扫一扫微信二维码

OpenStack悉尼峰会议题投票火热进行中 发布时间:2017-08-03

2017年OpenStack 悉尼峰会来临在即,目前大会筹备已经进入议题征选阶段,九州云作为中国OpenStack最早的专业技术公司也积极参与其中,并向大会组委会提交了多个精彩议题。现在所有峰会议题已在社区开放,九州云诚邀各位关注OpenStack的朋友为我们的议题投上宝贵的一票,让本次悉尼峰会参会者收获更多更为精彩的内容。
 
议题及投票链接:
Topic 1:OpenStack at SJTU: Predictive Data Mining in Clinical Medicine with Dynamical HPC
投票链接:
https://www.openstack.org/summit/sydney-2017/vote-for-speakers/#/19384
Topic 2:Composable Infrastructure ? Try Valence!
投票链接:
https://www.openstack.org/summit/sydney-2017/vote-for-speakers/#/19421
Topic 3:Dynamic NVMe Allocation for Bare Metal through PCIe in OpenStack
议题投票链接:
https://www.openstack.org/summit/sydney-2017/vote-for-speakers/#/19593
Topic 4:Kolla in practice for OpenStack production environment
议题投票链接:
https://www.openstack.org/summit/sydney-2017/vote-for-speakers/#/19968
Topic 5:Virtual Machine Intelligent Dynamic Resource Quota Strategy
议题投票链接:
https://www.openstack.org/summit/sydney-2017/vote-for-speakers/#/19338
Topic 6:What have we done with Tacker in NFV orchestration
议题投票链接:
https://www.openstack.org/summit/sydney-2017/vote-for-speakers/#/19277
Topic 7:Quick healing and better topology view of VNFs using Unified Event Stream.
议题投票链接:
https://www.openstack.org/summit/sydney-2017/vote-for-speakers/#/19598
Topic 8 :SDN hub platform unifies multi-vendor heterogeneous SDN powered openstacks:
议题投票链接:
https://www.openstack.org/summit/sydney-2017/vote-for-speakers/#/19507
Topic 9: Multi-Region OpenStack Swift Clusters: Lessons Learned from the Production Environments
议题投票链接:
https://www.openstack.org/summit/sydney-2017/vote-for-speakers/ - /20073
Topic 10:A Road to OpenStack in China
投票链接:
https://www.openstack.org/summit/sydney-2017/vote-for-speakers#/19738
Topic 11:All you need to know to build your GPU machine learning cloud
投票链接:
https://www.openstack.org/summit/sydney-2017/vote-for-speakers#/19622

投票流程:
1、打开议题对应的投票链接
2、登录或注册
\
3、点击投票
\
 议题简介:
Topic 1OpenStack at SJTU: Predictive Data Mining in Clinical Medicine with Dynamical HPC
演讲人:Shuquan Huang(99cloud)、 Dr Yih Leong Sun (Intel)、Luo Xuan(Shanghai Jiaotong University)
演讲内容
Shanghai Jiao Tong University (SJTU) is building an OpenStack-based HPC Cloud for clinicians from various hospitals & institutes to improve the efficiency of diagnostic, therapeutic,  and monitoring tasks. Clinicians can take advantage of cloud-based data mining technology to deal with huge amount of research data obtained from molecular medicine, such as genetic or genomic signatures and apply predictive data analytics with learning models in clinical medicine for patients' health. Predictive data mining is a typical HPC workload which is not easy to manage in OpenStack cloud. There are many challenges, trade-offs, and gaps in this cloud journey.
In this session, we’ll share:
How to build a HPC platform upon OpenStack infrastructure
What’s the key consideration of architecture design
How to dynamically provide a fully optimized HPC cluster with OpenHPC ingredients within OpenStack
How to guarantee the data mining workload performance in cloud environment

