This paper presents a novel model-free resource allocation framework for the downlink of 5G cellular networks to guarantee stringent QoS requirements in wireless applications. A Deep deterministic policy gradient (DDPG) agent with a modified Genetic Algorithm (GA) based resource allocation framework is proposed to balance the tradeoffs between reliability, latency, and data rate. Any feasible point in the rate-latency-reliability domain can be achieved with this approach. Compared to state-of-the-art approaches DDPG-Dual and DDPG-PSO, the proposed model achieves higher reliability and scalability in joint optimization with QoS constraints. Specifically, the proposed model guarantees the expected reliability with 25 % and 42.86 % improvement respectively over the compared models. In terms of conventional effective bandwidth approach, the proposed model achieves 30.82 % improvement of energy efficiency under the same QoS constraints. Moreover, the proposed model offers a practical solution, namely, three times faster convergence and only 6.7% of the scheduling time compared to the ground truth Dual decomposition optimization.

Spatio-temporal aware IoT applications need to deliver the Spatio-Temporal Data (STD), which depends on geographic location and time, to the users in real-time. We have proposed the STD retention system using crowds of vehicles to promote the immediate distribution and utilization of STDs in geographical proximity to the user. However, when the system retains a Large-size STD (LSTD) consisting of multiple packets such as videos and photos, the packet loss due to channel collisions should be suppressed for effective retention of an LSTD. In this paper, we propose a complementary data transmission control method with collision avoidance for LSTD retention and clarify the effectiveness of our proposed method by simulation.

IoT data includes Spatio-temporal data (STD) that is needed only at a specific time and place. In our previous research, we have proposed the STD data retention system (STD-RS), aiming to construct a novel architecture for STD distributing in a retention area by vehicles. However, since vehicles may distribute the STD in unexpected locations due to GPS errors or malicious behavior in a real environment. Therefore, we propose a reliability audit mechanism based on multi-layered blockchain managing and analyzing the history of STD distribution and vehicles' behavior in each retention area. Through simulation experiment, we demonstrated that our mechanism detects the malfunction of STD-RS effectively.

With the development and spread of IoT technology, the number of devices connected to the Internet is increasing. Some data generated by IoT devices include spatio-temporal data (STD) that depends on the location and time of data generation. Therefore, we have proposed the STD retention system (STD-RS) using vehicles as a network infrastructure for local production and consumption of STD. This paper proposes a transmission control method that can mitigate the fluctuation of RSS due to the influence of obstacles in the real environment and evaluates its effectiveness through experiments on actual devices.

Software-defined networking (SDN) has notably improved networks by providing Machine Learning-Powered programming capabilities at the control plane (CP), making it easier to dynamically manage the network resources according to varying traffic conditions. However, the geographically remote location of the CP from the data plane (DP) leads to significant round-trip delays in the order of milliseconds, which can ad-versely impact the performance of delay-sensitive and real-time traffic. To address this issue, this paper proposes a novel in-network reinforcement learning (RL) inference framework that extends programming capability from the CP to the DP for fine-grained control of network resources to meet the Quality of Service (QoS) demands of real-time applications. The in-network RL inference is achieved by adopting a match-action table mapping strategy in the DP and validating it through programming protocol-independent packet processors (P4). A P4 meter extern is utilized to allocate bandwidth to individual traffic flows based on their QoS requirements. Our proposed strategy achieves in-network RL inference at the line rate with negligible processing overhead while reducing packet loss rate and jitter by up to 92% and 57%, respectively, compared to the CP-based approach. Additionally, we evaluate the performance of our proposed bandwidth allocation framework using a state-of-the-art deep-deterministic policy gradient (DDPG)-based RL agent with a heuristic priority experience replay (hPER) technique. Our proposed DDPG agent achieves a faster convergence rate, higher reward, superior training stability, and up to 56 % reduction in operational cost compared to two alternative agents,

In this paper we propose a novel Stochastic Computing (SC) Finite Impulse Response (FIR) filter design to improve the overall accuracy. Since SC is generated probabilistically, it incurs inherent SC process error, i.e. the larger number of SC process blocks, the larger error probability. Moreover, generation of SC from lower binary radix value produces higher percentage of error as compared to generation from higher value. In our SC FIR design, we reduce the number of SC processing blocks by reducing the number of FIR stages using fast FIR algorithm (FFA). Further we improve the inner-product performance by re-ordering the FIR filter parameters to generate higher-value SC input. Confirmed by the numerical analysis and simulation results, our novel design can implement higher-order FIR filter without reducing its output performance.

