Summary Live migrations of virtual machines are required in cloud data centers in various contexts such as evacuating a host machine for maintenance, balancing the workloads on host machines, optimizing physical resource utilization, and meeting the custom demands of the user applications. Virtual machine migrations can, however, be costly in terms of both resources consumed for the migration as well as in terms of service level agreement (SLA) violations during the migration window. The cost is determined by many factors and they, in turn, are impacted by WHEN the migration happens during the lifetime of the virtual machine. Empirical studies show that if we have a window in which the migration is to be done (proactive rather than reactive where no such window exists), we can do it by carefully choosing the starting point within the window so that the SLA violation is minimum. In this paper, we propose a model to migrate virtual machines opportunistically to minimize the SLA violations. The idea is to find the time instants, called opportunities, at which, if the migration starts, the number and extent of SLA violations are minimal. Our experiments show that the migrations time savings with opportunistic migrations are up to 25%.

Energy efficiency is an important metric in the upcoming 5G scenario due to the increasing energy costs and the focus on sustainable technology. Paradigms like device-to-device communication and heterogeneous networks help in achieving high data rates, user capacity, and coverage promised by 5G, while simultaneous wireless information and power transfer (SWIPT) offers a green technique to have dual usage of the incoming RF wave in transferring energy and information simultaneously. SWIPT also helps in converting the harmful interference, incurred by the frequency reuse of D2D tier underlaying the Hetnet, into energy that can be stored for the future. This paper presents the use of two techniques of SWIPT, namely power splitting SWIPT and time splitting SWIPT, and gives a resource allocation scheme so as to improve the energy efficiency of the D2D tier. Minimum data rates are ensured for all users, cellular and D2D. Maximum bipartite matching is used for subcarrier allocation to cellular users and a stable many-to-one matching game is used for subcarrier allocation to D2D users. Extensive simulations are done to compare the performance of the two approaches.

In this study, the authors address the problem of joint resource block and power allocation in ultra‐dense device‐to‐device (D2D) networks underlaying cellular networks. They propose a fully distributed scalable scheme that ensures quality‐of‐service (QoS) of the cellular user equipments (CUEs) while maximising the sum rate of the D2D tier. The resource allocation problem is modelled as a Stackelberg game with pricing. The base station as the leader decides prices to be paid by the D2D users for reusing the resources of the CUEs, based on the QoS requirements of the CUEs, and broadcasts the prices. D2D pairs as followers use the pricing information and a completely uncoupled learning algorithm that guarantees Pareto optimal solution. Analytical proof is provided to establish the convergence of the learning algorithm. Simulation results show superior performance compared to two other existing schemes.

A Virtual Data Center (VDC) is a set of virtual machines (VMs) connected by a Virtual Network (VN) topology. Today's cloud data centers support dynamic requests for VDCs, by using software defined embedding strategies thatallow them to mesh multiple VDCs onto their Physical Data Center(PDC) network and machines. In this paper, we present a solution to this VDC embedding problem that achieves a higher acceptance rate by minimizing fragmentation, compared to existing strategies, while at the same time minimally disrupting the existing VDCs.

ABSTRACT In large‐scale wireless sensor networks, cost‐effective and energy‐efficient localization of sensor nodes is an important research topic. In spite of their coarse accuracy, range‐free (connectivity‐based) localization methods are considered as cost‐effective alternatives to the range‐based localization schemes with specialized hardware requirements.In this paper, we derive closed‐form expressions for the average minimum transmit powers required for the localization of sensor nodes, under deterministic path loss, log‐normal shadowing, and Rayleigh fading channel models. The impacts of propagation environment and spatial density of anchor nodes on the minimum transmit power for node localization are evaluated analytically as well as through simulations. Knowledge of the minimum transmit power requirements for localizability of a sensor node enables improving energy efficiency and prolonging lifetime of the network. We also propose a novel distance metric for range‐free localization in large‐scale sensor networks. The target and anchor nodes are assumed to be positioned according to two statistically independent two‐dimensional homogeneous Poisson point processes. Analytical expression for the average distance from a target node to its kth nearest neighbor anchor node is derived and is used for estimating the target‐to‐anchor node distances for localization. The Cramér–Rao lower bound on the localization accuracy for the new distance estimator is derived. Simulation results show the accuracy of the proposed distance estimate compared with some existing ones for range‐free localization. The results of our investigation are significant for low‐cost, energy‐efficient localization of wireless sensor nodes. Copyright © 2011 John Wiley & Sons, Ltd.

