Performance analysis of layer3 switch: case of flow-driven connection setup

This paper presents a traffic model for an ATM multiplexer providing video, voice, image, and data services. The traffic model classifies the input traffic into two types: real-time and non-real-time. The input process for real-time traffic is a Markov modulated Bernoulli process (MMBP), while that for non-real-time traffic is a Bernoulli process with batches. The z-transform of the cell interarrival time A/sub MMBP/(z) is obtained. The average traffic load /spl rho//sub MMBP/ and the squared coefficient of variation of the time between success cell C/sup 2//sub MMBP/ are then obtained from the respective derivatives of A/sub MMBP/(z). The autocorrelation coefficient of the cell interarrival time with lag 1 /spl psi/, can be also obtained. This paper presents an iterative algorithm that can be used to calculate the MMBP probability distribution of the cell interarrival time. So far few results about discrete-time finite buffer queues with correlated inputs are known. This paper analyzes the performance of an ATM multiplexer considering the correlation and burst in cell arrivals. For the MMBP-stream, the waiting time distribution and the cell loss probability for MMBP+M/sup [X]//D/1/K finite buffer system are all exactly calculated with the use of a recursive algorithm. Calculations show the influence of burstiness and correlation on both the loss probability and waiting time. The accuracy of the analytical results has been largely validated by means of a simulation approach.

How the traffic characteristics and the performance of a transport connection are affected at the link-transport layer is one of the important items to be studied to determine the end to end performance of a transport protocol. We model the link-transport layer that provides service to an XTP connection as a discrete time single server finite capacity queueing system at which two different arrival processes are allowed to merge, e.g., a Markov Modulated Bernoulli Process (MMBP) and a Bernoulli Process (BP). The MMBP models the bursty and correlated traffic from the designated XTP connection, and the BP models the external traffic, that is, the arrival of the packets from the rest of the coexisting connections at the link-transport layer. An exact analysis to obtain the queue length distributions, and the probability densities and the correlation coefficients of the interdeparture time distributions for the packets from the designated XTP connection is presented. Numerical examples are also investigated to show how the traffic characteristics of the designated XTP connection are affected by the amount of the external traffic, the degree of the burstiness and the correlation of the traffic source.KeywordsQueue LengthArrival ProcessInterarrival TimePacket ArrivalWait Time DistributionThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Network devices are fundamental items in computer network design. The selection of the best device must be in terms of price to performance ratio since there is a budget limit to build a computer network. In this paper, we compare the packet loss rate (PLR), throughput and jitter of switches from different vendors and with different prices. A standard methodology is implemented to evaluate these parameters and final comparisons are made. Packet size and transmission rates followed the recommendations of RFC2544 for full duplex Ethernet 100BaseTX. The difference in cost between layer-2 switches must be related to administrative features, number/type of ports, supported layer-2 protocols and head-of-line blocking. It is observed that the best layer-2 switch is not necessarily the most expensive one and depending on the administrative features needed for each network, less expensive model can be used and obtain the same result.

This paper presents analyses to study the persession performance of an ATM multiplexer which finds application in call admission control (CAC). We consider a scenario where a switching node is non-blocking and output buffered, links are of different speeds and the traffic is bursty with correlated arrivals. We model the session of interest as a Markov modulated Bernoulli process (MMBP) and the interfering traffic as an MMBP with batch arrivals. We propose an efficient method to obtain the exact steady-state performance results of a session. An approximation technique is also presented. This method could also be used to analyze any Markov renewal traffic model such as discrete-batch Markovian arrival process (D-BMAP), and study the transient behavior of a session.

In ATM networks, the time axis is divided into equal intervals called time slots, so most references assume that the Markov chain makes a transition from state i to state j in the boundary of time slots in view of Markov-modulated Bernoulli process (MMBP). This paper considers a more general situation, i.e., the Markov chain makes a transition from state i to state j in the boundary of frames whose length are an integral multiple of length of one time slot. Under this MMBP model, the paper develops a four-dimensional discrete-time Markov chain, and obtains results of the performance analysis of leaky bucket algorithms.