Internet of Things (IoT) is the reality of a new and powerful ubiquitous technology. One of its main driving forces is the 3rd Generation Partnership Project (3GPP) Long-Term Evolution (LTE), seeking to encompass all the applications of IoT. With this trend, 3GPP has finally made the Release 14 for LTE-based vehicle to everything (V2X) service. In this proposed paper, we evaluated the new LTE-based V2X architecture in regards to V2X message delivery and security requirements. We showed that a proper resource allocation and reference point (channel) selection could accommodate all types of V2X message deliveries. However, focusing more on security, we deemed that LTE-based V2X security falls short of meeting adequate security requirements, especially, to well preserve the privacy. Hence, we proposed a privacy preserving security for LTE-based V2X service. Considering the privacy as the top security requirement, we seamlessly integrate our security scheme with the specified LTE security architecture. Our scheme is scalable while fulfilling basic wireless message security requirements. We also provide the security and performance analysis to show the robustness and effectiveness of our proposed schemes.

Mobile cloud computing utilizing cloudlet is an emerging technology to improve the quality of mobile services. In this paper, to better overcome the main bottlenecks of the computation capability of cloudlet and the wireless bandwidth between mobile devices and cloudlet, we consider the multi-resource allocation problem for the cloudlet environment with resource-intensive and latency-sensitive mobile applications. The proposed multi-resource allocation strategy enhances the quality of mobile cloud service, in terms of the system throughput (the number of admitted mobile applications) and the service latency. We formulate the resource allocation model as a semi-Markov decision process under the average cost criterion, and solve the optimization problem using linear programming technology. Through maximizing the long-term reward while meeting the system requirements of the request blocking probability and service time latency, an optimal resource allocation policy is calculated. From simulation result, it is indicated that the system adaptively adjusts the allocation policy about how much resource to allocate and whether to utilize the distant cloud according to the traffic of mobile service requests and the availability of the resource in the system. Our algorithm outperforms greedy admission control over a broad range of environments.

In this paper, a novel resource allocation algorithm is proposed for secure mobile cloud computing systems. The mobile request for using cloud resource is classified according to its level of security requirement and the amount of required resource for remote computing. We formulate the resource allocation problem as a semi-Markov decision process under the average reward criterion, where the average reward of states is expected to be optimized. Through maximizing the long-term reward while meeting the system requirements of the blocking probability and the amount of resource requested with a security guarantee, the optimal resource allocation policy is calculated by using the linear programming. Simulation results demonstrate that the system adaptively modifies the resource allocation policy for cloud computing, and determines whether to utilize extra resource for security implementation according to the mobile request type, the current traffic, and the cloud resource availability.

Mobile applications are providing increasingly richer functionalities, which generally result in high computational complexity and thus high energy consumption of mobile devices. In this article, to alleviate the computational burden of mobile devices, we present a Dynamic Programming based Offloading Algorithm (DPOA) to quickly find the optimal partitioning between executing subcomponents of a mobile application at the mobile device and the cloud server, taking into account the CPU speed of mobile device, network performance, the characteristics of an application program, and the efficiency of cloud server. DPOA solves the offloading optimization problem with much lower complexity than the Branch & Bound used in [1][2], while significantly reducing the execution time of mobile application proved by the simulations.

One of the most important and yet most challenging issues in body area networks (BANs) is to provide diverse quality of service (QoS). Most physiological data monitoring applications require low rate periodic reporting while real-time entertainment applications require high rate continuous streaming. Emergency alarm, the most time-critical but unpredictable data, must be delivered instantaneously. A BAN should satisfy these diverse requirements since applications each with distinctive QoS requirement may run simultaneously. We propose PNP-MAC protocol that can flexibly handle variety of applications with diverse requirements through fast, preemptive slot allocation, non-preemptive transmission in the allocated slots, and flexible superframe adjustments. Performance evaluation using OPNET network simulator shows that PNP-MAC can satisfy diverse delay and throughput requirements of various applications such as continuous streaming, routine periodic monitoring, and time-critical emergency alarm.

