NEAT: NEtwork dATa plane measurements as first class primitive
Lead Research Organisation:
Queen Mary University of London
Department Name: Sch of Electronic Eng & Computer Science
Abstract
Data centre networks are poorly equipped to rapidly spot and address failures, resulting in countless well-documented application performance degradation or outages. This is because the investigation process is performed in centralised commodity servers (collectors) that do not have per-packet visibility, but instead aggregated and sampled statistics from the data plane. The NEAT project will address this deficiency by moving traffic analysis directly into switches that have per-packet visibility. Exploiting advances in programmable hardware, e.g. P4, NEAT will rethink data plane operation and will transform switches from just packet forwarder with limited monitoring capabilities to more intelligent systems capable of analysing traffic and exporting only relevant results. This will enable the level of fine-grained data plane visibility required to allow operators to rapidly identify and adapt to changes in network conditions, which hurts applications.
Planned Impact
The NEAT research programme targets one of the most critical part of today's Internet: data centres. Data centres are a fundamental part of UK infrastructure, underpinning a wide range of activities across government, business, and society. Specifically, NEAT focuses on improving the current monitoring and traffic analysis practices by leveraging data plane programmability on top of an SDN control plane. In this context, my impact plan will reach and influence two main two main constituencies: (1) industry and organisations providing the data centre infrastructure, platforms and services and (2) standardisation bodies developing and promoting open standards in both the SDN and programmable data plane area.
Enabling network traffic analysis techniques in the context of data centres impact a number of players spanning from industry to organisations. The letters of support accompanying this proposal indicate the serious and deep interest in NEAT from a broad array of players: Cisco Systems is a leading networking company, which designs, manufactures, and sells network equipment; Barefoot Networks is another leading networking company, which designs and produces programmable network switch silicon, systems and software. VMware is a subsidiary of Dell Technologies, that provides cloud and vitalisation software and services; CERN is a European research organisation that operates the biggest particle physics
laboratory in the world, where a large computing facility is used to store and analyse data from experiments, as well as simulate events. Those players, given the mixture of expertise ranging from data centre operation to switch design, are key elements to foster innovation in data plane assisted network traffic analysis. They are eager to provide support and benefit from the output of my research by helping a smooth transition from the theory to the practice.
NEAT project will contribute to open standards for programmable networking and monitoring. Standardisation efforts in SDN and data plane programmability are currently thriving with a broad commercial and academic support. I will take part in activities and working groups such as the Open Networking Foundation (ONF), P4 and the Internet Research Task Force (IRTF) to emphasise the importance of data plane assisted traffic analysis in the next generation data centres. My former status of ONF Research Associate and the engagement with Barefoot Networks (see letter of support) opens the door for a fruitful collaboration towards new open standards.
The open source NetFPGA platform is also of particular relevance to NEAT. NetFPGA is an open source hardware/software system with high impact. The NetFPGA platform is a joint project between Stanford University and University of Cambridge, providing an ideal means to disseminate NEAT technology developments to a wide community of researchers, adopters, and investors. Leveraging my status of active member of the official NetFPGA development team and to stimulate a quick uptake of the ideas and prototypes developed under NEAT, I will make the code available under an open source licence. My goal is to build an active user group to eventually provide the research community and industry with mature instruments for exploring theoretical and practical aspects of the NEAT vision.
The novelty, timeliness, and relevance of the NEAT project will ensure a widespread interest in its results from the international community of practitioners, e.g. RIPE, NANOG, MENOG. There will also be abundant opportunities for the technology to be exploited by the industrial supporters. Through those supporters, NEAT will provide significant benefits to the UK industry, increasing its competitiveness in the growing data centre marketplace. The project will lead to high-profile publications and to exciting opportunities for future research. This will further increase the UK's aspirations for the development and use of SDN-enabled technologies.
Enabling network traffic analysis techniques in the context of data centres impact a number of players spanning from industry to organisations. The letters of support accompanying this proposal indicate the serious and deep interest in NEAT from a broad array of players: Cisco Systems is a leading networking company, which designs, manufactures, and sells network equipment; Barefoot Networks is another leading networking company, which designs and produces programmable network switch silicon, systems and software. VMware is a subsidiary of Dell Technologies, that provides cloud and vitalisation software and services; CERN is a European research organisation that operates the biggest particle physics
laboratory in the world, where a large computing facility is used to store and analyse data from experiments, as well as simulate events. Those players, given the mixture of expertise ranging from data centre operation to switch design, are key elements to foster innovation in data plane assisted network traffic analysis. They are eager to provide support and benefit from the output of my research by helping a smooth transition from the theory to the practice.
NEAT project will contribute to open standards for programmable networking and monitoring. Standardisation efforts in SDN and data plane programmability are currently thriving with a broad commercial and academic support. I will take part in activities and working groups such as the Open Networking Foundation (ONF), P4 and the Internet Research Task Force (IRTF) to emphasise the importance of data plane assisted traffic analysis in the next generation data centres. My former status of ONF Research Associate and the engagement with Barefoot Networks (see letter of support) opens the door for a fruitful collaboration towards new open standards.
The open source NetFPGA platform is also of particular relevance to NEAT. NetFPGA is an open source hardware/software system with high impact. The NetFPGA platform is a joint project between Stanford University and University of Cambridge, providing an ideal means to disseminate NEAT technology developments to a wide community of researchers, adopters, and investors. Leveraging my status of active member of the official NetFPGA development team and to stimulate a quick uptake of the ideas and prototypes developed under NEAT, I will make the code available under an open source licence. My goal is to build an active user group to eventually provide the research community and industry with mature instruments for exploring theoretical and practical aspects of the NEAT vision.
The novelty, timeliness, and relevance of the NEAT project will ensure a widespread interest in its results from the international community of practitioners, e.g. RIPE, NANOG, MENOG. There will also be abundant opportunities for the technology to be exploited by the industrial supporters. Through those supporters, NEAT will provide significant benefits to the UK industry, increasing its competitiveness in the growing data centre marketplace. The project will lead to high-profile publications and to exciting opportunities for future research. This will further increase the UK's aspirations for the development and use of SDN-enabled technologies.
Organisations
- Queen Mary University of London (Lead Research Organisation)
- Polytechnic University of Turin (Collaboration)
- University College London (Collaboration)
- HARVARD UNIVERSITY (Collaboration)
- Purdue University (Collaboration)
- Sapienza University of Rome (Collaboration)
- Budapest University of Technology and Economics (Collaboration)
- Brno University of Technology (Collaboration)
- Facebook (Collaboration)
- VMware, Inc. (Collaboration)
- Boston University (Collaboration)
- University of Cologne (Collaboration)
- University of Rome Tor Vergata (Collaboration)
- NEC Corporation (Collaboration)
- New York University (Collaboration)
- Technical University of Madrid (Collaboration)
- DELL (United States) (Project Partner)
- European Organization for Nuclear Research (Project Partner)
- Barefoot Networks (Project Partner)
- Cisco Systems (United States) (Project Partner)
Publications
Miano S
(2023)
Fast In-kernel Traffic Sketching in eBPF
in ACM SIGCOMM Computer Communication Review
Matousek J
(2022)
ClassBench-ng: Benchmarking Packet Classification Algorithms in the OpenFlow Era
in IEEE/ACM Transactions on Networking
Chiesa M
(2021)
Fast ReRoute on Programmable Switches
in IEEE/ACM Transactions on Networking
Wang T.
(2022)
Isolation Mechanisms for High-Speed Packet-Processing Pipelines
in Proceedings of the 19th USENIX Symposium on Networked Systems Design and Implementation, NSDI 2022
Sanaee A.
(2022)
Backdraft: a Lossless Virtual Switch that Prevents the Slow Receiver Problem
in Proceedings of the 19th USENIX Symposium on Networked Systems Design and Implementation, NSDI 2022
Siracusano G.
