Optimizing IP networks in a hybrid IGP/MPLS environment

In this paper, we consider a traffic engineering (te) approach toip networks in a hybridigp/mpls environment. Thoughigp (Interior Gateway Protocol) routing has proven its scalability and reliability, effective traffic engineering has been difficult to achieve in public IP networks because of the limited functional capabilities of conventionalip technologies.mpls (Multi-Protocol Label Switching) on the one hand enhances the possibility to engineer traffic onip networks by allowing explicit routes. But on the other hand it suffers from the scalability (n-square) problem. Hybridigp/mpls approaches rely onip native routing as much as possible and usempls only if necessary. In this work we propose a novel hybrid traffic engineering method based on genetic algorithms, which can be considered as an offlinete approach to handle long or medium-term traffic variations in the range days, weeks or months. In our approach the maximum number of hops anlsp (Label Switched Path) may take and the number oflsps which are applied solely to improve the routing performance, are treated as constraints due to delay considerations and the complexity of management. We apply our method to the German scientific network (b-win) for which a traffic matrix is available and also to some other networks with a simple demand model. We will show results comparing this hybridigp/mpls routing scenario with the result of pureigp routing and that of a full meshmpls with and without traffic splitting.     

Bandwidth sharing under the assured forwarding Per-Hop behavior

The DiffServ’s Assured Forwarding (af) Per-Hop Behavior (phb) Group defines a differentiated forwarding of packets in four independent classes, each class having three levels of drop precedence. Specific end-to-end services based on thisphb are still being defined. A particular type of service that could assure a given rate to a traffic aggregate has been outlined elsewhere. In such a service, a fair distribution of bandwidth is one of the main concerns. This paper presents experimental work carried out to evaluate howaf distributes bandwidth among flows under different load conditions and traffic patterns. We focused on the effect that marking mechanisms have on bandwidth sharing among flows within a singleaf class. The traffic types we used includeudp flows, individual and aggregatedtcp flows, mix oftcp andudp, tcp sessions with heterogeneous round-trip times, as well as color-blind and color-aware re-marking at the aggregation point fortcp flows. Tests were performed on real and simulated networks. We have found certain conditions under whichaf distributes bandwidth fairly among nonadaptiveudp flows andtcp aggregates. Finally, we evaluate a basic rule for setting the parameters of the two-rate Three-Color Marker conditioning algorithm (trtcm) in order to achieve a better bandwidth distribution fortcp flows.     

nOBS: an ns2 based simulation tool for performance evaluation of TCP traffic in OBS networks

Performance evaluation of tcp traffic in obs networks has been under intensive study, since tcp constitutes the majority of Internet traffic. As a reliable and publicly available simulator, ns2 has been widely used for studying tcp/ip networks; however ns2 lacks many of the components for simulating optical burst switching networks. In this paper, an ns2 based obs simulation tool (nobs), which is built for studying burst assembly, scheduling and contention resolution algorithms in obs networks is presented. The node and link objects in obs are extended in nobs for developing optical nodes and optical links. The ingress, core and egress node functionalities are combined into a common optical node architecture, which comprises agents responsible for burstification, routing and scheduling. The effects of burstification parameters, e.g., burstification timeout, burst size and number of burstification buffers per egress node, on tcp performance are investigated using nobs for different tcp versions and different network topologies.     

IP-based transport of voice traffic in the UMTS radio access network: analytical study and empirical validation

In this paper, we investigate theIp protocol as a transport option for the user traffic in the UMTS Terrestrial Radio Access Network (Utran), where stringent delay bounds are to be met for both real-time and non real-time traffic. We focus on real-time voice traffic and present an analytical model for the multiplexing and transport of voice channels in theUtran usingIp. The novelty of our model is that it analytically includes and quantifies the performance of the timer used in multiplexing arriving Frame Protocol (Fp) frames into largerIp packets. We then validate our work through empirical results on a test-bed emulating theUtran transport functionalities. We show the trade-offs between performance, in terms of delay and link utilization, and quantify optimal values for the timer as well as the number ofFp frames perIp packet for a given output link capacity.     