This paper introduces a new IEEE standard, IEEE 802.15.5,which provides mesh capability for wireless personal area network (WPAN) devices. The standard provides an architectural framework enabling WPAN devices to promote interoperable, stable, and scalable wireless mesh topologies. It is composed of two parts: low-rate WPAN mesh and high-rate WPAN mesh. In this paper, we present only low-rate WPAN mesh because it is designed to support wireless sensor networks. IEEE 802.15.5 low-rate part is a light-weight scalable mesh routing protocol that caters well to the requirements of resource-constrained wireless sensor networks. By binding logical addresses to the network topology, IEEE 802.15.5 obviates the need for route discovery. This eliminates the initial route discovery latency, saves storage space and reduces the communication overhead and energy consumption. A distributed link state scheme is further built atop the block addressing scheme to improve the quality of routes, robustness, and load balancing. The routing scheme scales well with regard to various performance metrics. The standard also provides enhanced functions such as multicast, reliable broadcast, power saving, time synchronization, route tracing and portability. We also present the performance evaluation of major functions performed with a 50-nodes tested deployed over a whole floor (100 × 140 ft2) at CUNY Engineering building. The results testify that the IEEE 802.15.5 will serve well for wireless personal area networks and wireless sensor networks.

This article presents an in-depth introduction and analysis of the low-data-rate portion of the IEEE standard 802.15.5 that targets on providing mesh capabilities to low-rate wireless personal area networks. As major contributors of this standard, we seek to share our insights and motivations of the approaches adopted in the major components of the standard instead of presenting a la carte items drawn in the specification. We wish this article helps readers of the 802.15.5 standard to better understand the rationale and intent of the protocol design. This article also compares 802.15.5 to ZigBeetrade and 6LoWPAN, for users to select the best mesh network technology for their applications.

This paper focus on the testbed experimentation of a meshed tree routing algorithm with local link state for wireless personal area networks (WPANs) based on current IEEE 802.15.4 MAC and PHY. The meshed tree uses a block addressing scheme based on tree structure and local link state information for mesh routing. Our approach exploits the information from the global tree structure for the direction of packet forwarding and local link state for choosing the next hop toward the destination. It has two prominent features: scalability and fault tolerance. Each node maintains a local link state of k-hop(usually 2-hop) information regardless of network size, which makes the approach scalable. Being a mesh, our approach shows good fault tolerance and load balancing. Testbed experiments show that the meshed tree displays superior performance when compared with AODV and a tree based algorithm. Comparisons were made with respect to packet deliver ratio, energy consumption and memory usage. The proposed algorithm is being considered as the routing algorithm for IEEE 802.15.5 WPANMesh standard.

In wireless sensor networks, a periodic broadcast message or a beacon with strict time requirement can be con- tinuously collided when there are multiple nodes with the same period in the same space. This type of collision prevents a network from operating normally: a node is unable to join a network, and a network topology breaks if the periodic message contains critical information such as network and timing parameters as in IEEE 802.15.4. We show that the affected area by the collision occupies significant percentage of an original service area. This paper proposes a new algorithm to solve the permanent collision problem in periodic broadcasting or beaconing. The proposed stochastic beacon algorithm avoids the collision by incorporating randomness in beacon transmission while main- taining a fixed beacon period. The algorithm is evaluated by analysis and simulation based on IEEE 802.15.4, and compared with other 7 methods. The results show that our algorithm can avoid permanent beacon collision and reduces substantially the collision probability with very low added complexity.

In this paper, we propose a general framework for wireless network optimization across physical, medium access control (MAC), and routing layers, applicable to various multiaccess techniques and physical layer settings. The traffic on the entire network is specified by a sequence of routing matrices which together with crosslayer settings and network topology can determine the throughput for each node, each route, and the entire network. To illustrate the applicability, the framework is applied to a multihop ad hoc network based on CDMA, employing multiuser detection technique with retransmission diversity. It is also shown that the approach based on the proposed framework can improve the network throughput by exploiting retransmitted packets and interference cancellation.

In this paper, the effect of user cooperation on network performance is investigated in terms of throughput and stability region for CDMA random access networks where a large number of users randomly access an access point (AP). The network considered consists of two classes of users depending on the received power at the AP. A cooperation strategy with very low overheads is proposed, and the network performance gain obtained by adopting this strategy is characterized in closed-form formulas. The validity of our analysis is confirmed by simulation results.

