Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Telco Distributed DC with Transport Protocol Enhancement for 5G Mobile Networks: A Survey
Karlstad University, Faculty of Health, Science and Technology (starting 2013), Department of Mathematics and Computer Science. (istributed Systems and Communications Research Group (DISCO))
Karlstad University, Faculty of Economic Sciences, Communication and IT, Department of Computer Science. Karlstad University, Faculty of Health, Science and Technology (starting 2013), Department of Mathematics and Computer Science. (Distributed Systems and Communications Research Group (DISCO))ORCID iD: 0000-0003-4147-9487
2017 (English)Report (Other academic)
Abstract [en]

Distributed data center hosts telco virtual network functions, mixing workloads that require data transport through transport protocols with either low end-to-end latency or large bandwidth for high throughput, e.g., from tough requirements in 5G use cases. A trend is the use relatively inexpensive, off-the-shelf switches in data center networks, where the dominated transport traffic is TCP traffic. Today’s TCP protocol will not be able to meet such requirements. The transport protocol evolution is driven by transport performance (latency and throughput) and robust enhancements in data centers, which include new transport protocols and protocol extensions such as DCTCP, MPTCP and QUIC protocols and lead to intensive standardization works and contributions to 3GPP and IETF.

By implementing ECN based congestion control instead of the packet-loss based TCP AIMD congestion control algorithm, DCTCP not only solves the latency issue in TCP congestion control caused by the switch buffer bloating but also achieves an improved performance on the packet loss and throughput. The DCTCP can also co-exist with normal TCP by applying a modern coupled queue management algorithm in the switches of DC networks, which fulfills IETF L4S architecture. MPTCP is an extension to TCP, which can be implemented in DC’s Fat tree architecture to improve transport throughput and shorten the latency by mitigating the bandwidth issue caused by TCP connection collision within the data center. The QUIC is a reliable and multiplexed transport protocol over UDP transport, which includes many of the latest transport improvements and innovation, which can be used to improve the transport performance on streaming media delivery.

The Clos topology is a commonly used network topology in a distributed data center. In the Clos architecture, an over-provisioned fabric cannot handle full wire-speed traffic, thus there is a need to have a mechanism to handle overload situations, e.g., by scaling out the fabric. However, this will introduce more end-to-end latency in those cases the switch buffer is bloated, and will cause transport flow congestion.

In this survey paper, DCTCP, MPTCP and QUIC are discussed as solutions for transport performance enhancement for 5G mobile networks to avoid the transport flow congestion caused by the switch buffer bloating from overloaded switch queue in data centers. 

Place, publisher, year, edition, pages
Karlstad: Karlstads universitet, 2017. , p. 54
Keyword [en]
5G, mobile, transport protocol, data center, dc, tcp, mptcp, dctcp, ecn, quic
National Category
Telecommunications
Research subject
Computer Science
Identifiers
URN: urn:nbn:se:kau:diva-65371ISBN: 978-91-7063-830-5 (print)ISBN: 978-91-7063-925-8 (electronic)OAI: oai:DiVA.org:kau-65371DiVA, id: diva2:1164941
Projects
High Quality Networked Services in a Mobile World (HITS)
Funder
Knowledge Foundation
Note

Working paper, December 2017

Available from: 2017-12-12 Created: 2017-12-12 Last updated: 2017-12-21Bibliographically approved

Open Access in DiVA

fulltext(3345 kB)153 downloads
File information
File name FULLTEXT01.pdfFile size 3345 kBChecksum SHA-512
b23ef8ff48d3e357bc30a13c1ae0a918e23711068a8fb8d1e501c2da05d4d9d4cbe3784ef1d3caa72c66c5f22df6fab88ef1fc3a2772fee90892bc2020ef6572
Type fulltextMimetype application/pdf

Search in DiVA

By author/editor
Cheng, JunGrinnemo, Karl-Johan
By organisation
Department of Mathematics and Computer ScienceDepartment of Computer Science
Telecommunications

Search outside of DiVA

GoogleGoogle Scholar
Total: 153 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

isbn
urn-nbn

Altmetric score

isbn
urn-nbn
Total: 266 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf