Performance evaluation of bandwidth allocation methods in a geostationary satellite channel in the presence of internet traffic
Introduction
TCP/IP based protocols and related networks are the most rapidly spreading technology. Many new applications use these. On the other hand, satellite networks are an essential element in the establishment of long distance communications, and will have a major role in the implementation of the so called global information infrastructure in the future [1]. Therefore, Internet-based applications represent most traffic over satellite networks. As a consequence adapting a bandwidth allocation scheme which, in satellite environments, has a function as a fading countermeasure, is a key issue.
The reference network for this paper is based on a Geostationary (GEO) satellite accessed by a number (I) of earth stations. Earth stations may be affected by fading. One of them (or the satellite itself, if switching on board is allowed), called the master station, assigns the overall bandwidth among all earth stations. The assignment depends on a cost function whose value is related to the fading level and to the traffic load. The traffic is composed of guaranteed calls (e.g., CBR, Constant Bit Rate, phone calls) and of TCP and UDP flows. The bandwidth assigned to each single station is divided into two portions. One part is dedicated to guaranteed calls (CBR traffic). Once the bandwidth is fixed for them, the maximum number of guaranteed calls that can be accepted in one station is fixed and is used within a CAC scheme. In other words, just to make an example, if the bandwidth portion dedicated to guaranteed calls in one station is 1280 kbit/s and each CBR call has a bit rate of 128 kbit/s, it means that the station can accept 10 CBR calls at most. This rule is applied by the CAC that decides if CBR calls can be admitted in the station or not. The other bandwidth portion is dedicated to TCP/IP traffic, which is the main focus of the paper.
TCP/IP traffic may be divided into TCP and UDP flows. Their behaviour is very different concerning congestion countermeasures. UDP behaviour is independent of what happens in the network. TCP reduces its rate if congestion is detected. Many traffic models and resource allocation schemes do not consider the different reaction to congestion and do not distinguish between TCP and UDP traffic. They model Internet flows as a mere superposition of TCP/UDP sources. The assumption is correct if the number of flows is so large as to be assumed infinite but, in real conditions, the differentiation between TCP and UDP traffic in the bandwidth control scheme is a way to save bandwidth and to have it available in case of need (i.e., for a faded station).
The paper takes its origin from a bandwidth allocation scheme, called CAP-ABASC [2] and identified as CAP-1 in the following, where the TCP and UDP flows are not differentiated for bandwidth allocation and proposes a scheme, called E-CAP-ABASC (Extended-CAP-ABASC), where TCP and UDP are distinguished through two different packet loss probability models appearing in the cost function that determines bandwidth allocation.
Fading is modelled by assigning a probability of channel degradation to each satellite link, along with a weighting coefficient to “measure” the degradation itself. The degradation is seen as a bandwidth reduction. Technically it may be due to the use of fading compensation techniques such as Forward Error Correction (FEC) schemes.
The paper includes a detailed performance evaluation section that allows investigating the main characteristics of E-CAP-ABASC by varying traffic and channel conditions.
The paper is structured as follows: Section 2 contains the description of the network topology, of the channel model and of the general structure of the allocation scheme. Section 3 describes the system architecture and the features of E-CAP-ABASC. Section 4 shows the performance evaluation and Section 5 lists the conclusions.
Section snippets
Network topology
The GEO satellite network considered is composed of I earth stations connected through a mesh topology. One of them (or, the satellite) is the “master” and controls the bandwidth assignment to the single stations which gather traffic from the users. The study is not linked to a particular bandwidth choice but Ka-band (20–30 GHz) may be the technological reference because here the effect of fading is particularly important.
The channel model
The fading effect is modelled as bandwidth reduction. Mathematically, this
Bandwidth allocation algorithms and traffic source models
As said, the expression that appears in (8) actually characterizes the allocation algorithm. Three algorithms are compared in this paper: the first one has been introduced in [2]. The algorithm, called originally CAP-ABASC, is referenced as CAP-1 in the following. is computed without performing any distinction between TCP and UDP, whose flows are conveyed towards the same buffer.
The second proposal is called CAP-2. It is the first step towards the differentiation of the TCP
Performance comparison
The aim of this evaluation is to compare the performance of the E-CAP-ABASC allocation (called E-CAP in the reminder of the paper) in terms of packet loss probability and bandwidth allocation with CAP-1 and CAP-2. To reach this goal, the allocation strategies have been implemented in C++ language. The minimization algorithm used to solve the allocation problems is based on a dynamic programming procedure [10]. The performance evaluation is carried out by varying traffic and fading conditions.