Although network-wide flooding in reactive routing protocols like AODV facilitates the discovery of optimal routes, they engender large control overhead. Some efficient flooding methods cut down the overhead by minimizing the set of nodes participating in packet relays. When used for route discovery, however, these approaches may suppress the detection of optimal routes. A region-based routing (REGR) protocol was proposed in to effectively reduce route discovery overhead without sacrificing route optimality. REGR defines a pre-routing region and tries to find a high-quality route by broadcasting a route request message only within the region. Although the region-limited broadcasting reduces overhead, it is actually not efficient, and may experience frequent packet collision especially in dense networks. In this paper, we propose a progressive route calculation (PROQ) protocol to locate the optimal route around a preliminary route obtained from sketchy network scanning. We propose to use progressive Dijkstra's algorithm to make nodes on the preliminary route jointly derive a minimum cost spanning tree (MCST) between a source-destination pair. The topology information needed to derive the MCST is delivered along the preliminary route by using unicast packets, which makes the new protocol more efficient and reliable than broadcast approaches. We simulate PROC and compare it with various protocols. The simulation results verify the efficiency of our method.

This paper presents a new medium access control (MAC) scheme, which exploits the code division multiple access (CDMA) technology to provision many-to-one simultaneous wireless communication service without utilizing any time division multiple access (TDMA) scheme. This, combined with existing one-to-one (i.e., unicast) and one-to-many (i.e., multicast or broadcast) wireless communication services, brings into life a new communication paradigm, that is, mesh communications. The new scheme also supports accumulative receptions, namely, a receiver can buffer partially damaged (re-)transmitted frames and add those frames efficiently to form an error-free frame. Our simulation results show that the new scheme outperforms IEEE 802.11 in both non-mobile and mobile environments.

Multipath routing is typically utilized in wireless mesh networks for fault tolerance, throughput enhancement, and load balancing purposes. To achieve good route resilience and optimal end-to-end transmission performances, most multipath routing algorithms tend to establish link-disjoint or node-disjoint paths with minimized path costs. Although multiple paths created by these algorithms do not share common links or nodes, they are usually located close to each other. If all these paths are activated simultaneously to boost end-to-end throughput, serious inter-path interferences will happen. In this paper, we first introduce an AODV-based decoupled multipath routing protocol named AODV-DM. AODV-DM builds an insulating region around each established path to prevent new paths from being formed in the neighborhood of the established path. By this means, several "isolated" paths can be created and utilized for simultaneous data transmissions. We also present a novel destination-polling scheme to improve medium access efficiency in destination area, and at the same time alleviate the out-of-sequence problem caused by concurrent packet transmissions through multiple paths. By optimizing the functions crossing network and MAC layers, we are able to complete the design of an efficient multipath structure targeting at end-to-end throughput enhancement. Simulation results demonstrate that the decoupled multipath structure can achieve much higher throughput than single-path and coupled-path cases

Low rate wireless personal area networks (LR-WPANs) offer device level wireless connectivity. They bring to light a host of new applications as well as enhance existing applications. Due to their low cost, low power consumption and self-organization features, LR-WPANs are ideal for applications such as public security, battle field monitoring, inventory tracking, as well as home and office automation. Nevertheless, one critical issue, security, needs to be solved before LR-WPANs are commonly accepted. Pursuing security in LR-WPANs is a challenging task. On one hand, wireless communications are inherently susceptible to interception and interference. On the other hand, most devices in LR-WPANs are resource-constrained and lack physical safeguards. This paper presents a systematic analysis of the threats faced by LR-WPANs with respect to the protocol stack defined by IEEE 802.15.4 and the ZigBee Alliance. Attacks are modeled and their impacts are evaluated. Some security problems within the current LR-WPAN security architecture are identified and remedies are suggested. Countermeasures of various attacks are also given.

Designed and developed for military communications in their early stage, multi-hop wireless ad hoc network is attracting interest from commercial applications and becoming a research hot spot. Although the technology has been studied for about a decade, there still exist some problems to be solved. Route discovery latency for reactive routing protocols is one of them. In this paper, we propose a dynamic RTT-based algorithm which can dramatically reduce the route discovery latency. This algorithm can be generally applied to any source selection based reactive routing protocols. Our simulation study has also shown some interesting results about packet transmission time and route discovery latency in a contention-based channel like IEEE 802.11 which have not been discussed before.

The standard SCTP, RFC 2960, assumes that all packet losses are congestion losses. Thus, it does not perform well in wireless ad hoc networks where congestion losses and wireless losses coexist. This paper designs an alternative of SCTP, which is wireless-aware (W-A). The W-A SCTP is the first that simultaneously considers i) the nature of wireless losses (random or bursty losses), ii) the energy efficiency of battery powered ad hoc nodes, and iii) the goodput performance.