An evaluation oftcp with explicit congestion notification

We study the effect of Explicit Congestion Notification (ecn) ontcp for relatively large but finite file transfers inip networks, and compare it to other congestion avoidance mechanisms, namely Drop Tail (dt) and Random Early Detection (red). We use simulation to measuretcp performance for transfers initiated by a varying number of end hosts. In contrast to previous work, we focus on situations in which all nodes in the network operate uniformly under the same mechanism (dt orred orecn). Our results show that under such uniform conditionsecn does not necessarily lead to significant improvement intcp goodput, although in no case does it lead to an actual degradation in performance. Our results also show that, withecn, tcp flows benefit from lower overhead for unsuccessful transmissions. Furthermore, lockouts are largely avoided. In other words, in an all-ecn network resources are shared more fairly. Finally, we show that global synchronization is no longer an issue, and argue that currenttcp versions have essentially solved the problem, regardless of the queue management scheme employed.     

A flow-based approach to modeling ADSL traffic on an IP backbone link

Measurements from an Internet backbone link carryingtcp traffic towards differentadsl areas are analyzed in this paper. For traffic analysis, we adopt a flow-based approach and the popular mice/elephants dichotomy. The originality of the experimental data reported in this paper, when compared with previous measurements from very high speed backbone links, is that commercial traffic comprises a significant part due to peer-to-peer applications. This kind of traffic exhibits some remarkable properties in terms of mice, elephants and bit rates, which are thoroughly described in this paper. Global traffic is actually decomposed into several flow components on the basis of the mice/elephants dichotomy. Mice due p2p protocols and mice due to classical Internet applications such ashttp, ftp, etc. are analyzed separately. It turns out that by adopting a suitable level of aggregation, global traffic can be described by means of usual tele-traffic models based on M/G/∞ queues with Weibullian service times. The blobal bit rate can then be approximated by the superposition of Gaussian processes perturbed by a white noise.     

Securing communications overAtm networks: the remoteAtm private networks interconnection example

When remoteAtm sites communicate through anAtm public network, a number of security problems arise, such as hacking, eavesdropping and traffic tampering. This paper proposes three contributions to these security problems. Firstly, risks due toAtm technology usage are detailed. Secondly, a survey of existing techniques aiming at securingAtm communications is presented with emphasis on theAtm Forum’s security specifications. Thirdly, a new solution called Safe (which stands for Solution for anAtm Frequent communications Environment) developed in the Démostène project is described. Safe realizes both firewall’s filtering functions and communications protection over theAtm network. The main idea of Safe is to use signaling (Uni 3.1) as a means to exchange security information over the network. This idea has been implemented and introduced to theAtm Forum.     

A scheduler for relative delay service differentiation

We propose a new delay-based scheduler called asRD-VC (Relative Delay VirtualClock). Since it performs a delay-based service differentiation among flow aggregates, the quality at microflow level is the same as that at aggregate level. This is not easily achievable when the service differentiation is bandwidth-based or loss-based. Unlike theEDF (Earliest Deadline First) scheduler [1], our proposed scheduler self-regulates and adapts the delays according to load changes. This characteristic permits us to implement it in an AF-likePHB providing the relative quantification service in a DiffServ network. Finally, we compare our proposedrd-vc scheduler with two important existing propositions:WTP (Waiting Time Priority) [2, 3] andex-vc (Extended VirtualClock) [4]. Both these propositions are delay-based and have self-regulation property. All three schedulers (RD-VC, WTP andEX-VC) maintain the required service differentiation among aggregates and have comparable long term average performance like mean throughput per aggregate and packet loss ratio etc. However,RD-VC and WTP take an edge overEX-VC at short-term performance like jitter. Bothrd-vc andWTP have good long term and short-term performance. Our proposedrd-vc, compared to existingWTP, has two additional characteristics, i.e. unlike WTP which is limited to architectures with one queue per Qos class, it has no limitation on implementation scope (with or without separate queues per class) and it has lower complexity. This rendersRD-VC an interesting proposition.     