Conclusions
The paper presents three bandwidth allocation schemes for the satellite environment. One of them (CAP-1) considers UDP and TCP traffic in undifferentiated fashion. CAP-2 separates the two Internet traffic types but uses the same traffic model. E-CAP exploits traffic separation and applies a different model to compute the TCP packet loss probability. The performance evaluation contains an extensive comparison of the three approaches. The full distinction of TCP and UDP traffic gives evident
Igor Bisio was born in Novi Ligure (Alessandria), Italy in 1978. He got his “Laurea” degree at the University of Genoa, Italy in 2002. and his Ph.D. degree in “Information and Communication Sciences” at the University of Genoa in 2006. His Ph.D. course was funded by the Italian Consortium of Telecommunications (CNIT). He worked for both CNIT University of Genoa Research Unit and Selex Communications s.p.a., from 2002 to 2006 as research consultant. He is currently in a Research Assistant
References (11)
- et al.
The role of satellites in global IT: Trends and implications
IEEE Personal Communications
(2001) - R. Bolla, F. Davoli, M. Marchese, Adaptive bandwidth allocation methods in the satellite environment, in: Proceedings...
- et al.
On self-similar traffic in ATM queues: Definition, overflow probability bound, and cell delay distribution
IEEE/ACM Transaction on Networking
(1997) - et al.
On the self-similar nature of ethernet traffic
IEEE/ACM Transaction on Networking
(1994) - et al.
Wide-area traffic: The failure of poisson modeling
IEEE/ACM Transaction on Networking
(1995)
Cited by (28)
Performance analysis of the TCP behavior in a geo satellite environment
2001, Computer CommunicationsA Study of Transmission Overhead of a Hybrid Bundle Retransmission Approach for Deep-Space Communications
2022, IEEE Transactions on Aerospace and Electronic SystemsFlexible Resource Allocation with Inter-Beam Interference in Satellite Communication Systems with a Digital Channelizer
2020, IEEE Transactions on Wireless CommunicationsAnalysis of Licklider Transmission Protocol for Reliable File Delivery in Space Vehicle Communications with Random Link Interruptions
2019, IEEE Transactions on Vehicular TechnologyAn adaptive beam control technique for Q band satellite to maximize diversity gain and mitigate interference to terrestrial networks
2019, IEEE Transactions on Emerging Topics in ComputingQueueing analysis of DTN protocols in deep-space communications
2018, IEEE Aerospace and Electronic Systems Magazine
Igor Bisio was born in Novi Ligure (Alessandria), Italy in 1978. He got his “Laurea” degree at the University of Genoa, Italy in 2002. and his Ph.D. degree in “Information and Communication Sciences” at the University of Genoa in 2006. His Ph.D. course was funded by the Italian Consortium of Telecommunications (CNIT). He worked for both CNIT University of Genoa Research Unit and Selex Communications s.p.a., from 2002 to 2006 as research consultant. He is currently in a Research Assistant Position and he is member of the Telecommunication Networking Research Group and, in particular, of the Satellite Communications and Networking Laboratory research staff at the University of Genoa. He is also an IEEE Student Member since 2004 and IEEE Satellite and Space Communications Technical Committee Member since 2005.
He is author of more than 30 scientific papers and he is recipient of the “Globecom 2006 Best Student Paper Award” in the Physical Communications Systems Category.
His main research activity concerns: Resource Allocation and Management for Satellite Communication systems, Optimization Algorithms and Architectures for Satellite Sensor Networks, Traffic Modelling.
Mario Marchese was born in Genoa, Italy in 1967. He got his ”Laurea” degree cum laude at the University of Genoa, Italy in 1992 and the Qualification as Professional Engineer in April 1992. He obtained his Ph.D. (Italian “Dottorato di Ricerca”) degree in ”Telecommunications” at the University of Genoa in 1996. From 1999 to 2004, he worked with the Italian Consortium of Telecommunications (CNIT), by the University of Genoa Research Unit, where he was Head of Research. From February 2005 he has been Associate Professor at the University of Genoa, Department of Communication, Computer and Systems Science (DIST). He is the founder and still the technical responsible of CNIT/DIST Satellite Communications and Networking Laboratory (SCNL) by the University of Genoa, which contains high value devices and tools and implies the management of different units of specialized scientific and technical personnel.
He was the Official Representative of CNIT within the European Telecommunications Standard Institute (ETSI) from 1999 to 2005, he is active member in the Satellite Earth Station (SES) Broadband Satellite Multimedia (BSM). He is Chair of the IEEE Satellite and Space Communications Technical Committee and Senior Member of the IEEE. He is author and co-author of more than 150 scientific works, including international magazines, international conferences and book chapters. He is the author of the book “Quality of Service over Heterogeneous Networks”, John Wiley & Sons, Chichester, 2007. He is Associate Editor of the International Journal of Communication Systems (Wiley) and Technical Committee Co-Chair of various international conferences, including Globecom and ICC.
His main research activity concerns: Satellite and Radio Networks, Transport Layer over Satellite and Wireless Networks, Quality of Service over ATM, IP and MPLS, Data Transport over Heterogeneous Networks, Emulation and Simulation of Telecommunication Networks.