This paper introduces a resource management application of a probabilistic-based ant routing algorithm for mobile ad hoc networks (ARAMA) that is inspired from the ant's life . Mobile ad hoc networks (MANETs) are highly dynamic, self-configured and self-built networks. The goal of this paper is to present ARAMA ability to manage MANET's resources by achieving fair network resources distribution, while considering the dynamic characteristics of MANETs and the need for low control overheads. This paper provides a description for the algorithm. In this algorithm, the nodes' (node's energy, processing power, ...) and links' (bandwidth, ...) parameters are measured and collected in the nodes' indices. A path index is used to measure the path total resources and serves to minimize the forward control packet (ant) size. The concepts of negative backward ant destination trail are introduced to enhance the performance of the algorithm. The simulation results show the potential of ARAMA to achieve fair energy usage across the network nodes as an example of the network resource management. More, the results show the general ability of the algorithm to solve MANET's routing problem.

In wireless sensor networks, the multihop routes between sensors and sinks (data collectors) suffer breakage or damage because of the movement of targets and sinks, or the power exhaustion of some sensors. An efficient route maintenance and update protocol should give attention to both route rediscovery overhead and route optimization. In most flooding based routing protocols, route discovery/rediscovery packets are broadcast omni-directionally, which causes significant communication overhead as well as wireless channel contention. Query localization (QL) protocol (Castenada, R. et al., ACM/Kluwer Wireless Networks Journal, vol.8, no.2, p.137-51, 2002) limits the broadcast of route rediscovery packets within the local region of the old routes. However, as new routes formed under QL have to share most parts with old routes, route optimization is penalized. We propose to update the route by using the information of the old route, but at the same time, select a new route not necessarily overlapping the old one. This virtue is especially important to route-wide update cases, such as continuous movement of targets and sinks, and renovation of worn-out routes. Simulations show that, our new protocol always builds a fresher and better route with a very low control overhead.

Under many circumstances, broadcasting is the only way of service discovery, device paging, and even data transfer in wireless ad hoc networks. However, broadcasting may lead to "broadcast storm" problems if it is done by simple flooding in a single common channel. The problems include redundant rebroadcast (overlaying), contention and collision. When the network density is high or the network is getting congested, the problems become severe and may directly lead to degraded performance of the network. We propose a novel broadcast scheme that is simple, highly efficient and scalable. Most importantly, it is a fully distributed scheme which introduces no control overheads (in terms of control traffic), and can adapt to network movement at no extra cost. We evaluate our scheme by OPNET simulation and compare it with existing methods.

With the popularity of multi-homed hosts, which can be reached through multiple paths, a new challenge arises on how to efficiently and simultaneously use multiple paths between multi-homed communication hosts to explore potential benefits of multi-homing feature, such as to improve connection reliability, to increase throughput etc. Standard SCTP cannot support the multi-homing feature efficiently due to its per association congestion control structure. Per association congestion control structure makes it unfeasible to extend standard SCTP for load sharing, load balancing and other multi-homing applications. In this paper, we identified and analyzed limitations of the congestion control mechanism of standard SCTP in the context of using multiple paths simultaneously. We proposed an independent per path congestion control principle for standard SCTP to overcome these limitations. Algorithms are provided to realize independent per path congestion control in standard SCTP without modifying SCTP DATA chunk format and without introducing overhead to the network. SCTP with independent per path congestion control not only overcomes existing flaws of standard SCTP in supporting the multi-homing feature but also demonstrates its applicability to various multi-homing applications.

In this paper, we address various aspects related to extending and engineering stream control transmission protocol (SCTP) to utilize its multi-homing support for simultaneous transmission of data chunks on different path, while maintaining the congestion control on each path to ensure fair integration with other traffic in the network. The extended SCTP, referred to as load sharing-SCTP (LS-SCTP), is able to aggregate the bandwidth of all the active transmission paths between the communicating endpoints. It monitors the paths, and accordingly chooses the paths that are suitable for load sharing. In addition, LS-SCTP retransmission mechanism accelerates the delivery of missing data to the receiver to prevent stalling the transport connection while waiting for missing data chunks. Simulation results show that LS-SCTP is extremely beneficial for networks with limited bandwidth and failure prone links.

We propose a mechanism that combines retransmission-based error control with path diversity, to provide different levels of protection to interactive video in ad-hoc networks. The mechanism factors in the importance of the retransmitted packets to the reconstructed video quality as well as the end-to-end latency constraints to minimize the overhead and maximize the reconstructed video quality at the receiver. Simulation results show that the proposed retransmission mechanism maintains the video quality under different loss rates and with less overhead compared to error control methods that depend on controlling the intra-update rate. In addition, the mechanism is shown to be more robust to wireless losses than schemes that combine layered coding with path diversity.

