{"id":1170563,"date":"2026-05-05T09:00:00","date_gmt":"2026-05-05T16:00:00","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/?p=1170563"},"modified":"2026-05-07T19:25:04","modified_gmt":"2026-05-08T02:25:04","slug":"microsoft-at-nsdi-2026-advances-in-large-scale-networked-systems","status":"publish","type":"post","link":"https:\/\/www.microsoft.com\/en-us\/research\/blog\/microsoft-at-nsdi-2026-advances-in-large-scale-networked-systems\/","title":{"rendered":"Microsoft at NSDI 2026: Advances in large-scale networked systems"},"content":{"rendered":"\n<figure class=\"wp-block-image size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"1400\" height=\"788\" src=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1.jpg\" alt=\"NSDI \u201926 logo in white, centered on a smooth gradient background transitioning from blue to purple and pink.\" class=\"wp-image-1170639\" srcset=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1.jpg 1400w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1-300x169.jpg 300w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1-1024x576.jpg 1024w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1-768x432.jpg 768w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1-1066x600.jpg 1066w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1-655x368.jpg 655w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1-240x135.jpg 240w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1-640x360.jpg 640w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1-960x540.jpg 960w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1-1280x720.jpg 1280w\" sizes=\"auto, (max-width: 1400px) 100vw, 1400px\" \/><\/figure>\n\n\n\n<p>Large-scale networked systems underpin cloud computing, AI, and distributed applications and services. The <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.usenix.org\/conference\/nsdi26\" target=\"_blank\" rel=\"noopener noreferrer\">USENIX Symposium on Networked Systems Design and Implementation 2026<span class=\"sr-only\"> (opens in new tab)<\/span><\/a> (NSDI \u201926) is a leading forum where researchers and practitioners share new research, insights, and advances in the design and operation of these systems.<\/p>\n\n\n\n<p>Microsoft is proud to support NSDI \u201926 as a returning sponsor, reflecting our ongoing commitment to advancing systems and networking research and engaging with the broader community. Microsoft researchers and engineering leaders are also serving on the program committee and in other organizational roles.<\/p>\n\n\n\n<p>This year, 11 papers by Microsoft authors and collaborators were accepted to the conference, spanning datacenter and wide-area networks, AI systems, and cloud infrastructure. Together, they highlight advances in building and operating large-scale networked systems.<\/p>\n\n\n\n\t<div class=\"border-bottom border-top border-gray-300 mt-5 mb-5 msr-promo text-center text-md-left alignwide\" data-bi-aN=\"promo\" data-bi-id=\"670821\">\n\t\t\n\n\t\t<p class=\"msr-promo__label text-gray-800 text-center text-uppercase\">\n\t\t<span class=\"px-4 bg-white display-inline-block font-weight-semibold small\">Spotlight: Microsoft research newsletter<\/span>\n\t<\/p>\n\t\n\t<div class=\"row pt-3 pb-4 align-items-center\">\n\t\t\t\t\t\t<div class=\"msr-promo__media col-12 col-md-5\">\n\t\t\t\t<a class=\"bg-gray-300 display-block\" href=\"https:\/\/info.microsoft.com\/ww-landing-microsoft-research-newsletter.html\" aria-label=\"Microsoft Research Newsletter\" data-bi-cN=\"Microsoft Research Newsletter\" target=\"_blank\">\n\t\t\t\t\t<img decoding=\"async\" class=\"w-100 display-block\" src=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2019\/09\/Newsletter_Banner_08_2019_v1_1920x1080.png\" alt=\"\" \/>\n\t\t\t\t<\/a>\n\t\t\t<\/div>\n\t\t\t\n\t\t\t<div class=\"msr-promo__content p-3 px-5 col-12 col-md\">\n\n\t\t\t\t\t\t\t\t\t<h2 class=\"h4\">Microsoft Research Newsletter<\/h2>\n\t\t\t\t\n\t\t\t\t\t\t\t\t<p id=\"microsoft-research-newsletter\" class=\"large\">Stay connected to the research community at Microsoft.