CCF YOCSEF, China Computer Federation Young Computer Scientists & Engineers Forum, is a series of academic activities created by CCF since 1998. It is a vigorous event planned and organized by academics, entrepreneurs and other young elite of passion, ideas and social responsibility.
The CCF YOCSEF activities are mainly in the form of forums and academic reports, which develop the discussion on IT industry and social hot issues, and spread the latest academic thinking and technology trends to the society. In addition, academic evaluation, awards and poverty alleviation.
As vice chair, Intel is honored to sponsor Cloud Computing and Its Core Technology Seminar, 2012 CCF YOCSEF Shanghai on June 8th, 2012.
This forum is focused on Cloud Computing and its core technology. According to registration statistics, up to 89 participants attended this forum, among which, 36 came from universities and 53 came from industry.
Keynote speech 1:
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.
Presenter: Jinquan (Jason) Dai
(Principal Engineer of Intel Software and Services Group, managing the engineering team for building and optimizing Big Data and Cloud related platforms.)
Keynote speech 2:
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.
Presenter: BinYu Zang
(Dean of School of Software; director of Parallel Programming Institute , Fudan University.)
Keynote speech 3:
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.
Presenter: Jiangang Duan
(Manager of scalability lab in Intel Asia-Pacific Research & Development Ltd, working on enterprise solution tuning and optimization for 10+ years.)
Intel Zizhu Site visit and group photo taking
Keynote speech 4:
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.
Presenter: Haibing Guan
(Deputy 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.)
Keynote speech 5:
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.
Presenter: Yaozu (Eddie) Dong
(Software Architect of Intel Open Source Technology Center, working on Linux virtualization for 10+ years.)
Keynote speech Q & A
Go to dinner venue
Dinner & free discussion