The current stream control transmission protocol (SCTP) does not work well over lossy links. To make SCTP congestion control algorithms robust in lossy networks, the paper first introduces a fine-tuned explicit congestion notification (ECN) mechanism for SCTP in such environment, and then discusses the ECN-D SCTP that can differentiate noncongestion losses from congestion losses. We identify the optimal value of the congestion window for an SCTP source in response to ECN messages in order to maximize the throughput and maintain relatively small end-to-end delay. A simple and practical method to achieve the optimal value is developed by carefully choosing the threshold of queues that support ECN. Because the total goodput performance of SCTP associations is not sensitive to window reduction policies when the network load is heavy, and because fine-tuning SCTP or transmission control protocols congestion window in response to congestion indications using complicated methods may not be worth the increase in complexity of the protocol, the simplified method becomes attractive in achieving the optimal congestion window.

Bluetooth scatternets, integrating polling, and frequency hopping spread-sprectrum in their medium access control protocol, provide a contention-free environment for Bluetooth devices to access the medium and communicate over multihop links. Currently, most available scatternet formation protocols tend to interconnect all Bluetooth devices at the initial network startup stage and maintain all Bluetooth links thereafter. Instead of this "big scatternet" approach, we propose a scatternet-route structure to combine the scatternet formation with on-demand routing, thus eliminating unnecessary link and route maintenances. To the best of our knowledge, this is the first effort to address on-demand scatternet formation with every detail. We introduce an extended ID (EID) connectionless broadcast scheme, which, compared with original Bluetooth broadcast mechanism, achieves very much shortened route discovery delay. We also propose to synchronize the piconets along each scatternet route to remove piconet switch overhead and obtain even better channel utilization. Furthermore, we present a route-based scatternet scheduling scheme to enable fair and efficient packet transmissions over scatternet routes. Network performance analysis and simulations show that scatternet routes can provide multihop wireless channels with high network utilization and extremely stable throughput, being especially useful in the transmission of large batches of packets and real time data in wireless environment.

The regional foreign agent Mobile-IP approach was introduced to reduce the frequency by which registration with the remotely located home agent is needed. Providing an efficient system that support IP multicast, in an environment where the multicast group members frequently change their locations, is a challenge for systems providing mobility support. In addition, sensitive applications require a reliable multicast service such that each intended receiver is guaranteed to receive a copy of each transmitted packet. We propose an architecture that supports IP multicast in an environment where mobility is supported using regional foreign agents. The proposed scheme takes advantage of the inherent characteristics of the local registration to enhance the efficiency of the multicast support. Performance aspects such as multicast datagram delay, delivery cost and robustness are analyzed. Simulation results are used to describe the performance under various conditions. In addition, the proposed approach is extended to support reliable multicast. The proposed scheme tolerates the failure of regional foreign agents, and compensates for those datagrams undelivered due to such failures and for those due to the mobility of the receivers.

We study the link-sharing between the assured real-time traffic and the throughput intensive best-effort services. Our simulation shows that the throughput intensive best-effort services can degrade the QoS performance of real-time services if both are forwarded to a single FIFO queue. Therefore, it is advantageous to assign separate queues for individual classes of traffic. A novel link-sharing scheme (delay-sensing) is proposed to improve the QoS measures such as delay and jitter for the real-time traffic while preserving the fairness among different traffic classes. The new method guarantees an explicit delay bound for the real-time traffic when bandwidth borrowing is allowed. Another advantage of the method is that it can offer different priority classes with user specified fairness.

We investigate and evaluate the performance of these two TCP variants in a wireless mobile ad hoc network. The results in this paper are based on simulations using the NS2 network simulator from Lawrence Berkeley National Laboratory (LBNL), with extensions from the MONARCH project at Carnegie Mellon. We present our experimental results in two parts. The first part of the experiments does not include node movement. The effect of mobility and the link breakage are considered in the second part experiments. We consider one type of topology in the first part experiments: a string topology with 8 nodes as shown. The distance between any two-neighbor nodes is equal to 200 m, which lets a node can only connect to its neighbor node. In other words, only those nodes between which a line exists can directly communicate. In this performance study, we set up a single TCP connection between a chosen pair of sender and receiver nodes and measured the successively received packets over the lifetime of the connection. The TCP good-put result was measured for each connection, averaged over ten runs. We compare these two TCP variants and discuss the effect of TCP maximum window size window.

The Base Mobile-IP, the Mobile-IP with Route Optimization extension and Local Registration Mobile-IP are different approaches to supporting the mobility of IP hosts. Each one of those schemes has its own advantages and drawbacks. The route optimization solves the triangular route problem and the corresponding larger delay associated with the Base Mobile-IP with the expense of larger overhead for location management purposes. The Local Registration Mobile-IP scheme was introduced to enhance the performance by processing the MN's registration requests at a local agent. The local registration approach may effect other aspects of the Mobile-IP systems as the fault tolerance. The Base Mobile-IP and Mobile-IP with the Route Optimization schemes can be implemented on top of Hierarchical Local Registration Mobile-IP, thus it is important to tackle the issues affecting the performance in those environments. We present a mechanism to enhance the performance of the Route Optimization Mobile-IP systems and a platform to support foreign agents fault tolerance in systems supporting local registration.

With the growing interests in distributed applications, efficient and scalable multicast routing is of great concern. Source-based tree multicast routing protocols, however, become very inefficient when used with large number of sources over large scale internetworks. Shared-tree based multicast routing protocols like PIM and CBT being developed by the IDMR working group of the IETF, trade efficiency of multicast delivery trees with scalability, but there are some issues open to further study. The work presented in this paper, the core-manager based multicast routing (CMMR) architecture, adopts shared-tree based techniques to achieve good scalability. It worked out an innovative approach to issues like core selection and multicast delivery tree management. The basic idea behind CMMR is that the core-manager (CM) keeps tracking cores of each multicast delivery tree in order to maintain a brief image of the tree. Consequently, new members will almost always be directed to a nearby core so that tree cost is well confined. However, not every new member has to actually consult the CM thanks to a self-growing scheme, maintaining control overhead low. Simulations show that the CMMR outperforms other existing shared-tree based multicast routing protocols like PIM, and can do as well as expensive near-optimal Steiner tree heuristics when average costs of multicast delivery trees are compared. Further improving on scalability is sought by using a hierarchical architecture, which requires the multicast addresses be administratively scoped in accordance with the hierarchy. Management of candidate cores are restricted within individual hierarchical regions. In addition, the CMMR scheme is recursive, meaning that all routers run the same algorithm regardless of their hierarchical levels.

Through an intensive experimental study, the characteristics of cell delay variation (CDV) and CTD in an ATM network transporting constant bit rate (CBR) traffic under diverse operating conditions are evaluated. We examine the effect of changing network parameters, such as the number of nodes and background load conditions, on the performance parameters of both CDV and CTD. Our results revealed the distributions of CTD and CDV could not be fit to any of the standard distributions (normal, gamma, etc.). The results also present, the autocorrelation behavior of CTD as a function of both load and number of travelled hops. A verification study of the square root law (SRL) of the ATM Forum, on estimating the CTD standard deviation with increasing number of network nodes is also presented. Our results indicate that the SRL reveal very close estimates only under the assumption of a uniform load conditions across all nodes in the network.

This paper proposes a connection managing scheme, called the connection server (CS), for supporting intra-domain connectionless multicast over ATM. The intention of using CS is to relieve multicasting hosts from directly being involving in ATM call setup and release, so that the dynamic membership transition of a host group has minimal impact on all steady members in the group. It also eliminates the needs for dynamic address resolution, and enables virtual circuit (VC) reuse at membership management level. The paper presents the proof-of-concept implementation of the CS system, including the server/client signaling protocol and the ATM user datagram protocol (AUDP) which is designed as a typical CS application to provide multicast users with efficient transport services.

This paper discusses major issues involved in the design and implementation of a multipoint multimedia conference system, such as system architecture, conference management, session control, and intramedium and intermedia synchronization. In particular, emphasis is given to conference management and adaptive synchronization algorithms. The management of multiparticipants is based upon a distributed architecture for greater flexibility. The proposed synchronization algorithm is adaptive to network changes, eliminates the need for a global clock, and is immune to the clock frequency drift, while its realization is very simple and the involved overhead is minimal. The essence of the algorithm is partitioning the vicinity of the arrival epochs of multimedia objects into three regions and counting arrivals at each region. The function of the synchronizer is to shift the playback clock (PBC) according to the individual counter contents. The ideas proposed are implemented within a teleconference system on the Ethernet/FDDI using the TCP/UDP. Experimental results show that the proposed synchronization algorithm performs well in our network testbed environment.

