1、Unlocking LLM Performance with :Optimizing Training and Inference PipelinesYang Xiang,Yunshan NetworksOutline1.1.BackgroundBackground:Challenges in Training and Inference Efficiency2.2.Status QuoStat。
2、How Does KubeEdge Build the Tunnel Which Is Secure,Trusted,and Adaptable to Edge NetworksWei Hu|DaoCloudClement Richard|n-hopAbort UsWei HuSenior Software Engineer at DaoCloudKubeEdge MaintainerCleme。
3、Unlocking LLM Performance with :Optimizing Training and Inference Pipelines向陽,云杉網絡Outline1.1.背景背景:訓練和推理的效率挑戰2.2.現狀現狀:傳統解決方案和工具的問題3.3.方法方法:eBPF 構建零侵擾可觀測性4.4.實踐實踐:PyTorch 全棧剖析和追蹤 LLM 訓練開銷大、效率低GPT-4GPT-。
4、Unveiling the Future:Nurturing Openness in AI Development揭示未來:培育人工智能開放性發展 Anni LaiChair,Generative AI Commons,LF AI&DataHead of Open Source Operations,FutureweiMer Joyce FounderDo Big Good LLCCo-。
5、xRegistry Looking Beyond CloudEventsLeo Li-Red Hat Hello!Leo LiSoftware Engineering Intern Red HatKnative Eventing MaintainerLeo6LeoAgendaCloudEvents(CE):Intro&Community updates1.Understand CE:an。
6、What if Your System Experiences an Outage?Lets Build a Resilient Systems with Chaos EngineeringCNCF AmbassadorNamkyu ParkMaintainer LitmusChaos(CNCF incubating)namkyu1999namkyu1999Agenda:What is this。
7、Zhonghu Xu,HuaweiJianpeng He,TetrateWhat Is the Future of Service Mesh?Sidecar or Sidecarless CNCF TAG Network Tech LeadIstio Steering Committee MemberIstio MaintainerKubernetes MemberZhonghu XuHuawe。
8、Unlocking the Power of Kubernetes(for AI)AI-Driven Innovations for Next-Gen InfrastructureBrandon Kang Who am I?Brandon Kang(姜相鎭)Akamai Technologies,APJ CTGPrincipal Technical Solutions ArchitectBack。
9、Uniting Sustainability and Edge Computing:Kepler&Open Horizon on RISC-V and Heterogeneous SystemsSpeakers:Peng Hui Jiang(IBM)&David Yao(Persistent Systems)Date:Thursday,August 22,2024Time:16:。
10、Qizhi Wang01020304050607Pay-as-you-goNo need to pay for idle resourcesFreeOSM(OpenStreetMap)Twitter in 2009250M RU1200QPS2400QPS205.21QPS$121/m$267/mThe above calculations are based on read load scen。
11、Time Series Database on Kubernetes:Efficient management of massive Internet of Vehicles dataAbout MeIm a Time-series database expert in the HUAWEI CLOUD Database Innovation Lab and the Co-founder of 。
12、Jeremy Olmsted-Thompson(jeremyot)Hongcai Ren(RainbowMango)Jian Qiu(qiujian06)SIG Multicluster Intro&Deep DiveWell cover:What this SIG is aboutOur next problem spaces Current activitySIG-MC websit。
13、Working with Raw Disk Drives in Kubernetes YDBs ExperienceIvan BlinkovIvan BlinkovVP,Product and Open-Source在Kubernetes中使用原始磁盤驅動器YDB的經驗2Ivan Blinkov Over a decade of experience in the database manage。
14、為為PyTorchPyTorch原生模型提供自動并行的訓練框架原生模型提供自動并行的訓練框架veScale TeamByteDance2024-8-22PhD畢業于多倫多大學(導師:Gennady Pekhimenko)2022年3月加入字節跳動目前主要專注于大語言模型訓練架構相關工作關于我朱虹宇議程為什么需要VeScaleVeScale設計與實現初步測試結果未來展望單個model往往需要數周時。
15、Securing the Supply ChainA Practical Guide to SLSA Compliance from Build to RuntimeAugust 21 2024 Enguerrand Allamel,LedgerAbout MeEnguerrand Allamel-Academic Experience:Studied for one year at Tsing。
16、xRegistry Looking Beyond CloudEventsLeo Li-Red Hat Hello!Leo Li軟件工程實習生 紅帽Knative Eventing 維護者日程CloudEvents(CE):介紹&社區最新動態分享1.理解什么是 CE:物流包裹運輸的小故事2.社區最新動態分享3.CloudEvents社區下一步發展的側重方向xRegistry:事件驅動架構中。
