Ändra sökning
RefereraExporteraLänk till posten
Permanent länk

Direktlänk
Referera
Referensformat
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • apa.csl
  • Annat format
Fler format
Språk
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Annat språk
Fler språk
Utmatningsformat
  • html
  • text
  • asciidoc
  • rtf
The eXpress Data Path: Fast Programmable Packet Processing in the Operating System Kernel
Karlstads universitet, Fakulteten för hälsa, natur- och teknikvetenskap (from 2013), Institutionen för matematik och datavetenskap (from 2013). (DISCO)ORCID-id: 0000-0001-5241-6815
Red Hat.
Cilium.io.
Cilium.io.
Visa övriga samt affilieringar
2018 (Engelska)Ingår i: CoNEXT '18 Proceedings of the 14th International Conference on emerging Networking EXperiments and Technologies, Association for Computing Machinery (ACM), 2018, s. 54-66Konferensbidrag, Publicerat paper (Refereegranskat)
Abstract [en]

Programmable packet processing is increasingly implemented using kernel bypass  techniques, where a userspace application takes complete control of the  networking hardware to avoid expensive context switches between kernel and  userspace. However, as the operating system is bypassed, so are its  application isolation and security mechanisms; and well-tested configuration,  deployment and management tools cease to function.  To overcome this limitation, we present the design of a novel approach to  programmable packet processing, called the eXpress Data Path (XDP). In XDP,  the operating system kernel itself provides a safe execution environment for  custom packet processing applications, executed in device driver context. XDP  is part of the mainline Linux kernel and provides a fully integrated solution  working in concert with the kernel's networking stack. Applications are  written in higher level languages such as C and compiled into custom byte code  which the kernel statically analyses for safety, and translates into native  instructions.  We show that XDP achieves single-core packet processing performance as high as  24 million packets per second, and illustrate the flexibility of the  programming model through three example use cases: layer-3 routing, inline  DDoS protection and layer-4 load balancing.

Ort, förlag, år, upplaga, sidor
Association for Computing Machinery (ACM), 2018. s. 54-66
Nyckelord [en]
XDP, BPF, Programmable Networking, DPDK
Nationell ämneskategori
Datavetenskap (datalogi)
Identifikatorer
URN: urn:nbn:se:kau:diva-69639DOI: 10.1145/3281411.3281443ISI: 000455383800006ISBN: 978-1-4503-6080-7 (digital)OAI: oai:DiVA.org:kau-69639DiVA, id: diva2:1256181
Konferens
CoNEXT '18: International Conference on emerging Networking EXperiments and Technologies
Tillgänglig från: 2018-10-16 Skapad: 2018-10-16 Senast uppdaterad: 2019-11-09Bibliografiskt granskad
Ingår i avhandling
1. Bufferbloat and Beyond: Removing Performance Barriers in Real-World Networks
Öppna denna publikation i ny flik eller fönster >>Bufferbloat and Beyond: Removing Performance Barriers in Real-World Networks
2018 (Engelska)Doktorsavhandling, sammanläggning (Övrigt vetenskapligt)
Abstract [en]

The topic of this thesis is the performance of computer networks. While network performance has generally improved with time, over the last several years we have seen examples of performance barriers limiting network performance. In this work we explore such performance barriers and look for solutions.

The problem of excess persistent queueing latency, known as bufferbloat, serves as our starting point; we examine its prevalence in the public internet, and evaluate solutions for better queue management, and explore how to improve on existing solutions to make them easier to deploy.

Since an increasing number of clients access the internet through WiFi networks, examining WiFi performance is a natural next step. Here we also look at bufferbloat, as well as the so-called performance anomaly, where stations with poor signal strengths can severely impact the performance of the whole network. We present solutions for both of these issues, and additionally design a mechanism for assigning policies for distributing airtime between devices on a WiFi network. We also analyse the “TCP Small Queues” latency minimisation technique implemented in the Linux TCP stack and optimise its performance over WiFi networks.