(2022)
Re-architecting Traffic Analysis with Neural Network Interface Cards
in Proceedings of the 19th USENIX Symposium on Networked Systems Design and Implementation, NSDI 2022
Monterubbiano A
(2023)
Lightweight Acquisition and Ranging of Flows in the Data Plane
in Proceedings of the ACM on Measurement and Analysis of Computing Systems
Langlet J
(2021)
Zero-CPU Collection with Direct Telemetry Access
Kucera J
(2020)
Detecting routing loops in the data plane
Shahinfar F
(2023)
Automatic Kernel Offload Using BPF
Description | During the project, myself and the research team has developed new techniques to improve visibility of datacenter networks. We developed new software and hardware solutions that helps to better understand what's happening in the network at run-time. This can be useful to detect attacks, for troubleshooting or simply to check if everything is working as expected. |
Exploitation Route | Some of the techniques implemented during the project have been cited and used by other researchers as well. https://github.com/multitenancy-project/menshen has 32 stars on Github and the paper cited 25 times so far (according to Google Scholar). https://github.com/jonlanglet/DTA has already 12 stars on Github albeit it was presented only in September 2023. The paper "Re-architecting traffic analysis with neural network interface cards" presented at USENIX NSDI 2022 has been already cited 54 times according to Google Scholar. |
Sectors | Digital/Communication/Information Technologies (including Software) |
Description | Advanced Network Offloads on Programmable NICs |
Organisation | New York University |
Country | United States |
Sector | Academic/University |
PI Contribution | I contributed with expertise and intellectual input. |
Collaborator Contribution | The partners contributed with expertise, intellectual input and access to equipment. |
Impact | paper: https://www.usenix.org/conference/nsdi22/presentation/wang-tao |
Start Year | 2021 |
Description | Automatic Optimization of Software Data Planes |
Organisation | Budapest University of Technology and Economics |
Country | Hungary |
Sector | Academic/University |
PI Contribution | I contributed with expertise, intellectual input and access to equipment. |
Collaborator Contribution | The partners contributed with expertise and intellectual input. |
Impact | Poster publication: https://dl.acm.org/doi/abs/10.1145/3405837.3411379 Poster publication: https://dl.acm.org/doi/10.1145/3485983.3493349 Paper publication: https://dl.acm.org/doi/10.1145/3503222.3507769 |
Start Year | 2020 |
Description | Automatic Optimization of Software Data Planes |
Organisation | NEC Corporation |
Department | NEC Laboratories Europe GmbH |
Country | Germany |
Sector | Private |
PI Contribution | I contributed with expertise, intellectual input and access to equipment. |
Collaborator Contribution | The partners contributed with expertise and intellectual input. |
Impact | Poster publication: https://dl.acm.org/doi/abs/10.1145/3405837.3411379 Poster publication: https://dl.acm.org/doi/10.1145/3485983.3493349 Paper publication: https://dl.acm.org/doi/10.1145/3503222.3507769 |
Start Year | 2020 |
Description | Automatic Optimization of Software Data Planes |
Organisation | Polytechnic University of Turin |
Country | Italy |
Sector | Academic/University |
PI Contribution | I contributed with expertise, intellectual input and access to equipment. |
Collaborator Contribution | The partners contributed with expertise and intellectual input. |
Impact | Poster publication: https://dl.acm.org/doi/abs/10.1145/3405837.3411379 Poster publication: https://dl.acm.org/doi/10.1145/3485983.3493349 Paper publication: https://dl.acm.org/doi/10.1145/3503222.3507769 |
Start Year | 2020 |
Description | Building better virtual switches |
Organisation | Purdue University |
Country | United States |
Sector | Academic/University |
PI Contribution | I contributed with expertise and intellectual input. |
Collaborator Contribution | The partners contributed with expertise, intellectual input and access to equipment. |
Impact | paper: https://dl.acm.org/doi/10.1145/3452296.3472914 |
Start Year | 2020 |
Description | Building better virtual switches |
Organisation | VMware, Inc. |
Country | United States |
Sector | Private |
PI Contribution | I contributed with expertise and intellectual input. |
Collaborator Contribution | The partners contributed with expertise, intellectual input and access to equipment. |
Impact | paper: https://dl.acm.org/doi/10.1145/3452296.3472914 |
Start Year | 2020 |
Description | Detecting Routing Loops in the Data Plane |
Organisation | Brno University of Technology |
Country | Czech Republic |
Sector | Academic/University |
PI Contribution | I contributed with expertise, intellectual input and access to equipment. |
Collaborator Contribution | The partners contributed with expertise, intellectual input and access to equipment. |
Impact | Paper publication: https://doi.org/10.1145/3386367.3431303 |
Start Year | 2020 |
Description | Detecting Routing Loops in the Data Plane |
Organisation | Harvard University |
Country | United States |
Sector | Academic/University |
PI Contribution | I contributed with expertise, intellectual input and access to equipment. |
Collaborator Contribution | The partners contributed with expertise, intellectual input and access to equipment. |
Impact | Paper publication: https://doi.org/10.1145/3386367.3431303 |
Start Year | 2020 |