Optimisation de la quantification vectorielle codée par treillis: application au codage des paramètres LSF

Speech coders operating at low bit rates necessitate efficient encoding of the linear predictive coding (Lpc) coefficients. Line spectral Frequencies (Lsf) parameters are currently one of the most efficient choices of transmission parameters for theLpc coefficients. In this paper, an optimized trellis coded vector quantization (Tcvq) scheme for encoding theLsf parameters is presented. When the selection of a proper distortion measure is the most important issue in the design and operation of the encoder, an appropriate weighted distance measure has been used during theTcvq construction process. We further applied the optimizedTcvq system for encoding theLsf parameters of the us Federal Standard (Fs1016) 4.8 kbps speech coder. At lower bit rates, objective and subjective evaluation results show that the incorporatedLsf tcvq encoder performs better than the 34 bits/frameLsf scalar quantizer used originally in the fs1016 coder. The subjective tests reveal also that the 27 bit/frame scheme produces equivalent perceptual quality to that when theLsf parameters are unquantized.     

Analysis of flow transfer times in IEEE 802.11 wireless LANs

In this paper we present an integrated packet/flow level modelling approach for analysing flow throughputs and transfer times inieee 802.11wlans. It captures the statistical characteristics of the transmission of individual packets at themac layer and takes into account the system dynamics due to the initiation and completion of data flow transfers. In particular, at the flow level the system is modelled by a processor sharing type of queue, reflecting theieee 802.11mac design principle of distributing the transmission capacity fairly among the active flows. The integrated packet/flow level model is analytically tractable and yields a simple approximation for the throughput and flow transfer time. Extensive simulations show that the approximation is very accurate for a wide range of parameter settings. In addition, the simulation study confirms the attractive property following from our approximation that the expected flow transfer delay is insensitive to the flow size distribution (apart from its mean).     

EnhancingIEEE 802.11 performance in congested environments

IEEE 802.11 is the most deployed wireless local area networking standard nowadays. It uses carrier sense multiple access with collision avoldance (CSMA/CA) to, resolve contention between nodes. Contention windows (CW) change dynamically to adapt to the contention level: Upon each collision a node doubles itsCW to, reduce further collision risks. Upon a successful transmission theCW is reset, assuming the contention level dropped. However, contention level is more likely to change slowly, and resetting theCW causes new collisions and retransmissions before reaching the optimal value again. This wastes bandwidth and increases delays. In this paper we propose simple slowCW decrease functions and compare their performances to the legacy standard. We analyze them through simulation and show their considerable enhancement at all congestion levels and transient phases.     

Iterative successive interference cancellation for multi-user DS/CDMA detectors in multipath channels

This paper deals with uplink Direct-Sequence Code Division Multiple Access (DS-CDMA) transmissions over mobile radio channels. A new interference cancellation scheme for multiuser detection, calledSIC/RAKE, is presented. It is based on a modified multistage Successive Interference Cancellation (sic) structure that enables efficient detection in multipath propagation environments, thanks to a single userRAKE receiver incorporated in each unit of thesic structure. Furthermore, a modified version of thesic structure, calledSIC/MMSE, that ensures convergence to theMMSE detector rather than to the decorrelating detector has been suggested. The convergence of theSIC/RAKE andSIC/MMSE methods is proved. Simulation results for the Universal Mobile Telecommunication System (UMTS) have been carried out for flat fading Rayleigh multipath channels, showing that the proposed detector is resistant to the near-far effect and that low performance loss is obtained compared to the single-user bound.     

ATM and AAL layer issues concerning multimedia applications

This paper presents an overview of the facilities that should be provided by theAtm layer and the adaptation layer (Aal) in order to support multimedia communications. From the analysis of data flows generated by well known multimedia applications and the services that could be offered byAtm networks both in the public and the private domain, we propose solutions enabling the availability of such applications onAtm networks taking into account both economical andQos parameters.     

Modélisation de la technique temporelle asynchrone

The evolution towards the integrated broadband communication network is undertaken by several countries. In France, system designers ofCnet in Lannion have assembled an experimental integrated broadband communication network (Ibcn)called Prelude based upon a 4.5 Gbit/s switching matrix and a new and promising transfer mode called the asynchronous time-division technique (Atd).During the project, performance evaluation studies have been carried through, aiming to validate the Atd concepts and to dimension the switching matrix. This paper reviews the Ibcn lab experiments in the world, with operational characteristics;Atd concepts are detailed and the most important features of Ibcn modelling are given. Finally, the network dimensioning is carried through giving packet size, buffers length, loss probabilities and waiting times.     