<\/p>\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"wp-block-buttons justify-content-center justify-content-md-start\">\n\t\t\t\t\t<div class=\"wp-block-button is-style-fill-chevron\">\n\t\t\t\t\t\t<a href=\"https:\/\/info.microsoft.com\/ww-landing-microsoft-research-newsletter.html\" aria-describedby=\"microsoft-research-newsletter\" class=\"btn btn-brand glyph-append glyph-append-chevron-right\" data-bi-cN=\"Microsoft Research Newsletter\" target=\"_blank\">\n\t\t\t\t\t\t\tSubscribe today\t\t\t\t\t\t<\/a>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div><!--\/.msr-promo__content-->\n\t<\/div><!--\/.msr-promo__inner-wrap-->\n\t<\/div><!--\/.msr-promo-->\n\t\n\n\n<h2 class=\"wp-block-heading\" id=\"technical-sessions\">Technical sessions<\/h2>\n\n\n\n<p><em>Monday, May 4, 2:00\u20133:20 PM<\/em><\/p>\n\n\n\n<h3 class=\"wp-block-heading h5\" id=\"droidspeak-kv-cache-sharing-across-fine-tuned-model-variants\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/www.usenix.org\/conference\/nsdi26\/presentation\/liu-yuhan\">DroidSpeak: KV Cache Sharing Across Fine-tuned Model Variants<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h3>\n\n\n\n<p><em>Yuhan Liu, Yuyang Huang, Jiayi Yao, Zhuohan Gu, Kuntai Du, Hanchen Li, Yihua Cheng, and Junchen Jiang, University of Chicago; <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/shanlu\/\"><em>Shan Lu<\/em><\/a><em>, <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/madanm\/\"><em>Madan Musuvathi<\/em><\/a><em>, and <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/eschouks\/\"><em>Esha Choukse<\/em><\/a><em>, Microsoft<\/em><\/p>\n\n\n\n<p>DroidSpeak enables LLMs with the same architecture to share and partially reuse KV caches across models, delivering up to 4 times higher throughput and faster responses with minimal impact on output quality.<\/p>\n\n\n\n<div style=\"height:20px\" aria-hidden=\"true\" class=\"wp-block-spacer\"><\/div>\n\n\n\n<p><em>Monday, May 4, 3:50\u20135:30 PM<\/em><\/p>\n\n\n\n<h3 class=\"wp-block-heading h5\" id=\"eywa-automating-model-based-testing-using-llms\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/www.usenix.org\/conference\/nsdi26\/presentation\/mondal\">Eywa: Automating Model-Based Testing using LLMs<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h3>\n\n\n\n<p><em>Rajdeep Mondal, Rathin Singha, Todd D. Millstein, and George Varghese, UCLA; <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/rybecket\/\"><em>Ryan Beckett<\/em><\/a><em> and <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/sivakakarla\/\"><em>Siva Kesava Reddy Kakarla<\/em><\/a><em>, Microsoft Research<\/em><\/p>\n\n\n\n<p>Eywa uses LLMs to automatically build protocol models from natural language sources, enabling model-based testing. It uncovered 33 bugs, including 16 previously unknown, in widely used network protocol implementations.<\/p>\n\n\n\n<div style=\"height:20px\" aria-hidden=\"true\" class=\"wp-block-spacer\"><\/div>\n\n\n\n<p><em>Tuesday, May 5, 2:00\u20133:20 PM<\/em><\/p>\n\n\n\n<h3 class=\"wp-block-heading h5\" id=\"octopus-enhancing-cxl-memory-pods-via-sparse-topology\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/www.usenix.org\/conference\/nsdi26\/presentation\/zhong\">Octopus: Enhancing CXL Memory Pods via Sparse Topology<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h3>\n\n\n\n<p><em>Yuhong Zhong, Columbia University; <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/fkazhamiaka\/\"><em>Fiodar Kazhamiaka<\/em><\/a><em>, Pantea Zardoshti, Shuwei Teng and <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/rofons\/\"><em>Rodrigo Fonseca<\/em><\/a><em>, Microsoft Azure; Mark D. Hill, University of Wisconsin-Madison; <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/daberg\/\"><em>Daniel S. Berger<\/em><\/a><em>, Microsoft Azure and University of Washington<\/em><\/p>\n\n\n\n<p>Octopus&nbsp;introduces a switch-free design for disaggregated memory pods that reduces cost and scales to multi-rack pods. On a three-server hardware prototype, Octopus RPCs are 3.2x faster than in-rack RDMA and 2.4x faster than CXL switches.<\/p>\n\n\n\n<div style=\"height:20px\" aria-hidden=\"true\" class=\"wp-block-spacer\"><\/div>\n\n\n\n<p><em>Tuesday, May 5, 3:50\u20135:30 PM<\/em><\/p>\n\n\n\n<h3 class=\"wp-block-heading h5\" id=\"hedge-traffic-engineering-with-probabilistic-link-capacities\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/www.usenix.org\/conference\/nsdi26\/presentation\/devraj\">HEDGE: Traffic Engineering with Probabilistic Link Capacities<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h3>\n\n\n\n<p><em>Arjun Devraj, Cornell University; Bill Owens, NYSERNet; Umesh Krishnaswamy, Microsoft; Ying Zhang, Meta; Rachee Singh, Cornell University<\/em><\/p>\n\n\n\n<p>HEDGE mitigates wavelength-specific faults in optical networks by combining link-local and global network-wide resilience that maintain stable capacity and optimize traffic flow despite fluctuating link performance. It matches existing systems\u2019 throughput while reducing network disruptions.