Description | Disaggregated Sketches for Traffic Analysis |
Organisation | Boston University |
Country | United States |
Sector | Academic/University |
PI Contribution | I contributed with expertise and intellectual input. |
Collaborator Contribution | They contributed with expertise, intellectual input and equipment. |
Impact | Poster Publication: https://doi.org/10.1145/3386367.3431674 |
Start Year | 2020 |
Description | Disaggregated Sketches for Traffic Analysis |
Organisation | Harvard University |
Country | United States |
Sector | Academic/University |
PI Contribution | I contributed with expertise and intellectual input. |
Collaborator Contribution | They contributed with expertise, intellectual input and equipment. |
Impact | Poster Publication: https://doi.org/10.1145/3386367.3431674 |
Start Year | 2020 |
Description | Disaggregated Sketches for Traffic Analysis |
Organisation | University College London |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | I contributed with expertise and intellectual input. |
Collaborator Contribution | They contributed with expertise, intellectual input and equipment. |
Impact | Poster Publication: https://doi.org/10.1145/3386367.3431674 |
Start Year | 2020 |
Description | Disaggregated Sketches for Traffic Analysis |
Organisation | University of Rome Tor Vergata |
Country | Italy |
Sector | Academic/University |
PI Contribution | I contributed with expertise and intellectual input. |
Collaborator Contribution | They contributed with expertise, intellectual input and equipment. |
Impact | Poster Publication: https://doi.org/10.1145/3386367.3431674 |
Start Year | 2020 |
Description | Lightweight data structures for network monitoring |
Organisation | Sapienza University of Rome |
Country | Italy |
Sector | Academic/University |
PI Contribution | I contributed with expertise, intellectual input, hardware implementation and equipment. |
Collaborator Contribution | They contributed with expertise and intellectual input. |
Impact | Journal publication: https://dl.acm.org/doi/10.1145/3626775 |
Start Year | 2022 |
Description | Lightweight data structures for network monitoring |
Organisation | Technical University of Madrid |
Country | Spain |
Sector | Academic/University |
PI Contribution | I contributed with expertise, intellectual input, hardware implementation and equipment. |
Collaborator Contribution | They contributed with expertise and intellectual input. |
Impact | Journal publication: https://dl.acm.org/doi/10.1145/3626775 |
Start Year | 2022 |
Description | Lightweight data structures for network monitoring |
Organisation | University of Cologne |
Country | Germany |
Sector | Academic/University |
PI Contribution | I contributed with expertise, intellectual input, hardware implementation and equipment. |
Collaborator Contribution | They contributed with expertise and intellectual input. |
Impact | Journal publication: https://dl.acm.org/doi/10.1145/3626775 |
Start Year | 2022 |
Description | Scalable Network Collectors |
Organisation | |
Country | United States |
Sector | Private |
PI Contribution | I contributed with expertise, intellectual input and access to equipment. |
Collaborator Contribution | The partners contributed with expertise and intellectual input. |
Impact | paper: https://dl.acm.org/doi/10.1145/3484266.3487366 |
Start Year | 2021 |
Description | Scalable Network Collectors |
Organisation | Harvard University |
Country | United States |
Sector | Academic/University |
PI Contribution | I contributed with expertise, intellectual input and access to equipment. |
Collaborator Contribution | The partners contributed with expertise and intellectual input. |
Impact | paper: https://dl.acm.org/doi/10.1145/3484266.3487366 |
Start Year | 2021 |
Description | Scalable Network Collectors |
Organisation | University College London |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | I contributed with expertise, intellectual input and access to equipment. |
Collaborator Contribution | The partners contributed with expertise and intellectual input. |
Impact | paper: https://dl.acm.org/doi/10.1145/3484266.3487366 |
Start Year | 2021 |
Title | Backdraft: virtual switch |
Description | A new virtual switch for end-host losseless networking |
Type Of Technology | Webtool/Application |
Year Produced | 2022 |
Open Source License? | Yes |
Impact | paper accepted at a top-conference in networking/system area |
URL | https://github.com/Lossless-Virtual-Switching/Backdraft |
Title | Morpheus: Domain Specific Run Time Optimization for Software Data Planes - Artifact for ASPLOS'22 |
Description | This is the artifact for the "Morpheus: Domain Specific Run Time Optimization for Software Data Planes" paper published at ASPLOS'22. Morpheus is a system working alongside static compilers that continuously optimizes the targeted networking code. It introduces a number of new techniques, from static code analysis to adaptive code instrumentation, together with a toolbox of domain specific optimizations used to manipulate the code on-the-fly depending on runtime traffic patterns and control plane configurations. The Morpheus core exploits the LLVM compiler toolchain (v10.0.1) for code manipulation and run-time code generation. It works at the intermediate representation (IR) level as it allows to reason about the running code using a relatively high-level language framework without compromising on code generation time. This artifact contains the source code, the experimental workflow, and additional information to 1) compile and build Morpheus, 2) install the software dependencies and setup the testbed to run all the experiments, 3) the scripts that can be used to perform some of the experiments presented in the paper, and 4) the scripts to generate the plots based on the obtained results. For more information, please refer to https://github.com/Morpheus-compiler/Morpheus |
