[OSS 2018 Berlin] OpenStack Summit 2018 Berlin : Day 1 Keynotes

  • Tuesday, December 18, 2018, 11:25 pm
  • ACROFAN=Yong-Man Kwon
OpenStack Foundation (OSF) held 'OpenStack Summit Berlin' at Citycube Berlin, Germany from November 13 to 15. 'OpenStack Summit Berlin' was held under the theme of 'Open Infrastructure', with over 200 sessions and workshops in more than 35 projects with experts and representatives from over 50 countries.

‘OpenStack Summit Berlin’ addresses not only the cloud but also innovative technologies and cases of various types of ‘Open Infrastructure’ ranging from edge computing, CI/CD, AI, network virtualization (NFV). And more than 200 sessions and workshops were held in more than 35 open source projects. In particular, Germany, where the summit was held, is considered to be one of the most active areas where OpenStack is used. Large telecommunication companies, enterprises and public cloud service providers in Germany are actively utilizing the OpenStack-based environment, and it is also considered to be an active region in terms of community size and activity.

OpenStack Foundation has also released some new news through ‘OpenStack Summit Berlin’. First, with the current progress of four new ‘pilot projects’ which have been introduced at the OpenStack Summit in addition to the OpenStack project in the past year, the new project policy reflecting it was announced. Moreover, the application area of OpenStack has expanded to artificial intelligence and HPC, so various project updates for this were introduced. Meanwhile, the OpenStack Foundation also announced the plan to change the name of the Summit from 'OpenStack Summit' to 'Open Infrastructure Summit' from the next summit.

▲ Jonathan Bryce, executive director of OpenStack Foundation

▲ Now, OpenStack is based on all the forms of ‘Open Infrastructure’.

Jonathan Bryce, executive director of OpenStack Foundation, first introduced that OpenStack Summit has become a massive global event, and this summit is a summit held in Europe again after two years in Barcelona in 2016. And he stated that at the recent events, OpenStack showed ‘Open Infrastructure’ as an important message, and there is a need to think about what opportunities there are in terms of technology and community. In addition, 81% of respondents in various surveys answered that open source is expected to be more important in the coming years.

In addition, he mentioned ‘flexibility’ over the business model as a point that companies consider significant with considerations of open source, and this is consistent with the results of OpenStack Foundation’s user survey and the survey results from external agencies. This is due to the wide variety of options that the open source ecosystem brings to, and thanks to this, the OpenStack is used not only in many enterprise private clouds, but also in over 75 public cloud data centers around the world, which is a reflection of flexibility over the business.

At the same time, Bryce talked about the advantages provided by open source. It protects against vendor lock-in and allows to work across multiple environments, and it also offers advantages in integration and implementation. With open source, it is possible to innovate, in some instances, faster than the market. On top of that, ‘Open Infrastructure’ reflects requested changes in infrastructure over the last few years, such as the emergence of software-defined base, new workloads like Cloud Native and AI. And ‘Open Infrastructure’ combines with open source based technologies, so it has various features including flexibility, innovation, interoperability, and collective intelligence of communities.

According to Bryce, the OpenStack project is one of the most vibrant projects around the world, as it had 70,000 commits over the last year, and 182 changes per day in the terms of Rocky release. Now, OpenStack became a software-defined based ‘base infrastructure’ that connects new environments such as containers, and new applications such as AI, machine learning, and edge computing, in addition to VMs and data centers, and encompasses network environments of service providers. Therefore, as the OpenStack Summit covers not just OpenStack, but also a variety of projects, the next summit will be called ‘Open Infrastructure Summit’.

▲ In the case of Oerlikon, the manufacturing environment was optimized after collecting data from machines based on OpenStack.

▲ Kubernetes and OpenStack are now going together.

In the case of Oerlikon's ManMade Fibers segment, OpenStack was used to create an infrastructure environment for integrating legacy, container and third-party workloads. In the production of yarn at a fast speed, a little failure that occurs momentarily creates a big problem. So, in order to optimize the process, Oerlikon digitized the combination of machine learning, process integration and automation, and introduced AI technology based on edge computing. This allows 69,120 sensors on 120 machines of the yarn production for the seat belts located geographically far from each other to sense data, view fault graphs, make 350,000 online decisions per day, and optimize models for data collection and training.

Nick Barcet, senior director of OpenStack Product Management at Red Hat, stressed that OpenStack is a good tool for ‘Open Infrastructure’ that spans public, private cloud, service providers, NFV, etc. In addition, with many new workload including containers, AI/ML, big data, HPC and direct hardware access, a variety of deployment models, such as containers in VMs, containers on bare metal, and multi-tenant BMaaS are appearing. In these various cases, even OpenStack cannot respond to all of them with OpenStack’s internal technologies.

Therefore, when OpenStack and Kubernetes meet, they can be seen as a complement to each other, rather than a competitive one, and Barcet stated that the point where they meet should be the location of ‘Open Infrastructure’. This is also because OpenStack can provide an infrastructure environment for deploying Kubernetes in large environments. And Red Hat has announced ‘Red Hat OpenStack Platform 14’, based on the release of Rocky, at the Berlin Summit. As a feature of the release, reinforcement of OpenStack-based OpenShift capabilities was introduced. This version will be released to public in December.

Red Hat also introduced a new Virtual Central Office (VCO) for the edge environment. It is a flexible open pluggable framework that simplifies deployment and management of edge-based services, including OPNFV, OpenStack, Kubernetes, ODL, Ansible, etc. Besides, Red Hat mentioned ‘dynamic migration’, which brings workloads of VMware to an OpenStack-based environment with a GUI-based interface, and introduction of 3-month release cycle, which is faster that the general 6-month cycle for introducing latest functions and accelerating innovations.

