Details
Originalsprache | Englisch |
---|---|
Aufsatznummer | 5415864 |
Seiten (von - bis) | 59-86 |
Seitenumfang | 28 |
Fachzeitschrift | IEEE Communications Surveys and Tutorials |
Jahrgang | 12 |
Ausgabenummer | 1 |
Publikationsstatus | Veröffentlicht - Feb. 2010 |
Abstract
In recent years service curves have proven a powerful and versatile model for performance analysis of network elements, such as schedulers, links, and traffic shapers, up to entire computer networks, like the Internet. The elegance of the concept of service curve is due to intuitive convolution formulas that determine the data departures of a system from its arrivals and its service curve. This fundamental relation constitutes the basis of the network calculus and relates it to systems theory, however, under a different, so-called min-plus algebra. As in systems theory, the particular strength of the minplus convolution is the ability to concatenate tandem systems along a network path. This facilitates the notion of network service curve that has the expressiveness to characterize whole networks by a single transfer function. This paper surveys the state-of-the-art of the deterministic and the recent probabilistic network calculus. It discusses the concept of service curves, its use in the network calculus, and the relation to systems theory under the min-plus algebra. Service curve models of common schedulers and different types of networks are reviewed and methods for identification of a system's service curve representation from measurements are discussed. After recapitulating the state of knowledge on time-varying min-plus systems theory, stochastic service curve models are surveyed. These models allow utilizing the statistical multiplexing gain in a network calculus framework that features end-to-end network analysis by convolution of service curves.
ASJC Scopus Sachgebiete
- Ingenieurwesen (insg.)
- Elektrotechnik und Elektronik
Zitieren
- Standard
- Harvard
- Apa
- Vancouver
- BibTex
- RIS
in: IEEE Communications Surveys and Tutorials, Jahrgang 12, Nr. 1, 5415864, 02.2010, S. 59-86.
Publikation: Beitrag in Fachzeitschrift › Artikel › Forschung › Peer-Review
}
TY - JOUR
T1 - A Survey of Deterministic and Stochastic Service Curve Models in the Network Calculus
AU - Fidler, Markus
N1 - Funding information: Manuscript received 26 February 2009; revised 31 August 2009. This work has been supported by an Emmy Noether grant of the German Research Foundation (DFG). Markus Fidler is with the Institute of Communications Technology, Leibniz Universität Hannover. Digital Object Identifier 10.1109/SURV.2010.020110.00019
PY - 2010/2
Y1 - 2010/2
N2 - In recent years service curves have proven a powerful and versatile model for performance analysis of network elements, such as schedulers, links, and traffic shapers, up to entire computer networks, like the Internet. The elegance of the concept of service curve is due to intuitive convolution formulas that determine the data departures of a system from its arrivals and its service curve. This fundamental relation constitutes the basis of the network calculus and relates it to systems theory, however, under a different, so-called min-plus algebra. As in systems theory, the particular strength of the minplus convolution is the ability to concatenate tandem systems along a network path. This facilitates the notion of network service curve that has the expressiveness to characterize whole networks by a single transfer function. This paper surveys the state-of-the-art of the deterministic and the recent probabilistic network calculus. It discusses the concept of service curves, its use in the network calculus, and the relation to systems theory under the min-plus algebra. Service curve models of common schedulers and different types of networks are reviewed and methods for identification of a system's service curve representation from measurements are discussed. After recapitulating the state of knowledge on time-varying min-plus systems theory, stochastic service curve models are surveyed. These models allow utilizing the statistical multiplexing gain in a network calculus framework that features end-to-end network analysis by convolution of service curves.
AB - In recent years service curves have proven a powerful and versatile model for performance analysis of network elements, such as schedulers, links, and traffic shapers, up to entire computer networks, like the Internet. The elegance of the concept of service curve is due to intuitive convolution formulas that determine the data departures of a system from its arrivals and its service curve. This fundamental relation constitutes the basis of the network calculus and relates it to systems theory, however, under a different, so-called min-plus algebra. As in systems theory, the particular strength of the minplus convolution is the ability to concatenate tandem systems along a network path. This facilitates the notion of network service curve that has the expressiveness to characterize whole networks by a single transfer function. This paper surveys the state-of-the-art of the deterministic and the recent probabilistic network calculus. It discusses the concept of service curves, its use in the network calculus, and the relation to systems theory under the min-plus algebra. Service curve models of common schedulers and different types of networks are reviewed and methods for identification of a system's service curve representation from measurements are discussed. After recapitulating the state of knowledge on time-varying min-plus systems theory, stochastic service curve models are surveyed. These models allow utilizing the statistical multiplexing gain in a network calculus framework that features end-to-end network analysis by convolution of service curves.
KW - Min-plus systems theory
KW - Network calculus
KW - Network performance analysis
KW - Scheduling
KW - Service curves
UR - http://www.scopus.com/inward/record.url?scp=77249130718&partnerID=8YFLogxK
U2 - 10.1109/SURV.2010.020110.00019
DO - 10.1109/SURV.2010.020110.00019
M3 - Article
AN - SCOPUS:77249130718
VL - 12
SP - 59
EP - 86
JO - IEEE Communications Surveys and Tutorials
JF - IEEE Communications Surveys and Tutorials
SN - 1553-877X
IS - 1
M1 - 5415864
ER -