1 / 12

System Research in MSRA

System Research in MSRA. Bingsheng He. Overview. Research works within SRG Debugging tools Many-core operating system Personal research interests Architecture aware systems Cooperation opportunities. Goal. Debugging is challenging

betty_james
Télécharger la présentation

System Research in MSRA

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. System Research in MSRA Bingsheng He

  2. Overview • Research works within SRG • Debugging tools • Many-core operating system • Personal research interests • Architecture aware systems • Cooperation opportunities

  3. Goal • Debugging is challenging • Complex software: multi-threaded, distributed systems, legacy modules • Complex environment: hardware failures, network partition, disk error… • Goal: make it easy to debug complex, distributed systems • Correctness bugs • Performance bugs

  4. Our approaches • Static analysis: enforce coding discipline • HangWiz (EuroSys’08): soft-hang detection • RegionWiz (PLDI’08): memory-region safety verification • Runtime analysis • D3S (NSDI’08): distributed predicate checking • R2 (OSDI’08): deterministic replay to make bugs reproducible

  5. Multi-core: the trend • Chips with 2 or 4 cores are common • Chips with 10~100 cores will appear within 5 years. [A. Agarwal 2007] Corollary of Moore’s Law

  6. Problems in Existing OSes • Share data structures protected by locks • Scheduling queues. [Gough et al. 2007] • Directory lookups. [Veal and Foong 2007] • Codebases are too large and complex • Modifications are resource-intensive undertaking • Corey: the first research prototype of redrawing the multi-core OS architecture

  7. Corey Design • Fundamental principle • Let applications/LibOS control sharing • Lock/share carefully /rarely • Use cores creatively Mapreduce App Web Server App Mapreduce libOS Web Server libOS U K Corey Hardware Many-core C1 C2 C0 C3 Interconnect

  8. Architecture Aware Systems • Processor aware systems • Cache conscious automata [ICDE’05, TKDE’06] • GPGPU [SIGMOD’07, SIGMOD’08] • Disk aware systems • Flash-aware B+-tree [ICDE’09] • Power aware systems

  9. Cooperation Opportunities • Co-application for project funds from MSRA • Funding promising research projects interested by the University and MSRA • Targeting at top-rank publications (usually) • Internship/Visiting scholar • Exposing the students/faculties with the state-of-the-art research • Building a close relationship between MSRA and universities

  10. Who we are • Managers • Zheng Zhang • Lidong Zhou • Staff • Zhenyu GUO, Yang MAO, Wei LIN, Xuezheng LIU, Ming WU, Haoxiang LIN, Bing Su, Bingsheng He • Interns (MIT, UCSD…)

  11. Our extended family • Close collaborators in China • HUST (P2P VoD, Storage) • Fudan (manycore) • Tsinghua (HPC) • SJTU (Model checking/verification) • BJU (P2P) • … Lintao ZHANG (MSR-SVC) Frans Kaashoek (MIT) Geoff Voelker (USCD) Junfeng YANG (MSR-SVC/Columbia Univ)

  12. Q&A • Thanks.

More Related