Topic 2:Composable Infrastructure ? Try Valence!
演讲人:Shuquan Huang(99cloud)、 Hu Bian(Lenovo)、Nate Potter(intel)
演讲内容
Composable Infrastructure is a revolutionary, new architecture that optimize various software and hardware for innovation ideas. Valence was introduced to disaggregates compute, storage, and network resources based on Intel Rack Scale Design. Nowadays applications and other OpenStack services can take advantage of Valence to introduce the ability to more efficiently pool and utilize these resources. Valence complements OpenStack by dynamically composing workload-optimized hardware while at the same time allowing workloads to run on bare-metal and do it all with a single management console. Valence was started one year ago and evolves rapidly. It’s readier than ever to unlock the effectiveness. In this session, we’ll share:
What exciting features have been added since last cycle
Intel Rack Scale Design Roadmap
New Features, such as Pooled NVMe resources management, Multi-Podmanager, etc.
Integration with other OpenStack projects
Community involvement and ecosystem
Use case & Demo

Topic 3:Dynamic NVMe Allocation for Bare Metal through PCIe in OpenStack
演讲人:Shuquan Huang(99cloud)、Anusha Ramineni、Lin Yang(Intel)
演讲内容
With the evolution of storage in data center, NVMe adoption is increasing by moving the low latency storage closer to CPU with flexibility and scalability. NVMe over Fabric (NVMeoF) can rely on various protocals to access remote NVMe. However, there are differences between local NVMe and NVMeoF, such as Identifier, Discovery, Queueing and Data Transfers. Besides, performance overhead is also expected.
Valence is an OpenStack project leveraging Intel Rack Scale Design Pooled NVMe Controller to dynamically allocate NVMe through PCIe switches to provide large scale sharing of storage in Rack level, evolving storage architecture from physical aggregation to resource pool. This presentation will cover how we enable Cinder and Ironic to support dynamic NVMe volumes allocation on demand and attaching them to bare metal, and demonstrate this functionality on the RSD hardware. We’ll also introduce the upstreaming plan and further support on other resources pools, including FPGA accelerators.

Topic 4:Kolla in practice for OpenStack production environment
演讲人:Jeffrey Zhang(99cloud)、
演讲内容
Container is a revolution. It is changing how the deployment works and speeds up the whole IT world. Kolla is a containerized OpenStack solution. It is a result of dogfooding and provides production-ready OpenStack cluster.
We have multi production OpenStack environment deployed by Kolla. This topic includes
the benifit of using kola
best practise when using containerized OpenStack
problems we encountered, and how we solved it.

Topic 5:Virtual Machine Intelligent Dynamic Resource Quota Strategy
演讲人:Jeffrey Zhang(99cloud)、Ice Yao(Tencent)
演讲内容
OpenStack limits the virtual machine resources by setting the Nova Flavor metadata to achieve, Nova in the creation of virtual machines according to Flavor set the metadata generated by the corresponding resource limit field libvirt xml file, the final role in the kvm virtual machine. If the administrator wants to limit the virtual machine resources during the run of the virtual machine, OpenStack can only make Flavor changes with Nova resize, but resize the virtual machine during resize (actually cold migration).
Here we propose a cloud platform based on the monitoring system to intelligently limit the realization of virtual machine resources, according to the cloud platform to monitor the performance data collected at the peak of resource use, once the performance data exceeds the set threshold to limit the virtual machine resources Quota; if below the threshold, the resource quota limit is automatically lifted; there is no need to consider intervention in this process.

Topic 6:What have we done with Tacker in NFV orchestration
演讲人:Yong Sheng Gong(99cloud)、 Yan Xing'an(China Mobile)、wu jiangtao(China Mobile)
演讲内容
Abstract:
China mobile have some requirements about NFV in public cloud or telecommunication network scenarios, such as decoupling SDN and NFV, orchestrating and managing many vendors' diverse VNFs, and standardizing the interfaces between OSS, NFVO, and VNFM. We tried to use tacker to resolve this issues. We have realized some VNFs in tacker including vFW, vLB, vRouter, and also, we commited some specs and patchs to make tacker moving forward to production.
What's the problem or use case you’re addressing in this session?
How to use tacker to orchestrate and manage diverse VNFs.
What have we done in tacker upstream.
The plan of tacker development from china mobile.
One use case from china mobile and 99cloud to enable C-RAN via Open-O and tacker.

