中国计算机学会青年计算机科技论坛

CCF Young Computer Scientists & Engineers Forum

CCF YOCSEF上海

于2012年 6月8日(星期五)13:30-20:00

在上海英特尔亚太研发有限公司举行

专题论坛，敬请光临

论坛主题

云计算及其核心支撑技术论坛

程 序

13:30     签到

13:45-14:00   英特尔致开幕辞

14:00-14:30   主题演讲一：Optimizing Hadoop Deployments

14:30-15:00   主题演讲二：CloudVisor: Retrofitting Protection of Virtual Machines in Multi-tenant Cloud with Nested Virtualization

15:00-15:30   主题演讲三：Build public and private clouds with open source software

15:30-16:15   英特尔公司参观及合影

16:15-16:40   主题演讲四：Memvisor: Flexible High Avaliability Memory via Binary Translation

16:40-17:15   主题演讲五：High Performance Network Virtualization with SR-IOV

17:15-17:30   主题演讲问答

执行主席：董耀祖 英特尔亚太研发有限公司

CCF YOCSEF上海学术委员会副主席

执行主席：颜   历 英特尔亚太研发有限公司

执行主席：董笑菊 上海交通大学计算机科学与工程系副教授

  CCF YOCSEF上海学术委员会委员

执行主席：沈丕豪 上海翰纬信息科技有限公司

  CCF YOCSEF上海委员

主办：CCF YOCSEF上海、英特尔亚太研发有限公司

云计算及其核心支撑技术 云计算通过将各种计算机领域中的软硬件资源，如Infrastructure, platform, software等，以按需使用的方式向客户提供服务，从而提高资源的使用效率。云计算的核心支撑技术包括虚拟化，安全，大数据以及数据中心管理等。本研讨会同时邀请了来自英特尔和上海交大和复旦的相关专家，分享他们在云计算核心技术领域的研究和开发成果。其部分研究成果已经发表于SOSP, USENIX, HPCA和TACO等国际顶级会议和杂志，并被工业界广泛采用。同时，我们也邀请到国家云计算专家，作关于产业发展规划方面的讲解。
特邀讲者介绍及内容摘要 Opening and OverviewPresenter: Honesty YoungChief Research Officer of Intel Asia-Pacific Research & Development Ltd. Optimizing Hadoop DeploymentsPresenter: Jinquan (Jason) DaiPrincipal Engineer of Intel Software and Services Group, managing the engineering team for building and optimizing Big Data and Cloud related platforms. Abstract:Having moved beyond its origins in search indexing, Hadoop is evolving toward the de facto standard for Big Data storage and analytics. It allows the users to easily develop a program (using a simple two-phase primitive) that can be automatically scaled to thousands of nodes, without worrying about the details of parallelism. Instead, the Hadoop framework is responsible for scheduling and distributing the program to run on a scale-out, shared-nothing cluster, abstracting away the messy details of data partitioning, resource allocation, node communications and fault tolerance. Unfortunately, this abstraction makes it very difficult, if not impossible, for the users to understand the runtime behavior of Hadoop applications; consequently, it becomes a big challenge for the users to efficiently provision and fine-tune the Hadoop systems. This talk will review the challenges in the efficient provisioning and tuning for Hadoop, and present the possible methodologies and tools (e.g., HiBench, HiTune, etc.) to help address these challenges. CloudVisor: Retrofitting Protection of Virtual Machines in Multi-tenant Cloud with Nested VirtualizationPresenter: BinYu ZangDean of School of Software; director of Parallel Programming Institute , Fudan University. Abstract:Multi-tenant cloud, which usually leases resources in the form of virtual machines, has been commercially available for years. Unfortunately, with the adoption of commodity virtualized infrastructures, software stacks in typical multi-tenant clouds are non-trivially large and complex, and thus are prone to compromise or abuse from adversaries including the cloud operators, which may lead to leakage of security-sensitive data.ClousVisor proposes a transparent, backward-compatible approach that protects the privacy and integrity of customers’ virtual machines on commodity virtualized infrastructures, even facing a total compromise of the virtual machine monitor (VMM) and the management VM. The key of the approach is the separation of the resource management from security protection in the virtualization layer. A tiny security monitor is introduced underneath the commodity VMM using nested virtualization and provides protection to the hosted VMs. As a result, our approach allows virtualization software (e.g., VMM, management VM and tools) to handle complex tasks of managing leased VMs for the cloud, without breaking security of users’ data inside the VMs. Build public and private clouds with open source softwarePresenter: Jiangang DuanManager of scalability lab in Intel Asia-Pacific Research & Development Ltd, working on enterprise solution tuning and optimization for 10+ years. Abstract:With the rising of cloud computing, more and more companies consider to build their own data center to offer cloud services. OpenStack is one of the most popular open source projects to build an open, extensible, framework for managing the various resources in a cloud. In this talk, we will give an introduction about the openstack project, including history, core project, and technical challenge and existing solutions. We will also give a brief introduction about Intel effort in this new area on how to build a more efficient cloud. Memvisor: Flexible High Availability Memory via Binary TranslationPresenter: Haibing GuanDeputy Dean of School of Electronic, Information and Electronic Engineering of Shanghai Jiaotong University, and director of the Shanghai Key Laboratory of Scalable Computing and Systems. Abstract:Nowadays, most high availability strategies for memory focus on mirrored hardware solution or application-level backup technologies to tolerate memory failures, which bring an expensive cost for cloud server operator and rigorous hardware/software environmental requirements. We propose Memvisor, a modified hypervisor to achieve real-time flexible memory backup and dynamical recovery in clusters. Memvisor creates the mirrored space by the additional instructions (mirrored instructions) through binary translation for every memory write operation, which provides the runtime backup for application and kernel's memory. When memory error occurs, the corrupted data will be recovered from the mirrored memory by Memvisor. We have implemented Memvisor in Xen-3.4.2, and used Sysbench as a practical benchmark to evaluate the performance overhead incurred by Memvisor. High Performance Network Virtualization with SR-IOVPresenter: Yaozu (Eddie) DongSoftware Architect of Intel Open Source Technology Center, working on Linux virtualization for 10+ years. Abstract:Virtualization poses new challenges to I/O performance. The single-root I/O virtualization (SR-IOV) standard allows an I/O device to be shared by multiple Virtual Machines (VMs), without losing runtime performance. This talk proposes a generic virtualization architecture for SR-IOV devices, which can be implemented on multiple Virtual Machine Monitors (VMMs). With the support of this architecture, the SR-IOV device driver is highly portable and agnostic of underlying VMM. Because the Virtual Function (VF) driver with SR-IOV architecture sticks to hardware and poses a challenge to VM migration, this talk also proposes a dynamic network interface switching (DNIS) scheme to address the migration challenge. 报名 请在6月3日（周日）前，填写所附参会回执并email至 mei.yu@intel.com 。因名额有限，请您尽早完成报名。  细节事项 会议备有纸笔及饮用水。因会议时间有限，自由讨论安排在晚宴期间同时进行。实验室参观期间请不要使用相机，谢谢配合。地点与交通 地点：英特尔亚太研发有限公司大会议室BUM 1F05 上海市闵行区紫星路880号（紫竹科学园区内），近东川路公共交通：公交180，180区间，虹桥枢纽4路，江川3线均可抵达。 地铁：地铁5号线东川路站下，打车起步费即可抵达。地铁8号线航天博物馆站下，打车抵达。组委会力争安排班车在地铁口定时等候，详见会前参会须知。 英特尔公司方位示意图：