<\/p>\n\n\n\n<div style=\"height:20px\" aria-hidden=\"true\" class=\"wp-block-spacer\"><\/div>\n\n\n\n<p><em>Wednesday, May 6, 9:00\u201310:20 AM<\/em><\/p>\n\n\n\n<h3 class=\"wp-block-heading h5\" id=\"ava-towards-video-analytics-with-vision-language-models\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/www.usenix.org\/conference\/nsdi26\/presentation\/yan\">AVA: Towards Video Analytics with Vision Language Models<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h3>\n\n\n\n<p><em>Yuxuan Yan, Zhejiang University; <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/shijiang\/\"><em>Shiqi Jiang<\/em><\/a><em>, Microsoft Research; Ting Cao, Tsinghua University; <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/yifanyang\/\"><em>Yifan Yang<\/em><\/a><em>, Microsoft Research; Qianqian Yang and Yuanchao Shu, Zhejiang University; <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/yuqyang\/\"><em>Yuqing Yang<\/em><\/a><em> and Lili Qiu, Microsoft Research<\/em><\/p>\n\n\n\n<p>AVA supports open-ended video analytics by combining event knowledge graphs with agentic retrieval over vision-language models. Furthermore, to evaluate video analytics in ultra-long, open-world scenarios, the authors introduce AVA-100, a benchmark comprising eight videos each exceeding 10 hours and 120 manually annotated, diverse, and complex question\u2013answer pairs, on which AVA achieves 75.8% accuracy.<\/p>\n\n\n\n<div style=\"height:20px\" aria-hidden=\"true\" class=\"wp-block-spacer\"><\/div>\n\n\n\n<p><em>Wednesday, May 6, 9:00\u201310:20 AM<\/em><\/p>\n\n\n\n<h3 class=\"wp-block-heading h5\" id=\"smartnic-enabled-live-migration-for-storage-optimized-vms-with-pyrocumulus\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/www.usenix.org\/conference\/nsdi26\/presentation\/zhao-jiechen\">SmartNIC-Enabled Live Migration for Storage-Optimized VMs with Pyrocumulus<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h3>\n\n\n\n<p><em>Jiechen Zhao, University of Toronto and Microsoft Research Asia; <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/rashu\/\"><em>Ran Shu<\/em><\/a><em>, Lei Qu, Ziyue Yang, and <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/mrui\/\"><em>Rui Ma<\/em><\/a><em>, Microsoft Research Asia; Derek Chiou, Microsoft and UT Austin; Natalie Enright Jerger, University of Toronto; <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/pengc\/\"><em>Peng Cheng<\/em><\/a><em> and <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/yqx\/\"><em>Yongqiang Xiong<\/em><\/a><em>, Microsoft Research Asia<\/em><\/p>\n\n\n\n<p>Pyrocumulus enables fast, low-overhead live migration for storage-optimized VMs through hardware customizability and efficient network accessibility of the FPGA SmartNIC with LM protocol, architecture, and algorithm designs.&nbsp;<\/p>\n\n\n\n<div style=\"height:20px\" aria-hidden=\"true\" class=\"wp-block-spacer\"><\/div>\n\n\n\n<p><em>Wednesday, May 6, 10:50 AM\u201312:30 PM<\/em><\/p>\n\n\n\n<h3 class=\"wp-block-heading h5\" id=\"forestcoll-throughput-optimal-collective-communications-on-heterogeneous-network-fabrics\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/www.usenix.org\/conference\/nsdi26\/presentation\/zhao-liangyu\">ForestColl: Throughput-Optimal Collective Communications on Heterogeneous Network Fabrics<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h3>\n\n\n\n<p><em>Liangyu Zhao, University of Washington; Saeed Maleki, Independent Researcher; Yuanhong Wang, Tsinghua University; Zezhou Wang, University of Washington; Ziyue Yang, Microsoft Research; Hossein Pourreza, Microsoft; Arvind Krishnamurthy, University of Washington<\/em><\/p>\n\n\n\n<p>ForestColl constructs broadcast\/aggregation spanning trees as the communication schedule, achieving theoretical optimality. Its schedule generation runs in polynomial time and is highly scalable. It supports any network fabric, including both switching fabrics and direct accelerator connections.<\/p>\n\n\n\n<div style=\"height:20px\" aria-hidden=\"true\" class=\"wp-block-spacer\"><\/div>\n\n\n\n<p><em>Wednesday, May 6, 10:50 AM\u201312:30 PM<\/em><\/p>\n\n\n\n<h3 class=\"wp-block-heading h5\" id=\"heuristic-analysis-from-source-code-via-symbolic-guided-optimization\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/www.usenix.org\/conference\/nsdi26\/presentation\/karimi\">Heuristic Analysis from Source Code via Symbolic-Guided Optimization<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h3>\n\n\n\n<p><em>Pantea Karimi, MIT; <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/sivakakarla\/\"><em>Siva Kesava Reddy Kakarla<\/em><\/a><em> and <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/rybecket\/\"><em>Ryan Beckett<\/em><\/a><em>, Microsoft Research; Santiago Segarra, Rice University; <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/namyarpooria\/\"><em>Pooria Namyar<\/em><\/a><em>, Microsoft Research; Mohammad Alizadeh, MIT; <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/bearzani\/\"><em>Behnaz Arzani<\/em><\/a><em>, Microsoft Research<\/em><\/p>\n\n\n\n<p>MetaEase analyzes heuristics directly from source code to uncover worst-case performance scenarios, eliminating the need for complex formal modeling. It matches or outperforms state-of-the-art analyzers across domains and reveals previously unknown performance gaps in real-world systems.<\/p>\n\n\n\n<div style=\"height:20px\" aria-hidden=\"true\" class=\"wp-block-spacer\"><\/div>\n\n\n\n<p><em>Wednesday, May 6, 2:00\u20133:20 PM<\/em><\/p>\n\n\n\n<h3 class=\"wp-block-heading h5\" id=\"harvesting-spare-cpu-resources-in-container-systems\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/www.usenix.org\/conference\/nsdi26\/presentation\/hall\">Harvesting Spare CPU Resources in Container Systems<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h3>\n\n\n\n<p><em>Adam Hall and Anirudh Sarma, Georgia Institute of Technology; <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/eschouks\/\"><em>Esha Choukse<\/em><\/a><em>, Microsoft Azure Research; Umakishore Ramachandran, Georgia Institute of Technology; <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/samehe\/\"><em>Sameh Elnikety<\/em><\/a><em>, Microsoft Research<\/em><\/p>\n\n\n\n<p>HarvestContainers protects latency-sensitive containers from interference while using their spare CPU cores to run latency-tolerant workloads. It dynamically determines how many cores can be safely harvested and requires no changes to applications or the operating system. It enables up to 75% utilization of spare CPU while keeping tail latency within 4% of standalone performance.<\/p>\n\n\n\n<div style=\"height:20px\" aria-hidden=\"true\" class=\"wp-block-spacer\"><\/div>\n\n\n\n<p><em>Wednesday, May 6, 3:50\u20135:30 PM<\/em><\/p>\n\n\n\n<h3 class=\"wp-block-heading h5\" id=\"offloading-cloud-network-services-at-production-scale-with-sonic-dash-smartswitch\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/www.usenix.org\/conference\/nsdi26\/presentation\/wu-shaofeng\">Offloading Cloud Network Services at Production Scale with SONiC DASH SmartSwitch<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h3>\n\n\n\n<p><em>Community Award Winner<\/em><\/p>\n\n\n\n<p><em>Shaofeng Wu, The Chinese University of Hong Kong and Microsoft Research Asia<\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/zhniu\/\"><em>; Zhixiong Niu<\/em><\/a><em>, Microsoft Research Asia; Riff Jiang, Lawrence Lee, Junhua Zhai, <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/zegan\/\"><em>Ze Gan<\/em><\/a><em>, Vasundhara Volam, Prabhat Aravind, Prince Sunny, Prince George, Qi Luo, Evan Langlais, Soumya Tiwari, Venkat Satish Katta, Weixi Chen, Rishiraj Hazarika, Sachin Jain, Deven Jagasia, Michal Zygmunt, Avijit Gupta, Neeraj Motwani, and Pranjal Shrivastava, Microsoft; Qiang Su, The Chinese University of Hong Kong; Anil Reddy Pannala, Kristina Moore, James Grantham, Anupam Pandey, Xin Liu, Guohan Lu, Gerald De Grace, Rishabh Tewari, Lihua Yuan, Erica Lan, Deepak Bansal, and Dave Maltz, Microsoft; <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/yqx\/\"><em>Yongqiang Xiong<\/em><\/a><em>, Microsoft Research Asia; Hong Xu, The Chinese University of Hong Kong<\/em><\/p>\n\n\n\n<p>SONiC DASH SmartSwitch redesigns cloud network offloading with a hardware-friendly pipeline, unified switch architecture, and open development model while addressing key scalability and deployment challenges. Deployed at scale in Azure, it delivers high throughput and connection capacity while significantly improving power and space efficiency.