Type Of Technology | Software |
Year Produced | 2021 |
Open Source License? | Yes |
Impact | It is too early at the moment to result in notable impact. |
URL | https://zenodo.org/record/5830832 |
Title | RDMA APIs for Telemetry traffic |
Description | P4 code implementing our ideas described in the paper "Direct Telemetry Access" presented at ACM SIGCOMM 2023. |
Type Of Technology | Webtool/Application |
Year Produced | 2023 |
Open Source License? | Yes |
Impact | paper at a top-conference in the networking/system area |
Title | RMT extension for Isolation |
Description | Verilog code implementing an extension for RMT pipelines that enables isolation between programs |
Type Of Technology | Webtool/Application |
Year Produced | 2022 |
Impact | paper at a top-conference in the networking/system area |
URL | https://github.com/multitenancy-project/menshen |
Description | Guest Lecture at Purdue University |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Postgraduate students |
Results and Impact | I delivered a Guest lecture at Purdue University (CS 536). The purpose was to increase awareness of students on the performance implications around end-host networking and how programmable hardware can be used in this situation. The lecture sparked questions and a good discussion afterwards, with many students interested in the subject. |
Year(s) Of Engagement Activity | 2021 |
Description | Keynote talk at IEEE NetSoft |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Schools |
Results and Impact | Presentation about new primitives for telemetry collection as well as ways to improve telemetry network stacks. The talk sparked questions and a lengthy discussion afterwards |
Year(s) Of Engagement Activity | 2023 |
URL | https://netsoft2023.ieee-netsoft.org/program/keynotes |
Description | Presentation at Beyond SDN. Programmable data plane: Abstractions, Architectures and Applications (Workshop) |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Professional Practitioners |
Results and Impact | I presented the role of programmable hardware in the context of network monitoring. The talk sparked questions and a lengthy discussion afterwards |
Year(s) Of Engagement Activity | 2021 |
URL | https://www.drcn2021.polimi.it/program/workshops/workshop-3/ |
Description | Presentation at Brown University |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Postgraduate students |
Results and Impact | I presented the role of programmable hardware in the context of network monitoring. The talk sparked questions and a lengthy discussion afterwards |
Year(s) Of Engagement Activity | 2021 |
URL | https://systems.cs.brown.edu/sysread/2021-10-08%20%7C%20Gianni%20Antichi/ |
Description | Presentation at Cisco Switzerland |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Professional Practitioners |
Results and Impact | I presented the role of programmable hardware in the context of network monitoring. The talk sparked questions and a lengthy discussion afterwards |
Year(s) Of Engagement Activity | 2021 |
Description | Presentation at Compute-First Networking (Dagstuhl) |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Professional Practitioners |
Results and Impact | I presented the role of programmable hardware in the context of network monitoring. The talk sparked questions and a lengthy discussion afterwards |
Year(s) Of Engagement Activity | 2021 |
URL | https://www.dagstuhl.de/en/program/calendar/semhp/?semnr=21243 |
Description | Presentation at Huawei Research Centre (Munich) |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Industry/Business |
Results and Impact | I presented the role of programmable hardware in the context of network monitoring. The talk sparked questions and a lenghty discussion afterwards |
Year(s) Of Engagement Activity | 2020 |
Description | Presentation at Italian Networking Workshop |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | National |
Primary Audience | Professional Practitioners |
Results and Impact | Presentation about new primitives for telemetry collection. The talk sparked questions and a lengthy discussion afterwards |
Year(s) Of Engagement Activity | 2023 |
URL | https://inw2023.unibs.it/program/ |
Description | Presentation at Towards More Flexible and Automated Communication Networks (Dagstuhl) |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Professional Practitioners |
Results and Impact | I have talked about opportunities for understanding systems behavior and optimize their performance at run-time, automatically. The talk sparked questions and a lengthy discussion afterwards |
Year(s) Of Engagement Activity | 2022 |
URL | https://www.dagstuhl.de/en/seminars/seminar-calendar/seminar-details/22471 |
Description | Presentation at University of Trento |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Professional Practitioners |
Results and Impact | I presented the role of programmable hardware in the context of network monitoring. The talk sparked questions and a lengthy discussion afterwards |
Year(s) Of Engagement Activity | 2021 |