Various middleboxes are ubiquitously deployed in networks to perform packet processing functions, such as firewalling, proxy, scheduling, etc., for the flows passing through them. With the explosion of network traffic and the demand for multiple types of network resources, it has never been more challenging on a middlebox to provide Quality-of-Service (QoS) guarantees to grouped flows. Unfortunately, all currently existing fair queueing algorithms fail in supporting hierarchical scheduling, which is necessary to provide QoS guarantee to the grouped flows of multiple service classes. In this paper, we present two new multi-resource fair queueing algorithms to support hierarchical scheduling, collapsed Hierarchical Dominant Resource Fair Queueing (collapsed H-DRFQ) and dove-tailing H-DRFQ. Particularly, collapsed H-DRFQ transforms the hierarchy of grouped flows into a flat structure for flat scheduling while dove-tailing H-DRFQ iteratively performs flat scheduling to sibling nodes on the original hierarchy. Through rigorous theoretical analysis, we find that both algorithms can provide hierarchical share guarantees to individual flows, while the upper bound of packet delay in dove-tailing H-DRFQ is smaller than that of collapsed H-DRFQ. We implement the proposed algorithms on Click modular router and the experimental results verify our analytical results.

The Virtual Network Function (VNF) deployment problem in Network Function Virtualization (NFV) is of broad theoretical and practical interests. As the traffic surges in data centers, VNF deployment requires load balancing across the servers to avoid possible congestions caused by the uneven distribution of VNFs. In addition, VNFs always specify placement constraints, restricting them to run only on part of the servers. Therefore, we study the load balancing problem for VNF deployment with placement constraints. Despite the rich bodies of recent work on the constrained load balancing problem using the network flow algorithms, they all suffer exponential complexities, leading to unbearable running time in practical executions. In this paper, we propose a new load balancing policy termed Constrained Min-max Placement (CMMP) that schedules VNFs in a way similar to the max-min allocation, where we try to assign the most possible VNFs to the poorest loaded server. The online scheduler for CMMP has a logarithmic time complexity and is simple enough to implement in practice. Trace-driven simulations show that the online CMMP speeds up at least two orders of magnitude of running time comparing to other network flow algorithms.

Image aesthetic is a highly subjective task. Thus, generic aesthetics models may lead to inconsistent user agreements even on the same image. Personalized aesthetics models can be employed to remedy the inconsistency issue. In real situation, users shared very small number of annotated images, which makes this problem more challenging. To solve problems above, unlike previous works that focused on user interactive or extracting simple yet effective image features, we address this by meta-learning. Meta-learning is a framework designed for quick adaption of an existing model to a new task with limited labeled data samples. In this way, we can leverage a small amount of annotated data from user and generate an effective personalized aesthetics model quickly. In addition, we proposed a novel meta-learning strategy and a novel meta regularization for our task. Experimental results demonstrate that our approach can effectively learn personalized aesthetics preferences and outperform existing methods on quantitative comparisons with a strong generalization ability.

Summary Studying the influence of negative words that spread during election period is an important work in social media. Most of current methods rely on sentiment analysis of tweets to determine the users' preference. However, sentiment analysis can only makes use of emotional words (ie, adverbs and adjectives), which only take 30 percent of the context in the Internet. According to our empirical analysis based on real datasets, the bias on word selection largely reduced the accuracy of the context in the Internet. In order to address this critical problem, we propose a new method that makes use of nouns with emotional context to determine the election preference of each user. By collecting the frequencies of words in context, we weigh the impact of each supportive/objective noun to strengthen the determination of users' preference. Final results will further be integrated to examine the effectiveness and efficiency of our proposed method. To indicate this idea, we collect and adopt real datasets (UK Prime Minister 2017 and US President Campaign 2016) in the experiments. All the experiment results suggested that our integrated method largely outperformed previous prediction methods. In particular, the prediction results were quite similar to the final results of the UK and US election. Meanwhile, for UK election, we found that the daily approval rate is closely related to the event happened everyday.

As Network Function Virtualization (NFV) becomes reality and cloud computing offers a scalable pay-as-you-go charging model, more network operators would like to outsource their Service Function Chains (SFC) to the public clouds in order to reduce the operational cost. However, how to minimize the operational cost with Quality of Service (QoS) guarantee when outsourcing SFC is still an open problem. In this paper, we are to study this problem when there are large number of candidate cloud providers with diverse pricing schemes of network functions. In addition, extra delay is introduced as the result of outsourcing SFCs. Firstly, we formulate this problem as an Integer Linear Programming (ILP) model. Then we design an efficient heuristic algorithm named QoS-Guaranteed SFC Outsourcing algorithm (QGSO) based on Hidden Markov Model (HMM). The extensive simulations show that QGSO saves up to 75.8% cost compared with that of deploying network functions in local network. QGSO also achieves up to 42.6% cost savings compared with the result of first-fit based optimization algorithm.

This paper presents D2-CAS, a novel decentralized coflow scheduling system, to minimize the rate of deadline missed coflow for datacenter networks. To design D2-CAS, we first formulate the deadline-missed coflow minimization problem and show its equivalence with the well-known problem of minimizing the late jobs in a concurrent open shop, which is NP-hard in ordinary sense. Inspired by Moore-Hodgso's algorithm (MHA), the optimal solution for minimizing late jobs on a single machine, we design an efficient coflow schedule algorithm, CS-MHA, and further propose its decentralized implementation, D2-CAS. Basically, each sender in D2-CAS periodically runs a part of CS-MHA to get a local suggestion for flow priority assignment, and then multiple senders of a coflow negotiate for an orchestrated priority by leveraging their common data receivers. Via delivering each coflow's packets with its negotiated priority, senders finally carry out efficient deadline-aware coflow scheduling in a decentralized fashion. To the best of our knowledge, this is the first paper that theoretically investigates the deadline-aware coflow scheduling problem, while D2-CAS is the first decentralized solution. Real parameter driven simulations imply that, with the simple yet efficient mechanism, D2-CAS greatly outperforms all existing solutions on reducing the deadline-missed coflows (e.g., outperforms Varys more than 2x).

In current data centers, an application (e.g., MapReduce, Dryad, search platform, etc.) usually generates a group of parallel flows to complete a job. These flows compose a coflow and only completing them all is meaningful to the application. Accordingly, minimizing the average Coflow Completion Time (CCT) becomes a critical objective of flow scheduling. However, achieving this goal in today's Data Center Networks (DCNs) is quite challenging, not only because the schedule problem is theoretically NP-hard, but also because it is tough to perform practical flow scheduling in large-scale DCNs. In this paper, we find that minimizing the average CCT of a set of coflows is equivalent to the well-known problem of minimizing the sum of completion times in a concurrent open shop. As there are abundant existing solutions for concurrent open shop, we open up a variety of techniques for coflow scheduling. Inspired by the best known result, we derive a 2-approximation algorithm for coflow scheduling, and further develop a decentralized coflow scheduling system, D-CAS, which avoids the system problems associated with current centralized proposals while addressing the performance challenges of decentralized suggestions. Trace-driven simulations indicate that D-CAS achieves a performance close to Varys, the state-of-the-art centralized method, and outperforms Baraat, the only existing decentralized method, significantly.

In current data centers, an application (e.g. MapReduce) usually generates a collection of parallel flows sharing a common goal. These flows compose a coflow and only completing them all is meaningful. Accordingly, minimizing the average coflow completion time (CCT) becomes a critical objective for flow scheduling. In this topic, the state-of-the-art centralized method, Varys, achieves a good average CCT; but it has the scalability problem. Alternatively, the only existing decentralized method, Baraat, suffers from the head-of-line blocking problem. To solve these problems, we propose D-CAS, a preemptive, decentralized, coflow-aware scheduling system in this paper. D-CAS pursues coflow-level remaining-time-first (MRTF) principle by leveraging a simple negotiation mechanism between each coflow's data senders and receivers. As the MRTF principle is inherently preemptive and proven to be a near-optimal guideline to minimize average CCT, D-CAS avoids the head-of-line blocking problem and gets good performances. Through extensive simulations, we find that D-CAS achieves a performance close to Varys (gap <; 15%) and outperforms Baraat significantly (about 1.4-4×).

Server consolidation schemes whereby each server is replaced with a virtual machine (VM) and multiple such VMs are run on a single physical server can reduce the number of physical servers needed, and in turn, both the cost and energy consumption in datacenters. However, existing schemes have not fully exploited the flexibility in the usage and allocation of virtualization resources, so as to allow one application originally deployed on a single large VM (LVM) to be split and hosted by multiple smaller VMs (SVM). Using multiple SVM instead of a LVM enables resource allocation at a smaller granularity and hence may further increase the utilization and reduce the number of physical servers. However, a major challenge to be overcome when deploying multiple SVMs for one application is to preserve the performance of the application in terms of e.g., response delay. In this paper, we show through experiment based data analysis that in order to preserve the performance of the application, one needs to allocate sufficient resources to each SVM, and the total amount of resources required by all the SVMs will exceed that required by the LVM. Nevertheless, we also show that by using the proposed heuristic algorithm called VM Splitting and Assignment (VMSA), we can substantially improve the utilization and reduce the number of physical servers.

In OpenFlow-based SDN, flow tables are TCAM-hungry and commodity switches suffer from limited concrete flow table size. One method for coping with the limitations is to use aggregation schemes to reduce the number of flow entries required to represent the same forwarding semantics. Unfortunately, the aggregation retards table updates and lengthens the updating time. During which, the data plane is inconsistent with the control plane, forwarding errors such as Reachability Failures, Forwarding Loops, Traffic Isolation and Leakage are prone to occur. Since network updates take place frequently in practice, the aggregation scheme must be efficient enough. In this paper we propose offline FFTA (Fast Flow Table Aggregation) and its online improver iFFTA to shrink the flow table size and to provide practical fast updates. iFFTA is the first online non-prefix aggregation scheme. Extensive experiments demonstrate: (1) FFTA is about 200× faster than the previously published best non-prefix aggregation scheme without loss of compression ratio on offline aggregation; and (2) iFFTA achieves about 3× faster than FFTA on online update incorporations with a loss of an acceptable compression ratio per update. Thus the user could make a combination use of FFTA and iFFTA for table aggregations: call iFFTA usually and recall the efficient FFTA once the switch is running out of concrete flow table space.

Network virtualization facilitates sharing and efficient utilization of computing and bandwidth resources of an underlying substrate network. As network virtualization becomes popular, it is important to efficiently map a virtual infrastructure (VI) onto a substrate network, such that the survivability of the former can be guaranteed against failures in the latter. In this paper, we study a virtualization layer approach to survivability, whereby the virtualization layer customizes a VI request with redundant nodes and links according to its reliability requirements and then passes limited information about the augmented VI to the physical layer, where the mapping of the augmented VI takes place. More specifically, we develop a flexible scheme to enhance the original VI graph with K redundant nodes, in order to fight against an arbitrary substrate node failure. In addition, a scenario-based component group (SBCG) concept is proposed to describe resource sharing of enhanced VI requests at the physical layer. We also develop an efficient heuristic that takes advantage of the limited information on SBCG to reduce costs when mapping the enhanced VI to the substrate network. The efficiency of the proposed solution is compared using extensive simulation under various performance metrics. It is shown that the K-redundant-node scheme with SBCG information is more cost efficient than the existing 1-redundant-node solution.

Cloud based datacenters provide on demand services and resources, both transparently and cost effectively. These services and applications are typically hosted and run on the servers located in interconnected datacenters. The task or application request from users can be abstracted as a virtual network (VN) request. How to efficiently accommodate VN requests by mapping them onto the substrate network is a challenging problem. Current research only considers static VN requests, where the VN request and the demand for resources is fixed and does not change over time. However, most of the application requests submitted to datacenters present dynamic changing characteristics. In this paper, we address the issue of how to optimally reconfigure and map an existing VN while this VN request changes. As the VN provisioning problem is NP-hard, we propose heuristic algorithms for solving it efficiently. We evaluate the performance and effectiveness our algorithms by conducting simulations on realistic network. The simulation results show that our approach solves this problem efficiently and performs well than existing approaches.

In the cloud computing paradigm users access applications/services hosted and run on virtual machines in interconnected datacenters. Applications from the same user may need to interact and change data or information, thus, we may abstract the applications/services request from same user as a virtual network (VN). To improve performance and resource efficiency, it is critical that the VN request be optimally provisioned given the current resource state of the datacenters. However, all of the existing research about optimal VN provisioning conducted for deterministic services. In this work, we formulate the problem of optimal provisioning for elastic service oriented VN request as a mixed integer programming (MIP) model with the objective of maximizing the total revenue of infrastructure provider (InP), and propose a genetic algorithm based heuristic (GAH) algorithm for addressing the optimal provisioning for virtual network request with un-splittable flow (OPVNUF) and optimal provisioning for virtual network request with splittable flow (OPVNSF) problems of elastic services. We demonstrate the effectiveness of our approach for improving the total revenue, by conducting extensive simulations on real substrate network.

Intra-domain traffic engineering can significantly enhance the performance of large IP backbone networks. An important component of current methods for traffic engineering with link state routing protocols like OSPF is accurate knowledge of traffic matrix. However, the traffic matrix is unknown and varies with time. So it is important to obtain a traffic engineering method that is "robust" to variations in traffic matrix. In this paper, we use multi-topology routing (MTR) for providing robust traffic engineering in IP networks. We first formulate the problem of robust traffic engineering using MTR as a mixed integer programming problem (MIP). To make the problem solvable, we then decompose the problem into logical topologies design problem and traffic assignment problem, which are solved by a heuristic algorithm and a tractable linear programming (LP) model, respectively. Using simulation results on Rocketfuel topologies, we study and discuss effectiveness of the proposed robust traffic engineering approach. Simulation results show that robust traffic engineering based on MTR is promising.

We address the issue of traffic-oblivious routing (i.e., robust routing) in WDM networks dealing with both link and node failures under load-balancing architectures. Two distinct schemes are proposed. One is static with the goal of minimizing total network cost given a set of multiple failures. The other considers dynamic network environment (i.e., connection requests arrive one after another), and is designed against the failures aiming at the maximum network throughput. We manifest by simulation that both of these two schemes perform better than the previously proposed protection schemes as well as the unprotected ones.

On the uplink of a DS-CDMA system, successive interference cancellation (SIC) technique can be employed to reduce multiple access interference and improve system capacity. In such a system with K active users, there are K possible decoding orders of SIC and not every decoding order is feasible due to some constraints. It is highly time-consuming to examine the system feasibility directly by using the exhaustive search method (ESM) for a system with even a moderate number of users. In this paper, we propose an efficient approach for examining the feasibility of DS-CDMA systems with imperfect SIC. The proposed approach has significantly less computational complexity than that of ESM, thus benefits the quick decision of admission control and/or scheduling in DS-CDMA systems. Furthermore, under the decoding order obtained by the proposed approach, we prove that the system is able to achieve the lowest outage probability among all possible decoding orders. Simulation experiments and numerical results validate our analysis and demonstrate the effectiveness of our approach.

As multicast applications become popular, provisioning survivable multicast connections in WDM networks is an important issue. In this paper, we study the problem of multicast protection in sparse-splitting WDM networks, and propose an efficient protection algorithm called sparse-splitting constrained multicast protection (SSMP) algorithm. Differing from previous works, the backup paths derived by SSMP can share wavelengths with primary tree in sparse-splitting WDM networks. To achieve wavelength sharing between primary tree and backup paths, a layered graph model is developed. Simulation results show that SSMP can achieve better performance in terms of average network cost and blocking probability than existing algorithms.

It has been widely recognized in the literature that it is imperative to protect light-tree based multicast sessions against single link failures since a single fiber failure can disrupt the information dissemination to several destination nodes. In this paper, we address the problem of routing survivable multicast sessions (RSMS) in wavelength-division multiplexing (WDM) mesh networks. We consider the dynamic network environment, where multicast sessions arrive dynamically one after another. Our objective is to be able to consume as least cost (e.g., wavelengths) as possible for one-at-a-time arrivals and no priori knowledge of future arrivals. We present an approximation algorithm that outperforms previously proposed algorithms for the RSMS problem such as optimal path-pair-based shared disjoint paths (OPPSDP). To the best of our knowledge, this is the first approximation algorithm proposed for the RSMS problem. We prove that the cost of the solution obtained by the HCBA algorithm is at most 4 times that of the optimal solution. We show by simulation that our algorithm performs very close to the optimal solution obtained by solving a mathematical formulation for the RSMS problem. We also show that, compared with the OPPSDP algorithm, our algorithm can significantly reduce the average cost and the blocking probability.

In this work, we investigate the problem of protecting dynamic multicast sessions in mesh WDM networks against single link failures. We propose two efficient multicast session protecting algorithms, called optimal path pair based removing residual links (OPPRRL) and source leaf path based avoiding residual links (SLPARL), which try to reduce the usage of network resource by removing or avoiding residual links in the topology consists of "light-tree " and its backup paths. We compare the proposed algorithms with existing algorithms through simulation. Simulation results indicate that the two proposed algorithms have better performance than other existing algorithms in terms of wavelength links required and network blocking probability.

In this paper, we focus on the problem with the objective to maximize the system performance, while guaranteeing multiple QoS (quality of service) constraints for wireless data networks accommodating multiclass services with different quality requirements. First, we formulate and solve the opportunistic scheduling problem with multiple general long-term QoS constraints. Then, we generalize this problem to include short-term QoS constraints for real-time multimedia users and long-term QoS constraints for non-real-time data users simultaneously in multiclass services networks. Simulation results illustrate that the proposed scheduling schemes guarantee the different QoS constraints, and achieve high system performance.

In this paper, we address the problem of resource allocation with efficiency and quality of service (QoS) support in uplink for a wireless CDMA network supporting real-time (RT) and nonreal-time (NRT) communication services. For RT and NRT users, there are different QoS requirements. We introduce and describe a new scheme, namely the traffic aided uplink opportunistic scheduling (TAUOS). While guaranteeing the different QoS requirements, TAUOS exploits the channel condition to improve the system throughput. In TAUOS, the cross-layer information, file size information, is used to improve the fairness of NRT users. Extensive simulation results show that our scheme can achieve high system throughput in uplink wireless CDMA system, while guaranteeing the QoS requirements.

The advances in wavelength-division multiplexing (WDM) technology are expected to facilitate bandwidth-intensive multicast applications. A single fiber failure in such a network, however, can disrupt the information dissemination to several destination nodes in a "Iight-tree"-based multicast session. Thus it is imperative to protect the multicast sessions. In this paper, we propose a novel protection scheme, called multicast protection through spanning paths (MPSP), for resource efficient multicast protection with spare capacity sharing. Here, a spanning path is a path from a leaf node to any other leaf node of a multicast tree. The key idea of MPSP is first to identify a backup path for each spanning path and then to appropriately select parts of these backup paths to protect the primary multicast tree, so that the total bandwidth allocated to the primary multicast tree and its protection paths (or trees, etc.) is minimized. While previous studies only consider self-sharing and intra-request sharing, to the best of our knowledge, this is the first time to take inter-request sharing of spare capacity into consideration when protecting dynamic multicast sessions. We use simulations to demonstrate the performance of the MPSP scheme. It is shown that significant performance improvements are achieved in terms of average cost per multicast session and blocking probability. Compared with existing schemes, the average cost is reduced by about 22% and the blocking probability can be reduced by about 27% in average.