<\/p>\n\n\n\n<div style=\"height:20px\" aria-hidden=\"true\" class=\"wp-block-spacer\"><\/div>\n\n\n\n<p><em>Wednesday, May 6, 3:50\u20135:30 PM<\/em><\/p>\n\n\n\n<h3 class=\"wp-block-heading h5\" id=\"krakenguard-towards-fine-grained-ebpf-isolation\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/www.usenix.org\/conference\/nsdi26\/presentation\/patel\">KRAKENGUARD: Towards Fine-Grained eBPF Isolation<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/h3>\n\n\n\n<p><em>Jainil Patel, IIT Roorkee; Lucas Graeff Buhl-Nielsen, Quantco; <\/em><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/adrienghosn\/\"><em>Adrien Ghosn<\/em><\/a><em>, Microsoft; Marios Kogias, Imperial College London<\/em><\/p>\n\n\n\n<p>KRAKENGUARD enforces fine-grained, policy-based controls on eBPF programs at load time using symbolic execution, enabling safe use in multi-tenant environments without relying on coarse Linux capabilities. It prevents malicious behavior, detects vulnerabilities, and allows for secure execution of untrusted programs with strong isolation guarantees.<\/p>\n\n\n\n<div style=\"height:20px\" aria-hidden=\"true\" class=\"wp-block-spacer\"><\/div>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"symposium-organizers-from-microsoft\">Symposium organizers from Microsoft<\/h2>\n\n\n\n<h3 class=\"wp-block-heading\" id=\"program-committee\">Program Committee<\/h3>\n\n\n\n<p><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/ga\/\">Ganesh Ananthanarayanan<\/a><br><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/bearzani\/\">Behnaz Arzani<\/a><br><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/hiballan\/\">Hitesh Ballani<\/a><br><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/rybecket\/\">Ryan Beckett<\/a><br><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/ranveer\/\">Ranveer Chandra<\/a><br><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/pcosta\/\">Paolo Costa<\/a><br><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/rofons\/\">Rodrigo Fonseca<\/a><br><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/xefouk\/\">Xenofon Foukas<\/a><br><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/kevhsieh\/\">Kevin Hsieh<\/a><br><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" rel=\"noopener noreferrer\" target=\"_blank\" href=\"https:\/\/www.linkedin.com\/in\/ukrishnas\/\">Umesh Krishnaswamy<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><br><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/jingliu3\/\">Jing Liu<\/a><br><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/jonathanmace\/\">Jonathan Mace<\/a><br><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/dmaltz\/\">Dave Maltz<\/a><br><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/samani\/\">Sathiya Mani<\/a><br><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/dnarayan\/\">Dushyanth Narayanan<\/a><br><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/sumann\/\">Suman Nath<\/a><br><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/ramjee\/\">Ram Ramjee<\/a><br><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/ssaroiu\/\">Stefan Saroiu<\/a><\/p>\n\n\n\n<h3 class=\"wp-block-heading\" id=\"steering-committee\">Steering Committee<\/h3>\n\n\n\n<p><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/sujbanerjee\/\">Sujata Banerjee<\/a><br><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/lorch\/\">Jay Lorch<\/a><\/p>\n\n\n\n<div style=\"height:35px\" aria-hidden=\"true\" class=\"wp-block-spacer\"><\/div>\n","protected":false},"excerpt":{"rendered":"<p>Microsoft researchers share advances in building and operating large-scale distributed systems, spanning datacenters, networking, and the growing intersection with AI during  NSDI \u201926.