Topic 7:Quick healing and better topology view of VNFs using Unified Event Stream.
演讲人:Yong Sheng Gong(99cloud)、 Dharmendra Kushwaha、Tung Doan
演讲内容
Tacker provides a ETSI based generic VNF Manager & NFV Orchestrator to deploy VNFs.
Now important point to take care is deployed VNF health and its life cycle.
In a typical Service Assurance Environment, the VNF Provider defines and provides the format and protocol for sending telemetry data to SPs responsible for managing VNF health and lifecycle. The standards for telemetry data come in many varieties.
There can be multiple standard to do that and one of the best one is VNF Event Stream(VES).
VES is a standardized event framework that can be used by OpenStack and OPNFV projects.
In this talk, we would like to discuss about:
How we can leverage the usage of unified event streams with Tacker.
How quickly to heal and scale up/down the VNFs using Doctor.
Use of telemetry policies at onboarding level to automate the process of healing and scaling using doctor and VES
Updates on the goals and roadmap for the VNF Event Stream (VES) project in OPNFV and ONAP.

Topic 8 :SDN hub platform unifies multi-vendor heterogeneous SDN powered openstacks
演讲人:Yong Sheng Gong(99cloud)、Chen Xingbin(China Telecom)、Zhilan Huang(China Telecom)
演讲内容
China Telecom is gradually promoting the deployment of OpenStack in cloud data centers. While network is the key link of the entire cloud solution, the current use of commercial overlay sdn products has greatly limited the open compatibility due to vendor lock-in. With the continuous evolution of Software Defined Network (SDN), the enormous challenge of heterogeneous mixing has been placed in front of China Telecom’s cloud operation.
The SDN collaboration platform (SDN HUB) developed by China Telecom together with 99Cloud, Huawei, ZTE and other companies is dedicated for the above problem. It enhances OpenStack Neutron with the ability of operating multiple sdn networks in a single panel, provides interconnectivity & interoperability abilities among multiple vendor sdn solutions. In general, SDN Hub is a network orchestrator for openstack & other cloud service with cross-domain, cross-vendor & cross-sdn controller ability.

Topic 9: Multi-Region OpenStack Swift Clusters: Lessons Learned from the Production Environments
演讲人: Mingyu Li(OStorage)
演讲内容
The presentation will take a look at two production deployments of multi-region object storage clusters based on OpenStack Swift and discuss the lessons learned. Two cases will be discussed in this presentation:
 
1. A commercial bank using Swift to store images, videos, documents and other unstructured data to meet the Active-Active requirements;
 
2. One of China Unicom's online video storage clusters with nodes located in Beijing and Hong Kong which are about 2,000 km apart.

Topic 10:A Road to OpenStack in China
演讲人: Gangyi Luo(China Mobile)、Fred LI(Huawei)、Liyun Yang(CESI)
演讲内容:
OpenStack and Open Source Software are gradually accepted by large corporations and organizations in China. Chinese Industry leaders such as China Mobile, Industrial and Commercial Bank of China (ICBC), Huawei and States Grid have built large OpenStack cloud and migrated their business into OpenStack. In Dec 2016, a guide of how to build a Cloud based on OpenStack is drafted by China Electronics Standardization Institute (CESI), China Mobile, Huawei and other leading OpenStack Companies.
 
In this topic, we will discuss about nurturing cultures that make traditional companies accepting Open Source and OpenStack.
 
Topic 11:All you need to know to build your GPU machine learning cloud
演讲人: Lu Ye
演讲内容
GPU is becoming the new common, but GPU resources are still hard to access who wants to have a taste, also management is a new problem coming up. So how to build your GPU ML cloud?
Resource management & App templating
After your organization have purchased GPU devices. Environment isolation is always a problem. Also at the beginning the cloud is more used as a playground, so another consideration is usage rate of resources. How we use Kubernetes to solve those problems. How to use a wizard to generate ML apps.
Make the “customized changes” in immutable container be be played back.
The features of container is immutable, which is a double-edged sword. Changes made can be lost after recreation. How the env is saved and reuse?
Managing persistence storage in Kubernetes
Turn Ceph RBD served as hosted S3, So the data scientist can access their data both as a volume and through s3-standard api. Support the running ML app making online resize.
App model & permission control.

© 2012-2017 九州云信息科技有限公司 99Cloud 版权所有 咨询热线:400 006 0472 售后服务热线:400 670 7810 培训咨询热线:400 826 0070   ICP证:浙ICP备12032350号-1

网站建设:信达互联

北京网站建设公司