An algorithm related to synchronization problems arising in a multimedia multiparty conference is presented in this paper. The proposed algorithm determines the set of packets generated periodically from different participants that are arriving at a node, either for mixing at the master of a conference, or for simply playing back at a regular participant of a conference. No global synchronization of the clocks is assumed. In this paper, the statistical approach rather than the deterministic is used to determine the proper set of packets that have to be mixed at a given time. The statistics are derived from the time stamp each packet carries along with it. The essence of the proposed algorithm is to estimate the expected packet arrival time (or reference time) for each participant. With the reference time at hand, the maximum jitter and the optimum waiting time for a mixer to wait packets from all participants can be determined. By employing feedback, the proposed algorithm is shown to lead to an optimum waiting/delay time. The error of the proposed algorithm is enumerated by the Chernoff bound, demonstrated by simulation, and shown to be acceptable in practical application.

In this paper, an algorithm to determine the set of packets generated continuously and periodically from different participants that are arriving at a node either for mixing at the master of a conference, or for simply playing back at a regular participant of a conference, is proposed. The essence of the algorithm is to estimate the average packet arrival time (or reference time) for each participant. With the reference time at hand, the maximum jitter and the optimum waiting time for a mixer to wait packets from all participants can be determined. The error of the proposed algorithm is enumerated by the Chernoff bound and is shown to be acceptable in practical application.

In practical ATM switch design, a proper dimensioning of buffer sizes and a cost effective selection of speed-up factor should be considered to guarantee a specified cell loss requirement for a given traffic. Although a larger speed-up factor provides better throughput for the switch, increasing the speed-up factor involves greater complexity and cost. Hence, it may not be cost effective to increase the speed-up factor for 100% throughput. Moreover, with a given buffer budget, an increase in the speed-up factor beyond a certain value only adds to the cell loss. The paper addresses design trade-offs existing between finite input/output buffer sizes and speed-up factor in a nonblocking ATM switch. Another important issue is the adverse effect on cell loss performance caused by nonuniform traffic (different traffic intensity and unevenly distributed routing). The paper analyzes cell loss performance of ATM switches with nonuniform traffic, and examines the effect of each nonuniform traffic parameter. The authors also provide an algorithm for effective buffer sharing that alleviates the performance degradation caused by traffic nonuniformity.>

In this paper, an algorithm to determine the set of packets generated continuously and periodically from different participants that are arriving at a node either for mixing at the master of a conference, or for simply playing back at a regular participant of a conference, is proposed. The essence of the algorithm is to estimate the average packet arrival time (or reference time) for each participant. With the reference time at hand, the maximum jitter and the optimum waiting time for a mixer to wait packets from all participants can be determined. An enhancement to improve synchronization which deals with the estimation of the time offsets between the individual periods of the sources and the period of the receiver is also presented. The error of the proposed algorithm is enumerated by the Chernoff bound and is shown to be acceptable in practical application.>

The authors study the performance of a nonblocking space-division packet switch, given that the traffic intensities at the switch not only are nonuniform but also change as a function of time. A finite-state Markov chain is used as an underlying process to govern the time variation of traffic for the entire switch. The packet arrivals at each input form an independent Bernoulli process modulated by the underlying Markov chain. The output address of each packet is independently and randomly assigned with probability distributions, which are also modulated by the Markov chain. Provided that the traffic on each output is not dominated by individual inputs the service time of each output queue for sufficiently large switches can be characterized by an independent Markov modulated phase-type process. A matrix geometric solution for the resultant quasi-birth-death type queuing process is presented. The maximum throughput is obtained at the system saturation. The performance of the switch is numerically examined under various traffic conditions. A contention priority scheme to improve the switch performance is proposed.>

The performance of a nonblocking space-division packet switch is studied given that traffics are imbalanced at input and output. Analysis shows that the performance of packet queuing delay at switch input, as well as the entire throughput of the switch, can be adversely affected by such imbalances. The work is then extended to examine a transient imbalance case, where the switch experiences the alternation of two transient periods, each at a different traffic imbalance mode. The alternation is modeled by a two-state Markov chain. Both balanced and imbalanced cases can be viewed as the two extremes of the transient case. It is observed that the system throughput and the queuing performance in the transient case heavily depend on both mean sojourn time and steady-state probability at each imbalance mode.>