17、深入了解Cilium背后的熱潮:在阿里巴巴的實踐Liyi Huang,Isovalent&Bokang Li,Alibaba Cloud 議程Cilium 基本介紹淺談ACK上的network policy阿里云上的CNI阿里云上的規?;瘜嵺`在阿里云上完整的cilium有什么額外功能1.16 新版本發布的亮眼特性提問時間-Networking-Security-Observability。
18、Unlocking Heterogeneous AI Infrastructure K8s ClusterLeveraging the Power of HAMi MengXuan LiGithub archlitchi4ParadigmAbout usXiao Zhang software engineerGithub wawa0210 DaoCloudChanllange 1:Require。
19、Revolutionizing Scientific Simulations with Argo WorkflowsShuangkun Tian&Yashi Su Alibaba CloudAbout UsShuangkun,TianAlibaba Cloud Software EngineerArgo Community MaintainerYashi,SuAlibaba Cloud 。
20、A PyTorchA PyTorch-Native AutoNative Auto-Parallel Framework Parallel Framework for for Ease of UseEase of UseveScale TeamByteDance2024-8-22About MeReceived my PhD degree from University of Toronto(a。
21、The Journey of Next-Gen FinTech IDP at China Merchants BankJiahang Xu,China Merchants BankAgenda CMB cloud migration Changes after cloud migration Challenges from Complexity Platform Engineering As S。
22、Testing and Release Patterns for CrossplaneStefano Borrelli and Yury Tsarev,August 22th 2024Stefano has worked as a Systems Engineer for the finance and technology companies,as well as starting a clo。
23、OpenYurt&Dragonfly:Enhancing Efficient Distribution of LLMs in Cloud-Edge CollaborativeLinbo He,Alibaba cloudJim Ma,Ant GroupAgendaEdge Computing and AIOpenYurtDragonflyPracticeEdge Computing and。
24、Understanding the Buzz Around Cilium:Introduction and in Production at AlibabaLiyi Huang,Isovalent&Bokang Li,Alibaba Cloud AgendaCilium general introductionShallow dive from network policy with A。
25、Strengthening Container Security:A Collaborative JourneyBeltran Rueda,Sr Eng Manager,VMWare Tanzu/BroadcomYi Zha,Senior Product Manager,MicrosoftAbout usYi ZhaSr Product Manager at MicrosoftMaintaine。
26、經濟增安全從構建到運時的 SLSA 合規性實指南2024年821 安格羅(Enguerrand Allamel),Ledger關于我安格羅(Enguerrand Allamel)-學術經歷:曾在清華學學習年-本職位:Ledger 的級云安全程師-公司概述:Ledger 尖端于安全錢包硬件和尖端安全產品-當前關注:通過加強戶保護來增強供應鏈安全議程1.為什么供應鏈安全重要?2.什么是 SLSA(軟。
27、The Challenges of Kubernetes Data Protection-Real Examples and Solutions with VeleroWenkai Yin()Bo Zou()AgendaVelero IntroductionChallenges and solutionsSummaryWhat Is VeleroOpen source backup and re。
28、KubeSkoop-Deal with the Complexity of Network Issues and Monitoring with eBPFYutong Li,Alibaba Cloud Bingshen Wang,Alibaba CloudAgendaThe complexity of Kubernetes NetworkingIntroduction to KubeSkoopK。
29、Run LLM agents on self-hosted devicesMichael Yuanx:juntao github:juntaoWasmEdge Runtime:https:/ Demo:The easiest way to chat with an open-source LLM on your own deviceDownload and install the softwar。
30、唐晨杰使用內存引擎強化 TiKVTiKV 簡介TiKV 讀場景痛點內存引擎概述內存引擎實現細節01020304目錄內存引擎性能測試結果05TiKV 簡介TiKV 簡介 TiDB 的存儲層 分布式鍵值數據庫 提供滿足 ACID 約束的分布式事務(percolator)接口 Raft 協議保證多副本一致性及高可用GitHub:https:/ TiKV 下的多版本控制(MVCC):key1key1-v。
31、kebe7junnekomeowww5.47.16.05.44.26.44.38.86.28.55.95.14.56.74.87.35.86.97.04.84.16.34.07.96.17.26.55.23.96.04.68.15.57.46.15.34.36.54.48.56.38.66.05.24.66.84.97.45.97.06.94.94.26.44.18.06.07.16.85.04。
32、JDCloud cross-cluster large-scale application management practiceJDCloud2024XiaoFei WangJDCloudCloudNativeEJDCloud2024Table of ContentsDevelopment Road of JD Cloud Containers0103Federation Auto-scali。