Un environnement de conception de systèmes distribués basé sur UML

This paper introduces a new environment for developing distributed systems. It is based on theTurtle uml profile. Analysis and design phases, described in previous papers, have been extended with an additional deployment phase. In this new step,Turtle components are deployed over hardware execution nodes, and nodes are connected together throughout links,Turtle deployment diagrams are given a formal semantics inRt-lotos, therefore following the approach used forTurtle analysis and design diagrams. Moreover, the paper presents a Java code generator which outputs appropriate Java code forTurtle deployment diagrams. This code is automatically deployable on networks because it implements node communication using network protocols such asUdp orRmi. ttool, the turtle toolkit has been extended to support these new diagrams and code generators. The attack of protected data exchanged throughout securedHttp sessions serves as example.     

Performance evaluation of DCCP: A focus on smoothness and TCP-friendliness

Recent years have seen dramatic increases of the use of multimedia applications on the Internet, which typically either lack congestion control or use proprietary congestion control mechanisms. This can easily cause congestion collapse or compatibility problems. Datagram Congestion Control Protocol (Dccp) fills the gap betweenUdp andTcp, featuring congestion control rather than reliability for packet-switched rich content delivery with high degree of flexibility. We present aDccp model designed and implemented withOpnet Modeler, and the experiments and evaluation focused on largely the smoothness of the data rates, and the fairness between concurrentDccp flows andTcp flows. We foundDccp-ccid3 demonstrates stable data rates under different scenarios, and the fairness betweenDccp andTcp is only achieved under certain conditions. We also validated that the throughput ofDccp-Ccid3 is proportional to the average packet size, and relatively fixed packet size is critical for the optimal operation ofDccp. Problems in the slow start phase and insufficient receiver buffer size were identified and we hereby proposed solutions on this.     

Parallel interference cancellation efficiency for noisy signals in O-CDMA system

The aim of this paper is to evaluate the robustness of Parallel Interference Cancellation (Pic) to noise contribution for an optical Code Division Multiple Access system. The theoretical expression of thePic error probability is developed in the case of white additive Gaussian noise. From theoretical analysis, we show that, even with noise contribution, thePic receiver outperforms the Conventional Correlation Receiver (Ccr). Moreover, the results highlight that, for a given performance, thePic receiver relaxes not only the constraint on the code length, but also the Signal to Noise Ratio compared toCcr. Particularly, this is proofed in access network context, i.e. 30 users withBer lt; 10^?9.     

Bandwidth sharing with heterogeneous flow sizes

We consider a system with two heterogeneous traffic classes. The users from both classes randomly generate service requests, one class having light-tailed properties, the other one exhibiting heavy-tailed characteristics. The heterogeneity in service requirements reflects the extreme variability in flow sizes observed in the Internet, with a vast majority of small transfers (“mice”) and a limited number of exceptionally large flows (“elephants”). The active traffic flows share the available bandwidth in a Processor-Sharing (ps) fashion. Theps discipline has emerged as a natural paradigm for modeling the flow-level performance of band-width-sharing protocols liketcp. The number of simultaneously active traffic flows is limited by a threshold on the maximum system occupancy. We obtain the exact asymptotics of the transfer delays incurred by the users from the light-tailed class. The results show that the threshold mechanism significantly reduces the detrimental performance impact of the heavy-tailed class.     

Characterization and evaluation of TCP and UDP-based transport on real networks

StandardTcp (RenoTcp) does not perform well on fast long distance networks, due to its AMD congestion control algorithm. In this paper we consider the effectiveness of various alternatives, in particular with respect to their applicability to a production environment. We then characterize and evaluate the achievable throughput, stability and intra-protocol fairness of differentTcp stacks (Scalable,Hstcp,Htcp, FastTcp, Reno,Bictcp, hstcp-lp andLtcp) and aUdp based application level transport protocol (Udtv2) on both production and testbed networks. The characterization is made with respect to both the transient traffic (entry and exit of different streams) and the steady state traffic on production Academic and Research networks, using paths withRtts differing by a factor of 10. We also report on measurements made with 10 Gbit/secNics with and withoutTcp Offload Engines, on 10 Gbit/s dedicated paths set up forSc2004.