In the ATM network, the leaky bucket scheme is a promising method that regulates input traffics for preventive congestion control. In order to satisfy the loss probability constraint of more bursty input traffic, it is known that the leaky bucket scheme with static token generation interval requires larger data buffer and token pool size. This causes an increase in the mean waiting time for input traffic to pass the leaky bucket scheme, which would be inappropriate for real time traffics such as voice and video. The authors present the leaky bucket scheme with threshold-based token generation intervals in which the token generation interval changes according to buffer occupancy. The proposed leaky bucket scheme is analysed in the discrete-time case by assuming the arrival process to be a Markov-modulated Bernoulli process (MMBP). The authors show that the quality of service (QoS) of input traffics is satisfied with smaller data buffer and token pool size in the proposed leaky bucket scheme.

In this paper, we examine the queuing performance of terminals in a multichannel centralized S-ALOHA system with a finite terminal population and finite queue size in each terminal, employing a uniform backoff (UB) algorithm with retry limit for collision resolution. The performance evaluations focus on uplink traffic from web browsing, which is modeled as a Markov-modulated Bernoulli process with correlated arrivals. We analyze the system performance in terms of system throughput, mean queue length, mean delay, the probability that a packet is dropped by retry limit, the probability that a packet is blocked by a full queue, and the mean and variance of packet retransmission time, in relation to the source correlation, number of channels, window size, and retry limit. For comparison, we also consider bufferless terminals with correlated arrivals. In addition, we evaluate by simulations the performance of terminals with finite queue size when they have perfect knowledge of the backlog size. Results from our study allow the parameters of the UB algorithm to properly be chosen to meet the access-level quality-of-service requirements.

Optical burst switching (OBS) technology is a promising solution for the next generation Internet backbone. However, contention resolution and QoS support constitute critical issues in its services development. We propose in this paper a novel architecture for optical networks suitable for OBS switching, contention resolution, and QoS provisioning. The performance evaluation is realized based on an ad-hoc analytic model and a simulation. The method proposed is based on the re-design of optical switches to include optical delays lines and adapting the signaling protocol.

Continuous and explosive growth of the Internet has shown that current TCP mechanisms can obstruct efficient use of high-speed, long-delay networks. To address this problem we propose an enhanced transport-layer protocol called gHSTCP, based on Highspeed TCP proposed by Sally Floyd. It uses two modes in the congestion avoidance phase based on the changing trend of RTT. Simulation results show gHSTCP can significantly improve performance in mixed environments, in terms of throughput and fairness against the traditional TCP Reno flows. However, the performance improvement is limited due to the nature of TailDrop router, and the RED/ARED routers can not alleviate the problem completely. Therefore, we present a modified version of Adaptive RED, called gARED, directed at the problem of simultaneous packet drops by multiple flows in high speed networks. gARED can eliminate weaknesses found in Adaptive RED by monitoring the trend in variation of the average queue length of the router buffer. Our approach, combining gARED and gHSTCP, is quite effective and fair to competing traffic than Adaptive RED with Highspeed TCP.

We consider simple models of congestion control in high-speed networks, and develop diffusion approximations which could be useful for resource allocation. We first show that, if the sources are ON-OFF type with exponential ON and OFF times, then, under a certain scaling, the steady-state distribution of the number of active sources can be described by a combination of two appropriately truncated and renormalized normal distributions. For the case where the source arrival process is Poisson and the service times are exponential, the steady-state distribution consists of appropriately normalized and truncated Gaussian and exponential distributions. We then consider the case where the arrival process is a general renewal process with finite coefficient of variation and service-time distributions that are phase type, and show the impact of these distributions on the steady-state distribution of the number of sources in the system. We also establish an insensitivity to service-time distribution when the arrival process is Poisson. We use these results to relate the capacity of a bottleneck node to performance measures of interest for best effort traffic.