33、Safeguarding Cloud Native Supply Chain Notary Project Intro&Whats NextYi Zha,Senior Product Manager,MicrosoftMostafa Radwan,Principal Consultant,CloudRoadsAbout usYi ZhaSr Product Manager at Micr。
34、Unlocking Scalability and Simplifying Multi-Cloud Management with Karmada and PipeCDHongcai Ren,Huawei,Karmada CommunityKhanh Tran,CyberAgent,PipeCD CommunityAgendaWhy multi-clusterIntroduction of Ka。
35、Observability SuperchargerBuild the Traffic Topology Map for Millions of Containers with Zero CodeLim Teck ChuanChow Sheng WeiEngineering Infra,ShopeeWhat is a topology map?ServiceKafkaDatabaseWhy is。
36、Rollout Patterns:Smoothly Migrate and Roll Out Your MicroservicesTim Xiao,DaoCloudAbout meTim muma378DaoCloud,Software Engineer,DevOps Platform LeaderArgo Project ContributorHiking、BadmintonandCoffee。
37、Redefining Service MeshLeveraging eBPF to Optimize Istio Ambient Architecture andPerformanceTechnical Expert,Alibaba CloudIstio&Envoy member,has rich experiences in cloud native fields such as Ku。
38、Unified Management,Continuity,Compliance in Multi-Clouds with Service MeshKebe Liu/DaoCloud/Istio Steering Committee MemberAgendaMulti-cloud scenarios and challengesIstio+KarmadaSolutionsCase studies。
39、From Novice to LitmusChaos MaintainerLearn to Make Valuable Contributions to Open Source Projects through LFX MentorshipCNCF AmbassadorNamkyu ParkMaintainer LitmusChaos(CNCF incubating)namkyu1999namk。
40、通過 Volcano 增強的智能基礎設施優化LLM作流程李鑫,Qihoo360 常旭征,Huawei Cloud Technologies Co.,LTD錄1.背景2.現狀3.存在問題4.如何解決背景LLM關鍵詞趨勢 從 2023 年開始,LLM 受到的關注越來越多 使 Kubernetes 的 LLM 基礎設施越來越多 kubernetes 對于 LLM 的持越來越好OpenAI博客文章201。
41、王琦智01020304050607按需付費無需為閑置資源付費免費OSM(OpenStreetMap)2009 年的Twitter2.5億 RU1200QPS2400QPS205.21QPS$121/m$267/m以上計算結果使用讀負載場景進行估算,實際支出請以 TiDB Cloud Billing 計算結果為準既可以向量搜索又不限數據量級這不巧了嘛這不是,歡迎體驗 TiDB Serverless。
42、1OpenTelemetry Amplified:Full Observability with EBPF-Enabled Distributed TracingKai Liu1,Wanqi Yang21Alibaba Cloud,2Sun Yat-sen University2Agenda31.Observability in Kubernetes2.End-to-End Request Tr。
43、Several Lessons I Learned From Project Governance to Corporate OSS GovernanceXu Wang,AntGroup&Kata Containers etc.My Brief ExperiecesOSS Startup:2015.Co-Found hyper.sh and projects:hyper-containe。
44、0NanoVisor:Revolutionizing FaaS Cold Start Performance with Secure,Lightweight Container RuntimeTianyu Zhou 1ContentsNanoVisor Intro&Role in FaaSCold Start OptimizationEvaluationSummary23NanoViso。
45、Scaling Kubernetes:Best Practices for Managing Large-Scale Batch Jobs with Spark and Argo WorkflowYu Zhuang Alibaba CloudJiaxu Liu Alibaba CloudAbout UsAlibaba Cloud Container Kubernetes Service(ACK)。
46、多集群助力小紅書打造面向混合云的高可用彈性架構熊峰Agenda為什么引入多集群&選型思考應用如何分發與遷移Kubernetes API 兼容多集群場景下彈性能力建設控制面自動化運維規劃與展望解決什么問題規模限制管理難度資源效率混合云-多云戰略單集群約5000節點的規模上限集群數量眾多,持續快速增長無法滿足互聯網超大規模應用對業務發布與運維管理帶來挑戰業務與集群深度綁定,資源分配效率低下資。
47、Redefining Service MeshLeveraging eBPF to Optimize Istio Ambient Architecture andPerformanceTechnical Expert,Alibaba CloudIstio&Envoy member,has rich experiences in cloud native fields such as Ku。
48、Kubernetes Community Panel:A Decade of Evolution and Future Trends演講嘉賓Klaus MaSIG Scheduling ex-ChairCNCF TAG-Runtime TLCNCF WG-Batch co-chairk82cnKevin WangCNCF TOC SIG Scheduling&SIG multi-clus。