▲ Sweden SBAB took advantage of active automation and 'micro services' on OpenStack basis.

▲ Hybrid service with public cloud stack on individual hardware was introduced.

In the case of SBAB Bank in Sweden, it was introduced that change in the IT architecture that supports tasks was essential for speeding up important tasks. To this end, many changes were made in the architecture as follows. By dismantling the ‘monolithic’ approach in which all of the existing components were unified, it was possible to focus on speed and customer-facing applications through micro-services based on OpenStack platform. In addition, by applying automation to every possible step, waiting time has been minimized. There was also a great effect in reducing the downtime caused by modification in the deployment, which was a problem in the existing environment.

Deutsche Telekom stated that OpenStack is large and sometimes complex, but it is a very flexible platform for all aspects of the cloud. Deutsche Telekom showed demonstration of changing common images into the images with the feel of famous artists, using AI technology serviced by OpenStack-based infrastructure. And it is said that OpenStack-based ‘Open Telekom Cloud’ is now being transformed from infrastructure to platform. In terms of supporting containers, Deutsche Telekom integrates the functions related to the operation of the Kubernetes with OpenStack-based platform and provides enhanced data protection through ‘Kata Container’.

Moreover, ‘Day-2-Operations’, which provides professionalism in a reliable operation, offers the best user experience through accurate capacity management based on the accurate analysis and monitoring of the infrastructure. Besides, in aspect of security, the concept of ‘hybrid’ was introduced as taking advantages of both the benefits of a public cloud and the security of separated hardware. It is a full managed hybrid cloud running on separated hardware with OpenStack environment based, and it can be installed and operated in a customer’s data center.

▲ Tilman Schulz, lead, Private Cloud Services, Volkswagen AG, and Lauren Sell, vice president of marketing and community services, OpenStack Foundation

▲ ‘Fusion Cloud’ of Huawei suggests a full-stack cloud solution.

Volkswagen group is a long-time user of OpenStack and has been successfully running an OpenStack-based cloud infrastructure since 2015. It introduced that the recent changes in the automobile industry have been extended not only to production and service provision of automobiles, but also to ‘mobility services’, and it was pointed out that it is important to respond cost efficiently and effectively. And in response to this demand, OpenStack-based environment has high performance, reliability, manageability, cost-efficiently, and flexibility and expandability to cope with change. For future tasks, Volkswagen group mentioned performance of upgrades that has less impact on the usage environment.

Frank Weyns, chief system architect of Cloud Computing Western Europe in Huawei Technologies Co., Ltd. stressed that OpenStack is now the perfect ‘Open Infrastructure’ platform to glue together new technologies, software, hardware, etc. He also highlighted that OpenStack is useful not only in the cloud but also in IoT, and it plays an important role in linking technologies for ‘digital transformation’. What’s more, ‘Full Stack Cloud’ is a model where resource convergence, data intelligence, and service innovation gather together to create the revolution of the general cloud, and the core technologies for this are AI, big data, cloud, DevOps, and edge computing.

The introduction of AI technology in the cloud reduces mistakes by human in more complex cloud environments and implements ‘self-healing’ by automatically responding to abnormal situations. Furthermore, in the big data section, he explained that various databases can be combined as needed on an OpenStack basis. In case of the cloud, with a full-stack private cloud that has a feeling of public cloud, IT operations will be simplified, and the service roll-out periods will be dramatically reduced through OpenStack and Kubernetes’s automated orchestrations. Besides, in the edge computing section, the viewpoint of intelligence, mobilization, edge automation, and edge-cloud integration were introduced.

▲ The addition of functionality for Ironic’s infrastructure operators should be the results of project contributions from the operators.

▲ A case of OpenStack infrastructure upgrade without downtime was shared.

Meanwhile, OpenStack is now being operated in more than 10 million cores worldwide, and well utilized in a production environment with a variety of new use cases are coming out. Also, the contribution to development is happening all over the world, and in the latest ‘Rocky’ cycle, there were 32,800 changes. A noteworthy trend in terms of contributions is that the percentage of ‘operator’ participation is increasing. With the increase of collaboration with operators in the project team and the increase of operator’s participation in PTG, it was explained that 25% of all commits are contributed by operators.

At the same time, it was introduced that the latest update of Ironic allows operators to make changes to their BIOS settings with software tools to make it easier to deploy OpenStack’s bare metal. In addition, the new hardware architecture supports ARM architecture as well as x86, and the Cyborg project enabled direct use of GPU, FPGA to accelerate workloads of AI and machine learning. Besides, the size of the deployment has been further reduced to make it more versatile, and Magnum now became the official Kubernetes installer in terms of infrastructure support over compute.

It was introduced that OVH was mentioned as the operating example of an OpenStack-base public cloud. OVH operates global scale bare metal, private and public cloud at 28 data centers and 360,000 servers. And OpenStack is not only the main layer of their public cloud, but also the abstract layer of bare metal and internally the foundation of its own infrastructure. In an OpenStack-based environment, they complete an initial configuration of a new region in about eight hours, and successfully completed a region-level version upgrade where up to 50,000 instances are operated in four hours without downtime with a year of R&D. In addition, NVIDIA Tesla-based GPU cloud was implemented as a hardware direct connection in an OpenStack basis via PCI passthrough.