Segment protection has been recognized as an efficient way to avoid ldquotrapsrdquo in survivable WDM mesh network. In this paper, we present some insights on this kind of protection. We first prove the correctness of segment protection used for avoiding traps. We then investigate the effect of the number of segments on the blocking performance in several typical real networks (includes middle and large networks). We find that, in most cases, it is sufficient to use two backup segments to protect a given active path. Based on this observation, we then propose a novel and efficient segment protection algorithm, called two-segment (TS), which has a lower computational complexity and comparative performance than existing algorithms.

The advances in wavelength-division multiplexing (WDM) technology are expected to facilitate bandwidth-intensive multicast applications. A single fiber cut on such a network, however, can disrupt the transmission of information to several destination nodes on a "light-tree"- based multicast session. Thus it is imperative to protect multicast sessions. In this article we propose a novel protection scheme called multicast protection through spanning paths (MPSP) for efficient multicast protection. Here, a spanning path is a path from a leaf node to any other leaf node of a multicast tree. The key idea of MPSP is to derive a backup path for each spanning path and then appropriately select part of these backup paths to protect the primary multicast tree such that the total bandwidth allocated to the primary multicast session and its protection paths (or trees, etc.) for all the multicast sessions is minimized. Simulation results are used to demonstrate the good performance of the proposed protection scheme in reducing the spare capacity for protection. Compared with existing schemes, the decrease in average cost can be quite high (e.g., decrease about 20%) and the blocking probability can be reduced approximately 15%.

Connection reliability is one important service level agreement (SLA) parameter for a customer and should be carefully considered in survivable WDM networks. A sound scheme should guarantee customers' reliability and simultaneously benefit a service provider in resource efficiency. Under the SLA constraints and the assumption of shared risk link group (SRLG) failures, a novel dynamic differentiated shared path-protection algorithm (DDSP) in WDM mesh networks is proposed. Based on the basic ideas of the K-shortest path algorithm and partial SRLGdisjoint protection, DDSP can provide differentiated services for customers according to their SLA parameters while optimizing resource utilization. Simulation results show that DDSP not only can efficiently guarantee the specific SLA requirements of customers, but also can achieve significant performance gain and lead to remarkable reduction in blocking probability.

A k-ary n-cube direct interconnection structure is originally introduced and developed in the context of massively parallel computer (MPC) architectures. In recent years, with the development of VLSI (very large scale integration) technology and requirements of high performance Internet routers/switches, the k-ary n-cube direct interconnection network is considered to build Terabit routers. In this paper, the analysis model for three-dimensional torus as switching fabric of Internet routers is presented. Delay/throughput characteristics are measured in terms of four different routing algorithms, buffer length and the size of fabrics. Also a new algorithm ladder is proposed. Simulations have been done on OPNET. Numeric results show that this architecture can be used to build high performance routers. Comparing with other routing algorithms, the ladder has better performance.

In this paper, we propose a fair scheduling algorithm combined with reservation resource compensation for wireless networks. The proposed algorithm is capable of supporting differentiate application-level services and uses different queuing strategies for each class of services. We also introduce a compensation mechanism and a resource reallocation method to achieve fairness. The compensation that a lagging flow receives is proportional to the original transmission speed of the flow.

A new heuristic algorithm, called virtual topology mapping for design protection based on layered graph (LG VTMDP), used in design protection for WDM optical networks is proposed in this paper. The algorithm considers the problem of routing and wavelength assignment simultaneously as well as design protection. Load balancing and capacity constraints of physical links are also considered. The LG VTMDP algorithm is shown to perform better than the combination of the disjoint alternate path (DAP) algorithm and existing wavelength assignment algorithms. Then an efficient wavelength converter placement (WCP) algorithm is presented. By only placing a small number of wavelength converters at some "key" nodes, this sparse conversion scheme can obtain sufficiently high performance. Finally, the performance of our algorithms is studied using network examples.