<\/p>\n","protected":false},"author":44124,"featured_media":1170639,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","msr-author-ordering":[{"type":"user_nicename","value":"Sujata Banerjee","user_id":"43911"}],"msr_hide_image_in_river":0,"footnotes":""},"categories":[1],"tags":[],"research-area":[13556,13547],"msr-region":[],"msr-event-type":[],"msr-locale":[268875],"msr-post-option":[],"msr-impact-theme":[],"msr-promo-type":[],"msr-podcast-series":[],"class_list":["post-1170563","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-research-blog","msr-research-area-artificial-intelligence","msr-research-area-systems-and-networking","msr-locale-en_us"],"msr_event_details":{"start":"","end":"","location":""},"podcast_url":"","podcast_episode":"","msr_research_lab":[199560,199565],"msr_impact_theme":[],"related-publications":[],"related-downloads":[],"related-videos":[],"related-academic-programs":[],"related-groups":[144812,144899,144927,282170,442950,815140],"related-projects":[1017939,982158],"related-events":[],"related-researchers":[{"type":"user_nicename","value":"Sujata Banerjee","user_id":43911,"display_name":"Sujata Banerjee","author_link":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/sujbanerjee\/\" aria-label=\"Visit the profile page for Sujata Banerjee\">Sujata Banerjee<\/a>","is_active":false,"last_first":"Banerjee, Sujata","people_section":0,"alias":"sujbanerjee"}],"msr_type":"Post","featured_image_thumbnail":"<img width=\"960\" height=\"540\" src=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1-960x540.jpg\" class=\"img-object-cover\" alt=\"NSDI \u201926 logo in white, centered on a smooth gradient background transitioning from blue to purple and pink.\" decoding=\"async\" loading=\"lazy\" srcset=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1-960x540.jpg 960w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1-300x169.jpg 300w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1-1024x576.jpg 1024w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1-768x432.jpg 768w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1-1066x600.jpg 1066w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1-655x368.jpg 655w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1-240x135.jpg 240w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1-640x360.jpg 640w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1-1280x720.jpg 1280w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/05\/NSDI2026-BlogHeroFeature-1400x788-1.jpg 1400w\" sizes=\"auto, (max-width: 960px) 100vw, 960px\" \/>","byline":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/sujbanerjee\/\" title=\"Go to researcher profile for Sujata Banerjee\" aria-label=\"Go to researcher profile for Sujata Banerjee\" data-bi-type=\"byline author\" data-bi-cN=\"Sujata Banerjee\">Sujata Banerjee<\/a>","formattedDate":"May 5, 2026","formattedExcerpt":"Microsoft researchers share advances in building and operating large-scale distributed systems, spanning datacenters, networking, and the growing intersection with AI during NSDI \u201926.","locale":{"slug":"en_us","name":"English","native":"","english":"English"},"_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/posts\/1170563","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/users\/44124"}],"replies":[{"embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/comments?post=1170563"}],"version-history":[{"count":23,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/posts\/1170563\/revisions"}],"predecessor-version":[{"id":1170712,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/posts\/1170563\/revisions\/1170712"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media\/1170639"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=1170563"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/categories?post=1170563"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/tags?post=1170563"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=1170563"},{"taxonomy":"msr-region","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-region?post=1170563"},{"taxonomy":"msr-event-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-event-type?post=1170563"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=1170563"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=1170563"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=1170563"},{"taxonomy":"msr-promo-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-promo-type?post=1170563"},{"taxonomy":"msr-podcast-series","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-podcast-series?post=1170563"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}