Finally, we explore how high-speed network processing can be enabled in software, by looking at the eXpress Data Path framework that has been gradually implemented in the Linux kernel as a way to enable high-performance programmable packet processing directly in the operating system’s networking stack.

A special focus of this work has been to ensure that the results are carried forward to the implementation stage, which is achieved by releasing implementations as open source software. This includes parts that have been accepted into the Linux kernel, as well as a separate open source measurement tool, called Flent, which is used to perform most of the experiments presented in this thesis, and also used widely in the bufferbloat community.

Abstract [en]

The topic of this thesis is the performance of computer networks in general, and the internet in particular. While network performance has generally improved with time, over the last several years we have seen examples of performance barriers limiting network performance. In this work we explore such performance barriers and look for solutions.

Our exploration takes us through three areas where performance barriers are found: The bufferbloat phenomenon of excessive queueing latency, the performance anomaly in WiFi networks and related airtime resource sharing problems, and the problem of implementing high-speed programmable packet processing in an operating system. In each of these areas we present solutions that significantly advance the state of the art.

The work in this thesis spans all three aspects of the field of computing, namely mathematics, engineering and science. We perform mathematical analysis of algorithms, engineer solutions to the problems we explore, and perform scientific studies of the network itself. All our solutions are implemented as open source software, including both contributions to the upstream Linux kernel, as well as the Flent test tool, developed to support the measurements performed in the rest of the thesis.

Ort, förlag, år, upplaga, sidor
Karlstad: Karlstads universitet, 2018
Serie
Karlstad University Studies, ISSN 1403-8099 ; 2018:42
Nyckelord
Bufferbloat, AQM, WiFi, XDP, TSQ, Flent, network measurement, performance evaluation, fairness, queueing, programmable packet processing
Nationell ämneskategori
Datavetenskap (datalogi)
Forskningsämne
Datavetenskap
Identifikatorer
urn:nbn:se:kau:diva-69416 (URN)978-91-7063-878-7 (ISBN)978-91-7063-973-9 (ISBN)
Disputation
2018-11-23, 21A342, Eva Erikssonsalen, Karlstad, 09:15 (Engelska)
Opponent
Handledare
Projekt
HITS, 4707
Forskningsfinansiär
KK-stiftelsen
Anmärkning

Paper 6 was published as manuscript in the thesis.

The revised fulltext is identical to the original version with the exception that printing errors have been removed.

Tillgänglig från: 2018-10-26 Skapad: 2018-09-27 Senast uppdaterad: 2020-06-09Bibliografiskt granskad

Open Access i DiVA

fulltext(2320 kB)2687 nedladdningar
Filinformation
Filnamn FULLTEXT01.pdfFilstorlek 2320 kBChecksumma SHA-512
a501f54189aff2a0484c5de0377231ce652088667efb867dcc534794691b8246cf8920419323ff8ef8045cff2c4bb4511f452246709faefb9bc2bd29d48ec009
Typ fulltextMimetyp application/pdf

Övriga länkar

Förlagets fulltext

Person

Høiland-Jørgensen, Toke

Sök vidare i DiVA

Av författaren/redaktören
Høiland-Jørgensen, Toke
Av organisationen
Institutionen för matematik och datavetenskap (from 2013)
Datavetenskap (datalogi)

Sök vidare utanför DiVA

GoogleGoogle Scholar
Totalt: 2690 nedladdningar
Antalet nedladdningar är summan av nedladdningar för alla fulltexter. Det kan inkludera t.ex tidigare versioner som nu inte längre är tillgängliga.

doi
isbn
urn-nbn

Altmetricpoäng

doi
isbn
urn-nbn
Totalt: 3085 träffar
RefereraExporteraLänk till posten
Permanent länk

Direktlänk
Referera
Referensformat
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • apa.csl
  • Annat format
Fler format
Språk
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Annat språk
Fler språk
Utmatningsformat
  • html
  • text
  • asciidoc
  • rtf