This paper studies multihop optical networks in which nodes employ wavelength routing switches that enable the establishment of wavelength division multiplexed (WDM) channels, called lightpaths, between node pairs. In fact, most optical networks are multifiber networks. The problem of dynamical routing and assignment of wavelength (RAW) in such networks is studied in this paper. Two resource assignment strategies, PACK and SPREAD, are proposed. By virtue of a layered-graph, routing and assignment of wavelength subproblems can be considered simultaneously. These two strategies can be used to solve the RAW problem in networks with even links as well as that in networks with uneven links. Simulation shows that layered-graph-based RAW algorithms perform better than the existing ones. It also shows that SPREAD with distributive use of network resources can achieve better performance than PACK with collective use of resources in multifiber networks. The layered-graph-based algorithms can effectively deal with the failure of the fiber/link and node. By making use of the special structure of the layered-graph, we propose a shortest path algorithm, whose complexity is lower than that of the standard shortest path algorithm.

Two algorithms are proposed for the dynamic routing and assignment of wavelength problem in multi-fiber wavelength division multiplexing all-optical networks. By virtue of the layered graph, the routing and assignment of wavelength subproblems can be considered simultaneously. Simulation shows that layered-graph-based RAW algorithms perform better than the existing ones. Making use of the special structure of the layered graph, we propose a shortest path algorithm, whose complexity is lower than that of the standard shortest path algorithms.

In order to support multimedia in the wireless networks, the ATM technology will be adopted. CBR traffic should also be supported, and then the clock recovery for CBR traffic is one of the key techniques in wireless ATM networks. Due to the absence of a common network clock, higher cell delay jitter and higher cell loss in wireless network, the new clock recovery technique should be used to deal with the above cases. We propose a new clock recovery technique named asynchronous residual time stamp (ARTS). The ARTS is similar to SRTS used in B-ISDN/ATM networks, but the reference clock in ARTS is independent of the source and the destination. In addition, the pulse positive/negative stuffing technique is used to improve the performance of ARTS, and an ARTS first-in-first-out (FIFO) buffer is implemented to absorb the ARTS delay jitter at the destination. On the other hand, the difference regularity of the previously received ARTSs is used to estimate and then compensate the loss of the ARTSs. Analysis and simulation results indicate that the proposed clock recovery technique for CBR traffic is adequate for wireless ATM networks, and is compatible with that of B-ISDN/ATM networks.

In this paper, a novel multiple access scheme named DSAMA protocol is proposed for wireless ATM networks. DSAMA is the combination of reservation multiple access protocol and dynamic slot allocation algorithm. The orthogonal code is used for reservation in the DSAMA system. Hence, the packet collisions are prevented while several packets apply for reservation slot simultaneously. In addition, due to bandwidth limitation, the dynamic slot allocation algorithm with bound variable for bursty traffic is also proposed in this paper, so that the QoS of multimedia traffic is guaranteed in wireless ATM networks. A computer simulation is utilized to validate the DSAMA protocol. From the analysis and simulation results, we know that the slot utilization in the DSAMA system is higher than that of the other well-known integrated services multiple access (e.g., PRMA, PRMA++). Furthermore, the performance of bursty traffic in the DSAMA system outperforms the case without dynamic slot allocation. We are sure that DSAMA protocol is a competitive candidate to be adopted in the wireless ATM networks.

The maximum throughput of an N/spl times/N nonblocking packet switch with input queues and two priority classes is analyzed. Packets are of fixed length and the switch operation is slotted. Packets of both priority classes are queued when waiting for service. High-priority packets preempt low-priority ones and move ahead of all low-priority packets waiting in the queue. A new method of analysis is employed. The calculated results of the maximum throughput obtained are close to the simulation results.>