A protocol for setting up a tree connection for the purpose of multicast communication over a high-speed packet switched network is presented. The tree connection is based on the use of local identifiers, which are swapped at every intermediate node of the tree. Local identifiers are simple to manage, provide fast access to the routing tables, and are very efficient in terms of the size of the resulting routing space. The authors consider a high-speed network in which the local label swapping on data messages is performed in hardware by the switching subsystem, while the connection set-up is done in software by the switch control subsystem. A reliable protocol for setting up the tree connection is presented, and its correctness is formally proved. The protocol ensures the integrity of the tree and when failures occur, it is gracefully degraded into a smaller tree. >

Network traffic measurement provides scientific basis for designing, building, and managing the next generation Internet, and is especially important for monitoring network behavior. There are many challenges in high-speed network traffic measurement. One of the most important bottlenecks is traffic capturing. Because of hardware capability and operating system overhead limitations, the existing network traffic measurement tools based on software can only perform well at low speed network with the link rate below 100 Mbps. This paper mainly focuses on the key issue to perform traffic capture for high-speed network based on software. To achieve this point, Distributed network traffic measurement system is discussed, and a traffic capture prober was designed under general PC server with Linux operating system. We also analyze crucial problems on prober’s implementation, building an experiment environment and running tests on it. The results show that our system satisfies all requirements imposed by real time measuring network traffic behavior. Furthermore, the performance of traffic capture has been improved significantly and can capture/process nearly all packets at line speed under Gigabit network.

Abstract. Hydrological observatories bear a lot of resemblance to the more traditional research catchment concept, but tend to differ in providing more long-term facilities that transcend the lifetime of individual projects, are more strongly geared towards performing interdisciplinary research, and are often designed as networks to assist in performing collaborative science. This paper illustrates how the experimental and monitoring set-up of an observatory, the 66 ha Hydrological Open Air Laboratory (HOAL) in Petzenkirchen, Lower Austria, has been established in a way that allows meaningful hypothesis testing. The overarching science questions guided site selection, identification of dissertation topics and the base monitoring. The specific hypotheses guided the dedicated monitoring and sampling, individual experiments, and repeated experiments with controlled boundary conditions. The purpose of the HOAL is to advance the understanding of water-related flow and transport processes involving sediments, nutrients and microbes in small catchments. The HOAL catchment is ideally suited for this purpose, because it features a range of different runoff generation processes (surface runoff, springs, tile drains, wetlands), the nutrient inputs are known, and it is convenient from a logistic point of view as all instruments can be connected to the power grid and a high-speed glassfibre local area network (LAN). The multitude of runoff generation mechanisms in the catchment provides a genuine laboratory where hypotheses of flow and transport can be tested, either by controlled experiments or by contrasting sub-regions of different characteristics. This diversity also ensures that the HOAL is representative of a range of catchments around the world, and the specific process findings from the HOAL are applicable to a variety of agricultural catchment settings. The HOAL is operated jointly by the Vienna University of Technology and the Federal Agency for Water Management and takes advantage of the Vienna Doctoral Programme on Water Resource Systems funded by the Austrian Science Funds. The paper presents the science strategy of the set-up of the observatory, discusses the implementation of the HOAL, gives examples of the hypothesis testing and summarises the lessons learned. The paper concludes with an outlook on future developments.

This paper proposes a new lightweight QoS/CoS control method called VWRR (variably weighted round robin) for use on high-speed backbone networks. We evaluate the efficiency of VWRR by analysis and simulation. Queueing methods that can accurately guarantee the bandwidth like WFQ (weighted fair queueing) usually put high processor loads; this makes them a poor choice for use on high-speed lines. In VWRR, the weight of the WRR (weighted round robin) router varies adaptively depending on the average packet length; therefore, the router can adaptively control bandwidth allocation without drastically increasing the processing load. VWRR can trade the attained fairness for the processing load imposed. In an ideal case where no control delay exists, fairness of VWRR varies from that of WRR to that of DRR (deficit round robin). On the other hand in art actual case where control delays exist, the fairness of VWRR becomes better than WRR if the measurement interval is appropriately determined.

A practical bytecode interpreter for programmable routers on IXP network processors