In this paper a cross-layer design of joint congestion control and power control is proposed for wireless sensor networks (WSNs). The multiple access communication for the WSNs is assumed to be ultra-wide bandwidth (UWB) time-hopping spread-spectrum (TH-SS) impulse radio. The joint congestion and power control is formulated as an optimization problem which is then solved by distributed iterative algorithms for flow rate adaptation and power adaptation on every link. The transport layer controls the congestion using a hop-by-hop flow rate assignment. Power control algorithm allocates the right amount of power at the right nodes to alleviate the bandwidth bottlenecks by increasing capacity, while maintaining the signal to interference-and-noise ratio above a given threshold for every link. Simulations show that considerable improvement in throughput can be obtained while also transmitting at lower power levels.

In this paper a performance study of the service differentiation mechanism 'stateless wireless adhoc networks (SWAN)', proposed for mobile adhoc networks (MANETs), is first carried out with adhoc on-demand distance vector (AODV), dynamic source routing (DSR) and associativity based routing (ABR) protocols. The paper then investigates any possible performance advantage by making SWAN benefit out of the features of the routing protocol The end-to-end probing mechanism of SWAN is reduced to probing the routing protocol within the source node. The burden of estimating the bottleneck bandwidth is thus passed on to the routing protocol.

Proceedings - Conference on Local Computer Networks, LCN

Achieving improved TCP performance together with ensuring end-to-end security necessitates the co-existence of security mechanisms like IPSEC and performance enhancing solutions. However, IP security and TCP performance have been traditionally dealt with in a mutually exclusive manner. We propose an innovative mechanism, secure performance enhancing proxy (SPEP) to address the seemingly arduous problem of enhancing TCP performance over wireless networks, preserving end-to-end TCP semantics as well as ensuring end-to-end security. The proposed SPEP scheme decouples loss detection and loss distinction mechanism from loss recovery mechanism which not only facilitates in performance improvement but also offers multifarious advantages discussed in the paper. We have implemented the proposed scheme in FreeBSD 4.5 and conducted experiments in a controlled test bed setup. Our results show improved TCP performance in a secured environment with introduction of minimal overhead.

Streaming multimedia (video/audio) over wireless access networks has problems of lower throughput and poor performance due to inherent wireless network characteristics of packet corruption. Solutions such as UDP-Lite are available that partially mitigate the effect of packet corruption over wireless hop. Meanwhile, because streaming applications preferably use unregulated UDP as transport protocol, there is an increasing need to regulate these streams for being friendly to other simultaneous flows. Recently, several slowly responsive congestion control algorithms such as TCP-Friendly Rate Control (TFRC) appropriate for real time streaming applications were proposed. However, this loss based TFRC doesn't work well in wireless scenario because it was designed for wired networks and it accounts wireless losses also as congestion losses. TFRC also suffers from degraded performance due to packet reordering in the Internet. Suggestions have been made to have a rate control mechanism based on ECN marking, which works well in wireless scenarios. We propose that ECN-based TFRC can also be robust to packet reordering. We first tested a combination of UDP-Lite and loss based TFRC for wireless streaming. Subsequently, we performed several experiments to study the robustness of ECN-based TFRC to both the losses over wireless hop and packet reordering in the Internet.

Depending on the bandwidth and round trip time of a network path, certain network connections may need different socket buffers than others. However, most TCP implementations use default socket buffer size for all the connections. This leads to a wastage of kernel memory for low bandwidth, long round trip time network paths and inefficiency for long fat W. Rechard Stevens (1994) network paths. To tackle this problem, we implement a TCP socket buffer auto-tuning daemon. It will periodically tune the socket buffer size of each connection to an optimal value according to the network status of each connection, without the intervention of user or change of network application. The tuning daemon can run on different machines and communicate with each other, send tuning request to the remote daemon and exchange the tuning information. Experiment results are presented to show that this daemon can improve the throughput for network application.

In recent years, a variety of new routing protocols for mobile ad hoc wireless NETworks (MANETs) have been developed. Performance evaluation and comparison of many of these routing protocols have been done using detailed simulation models. Zone routing protocol (ZRP) is one of these routing protocols, which is a hybrid routing protocol that proactively maintains routing information for a local neighborhood (routing zone), while reactively acquiring routes to destinations beyond the routing zone. The studies on ZRP have assumed homogeneous scenarios where all mobile nodes are statistically identical, lacking the studies on heterogeneous scenarios where mobile nodes behave differently in the same network. We study the performance of ZRP in such scenarios. We propose an efficient scheme for ZRP to adapt to the heterogeneous mobility scenario and study its performance for different mobility scenarios, network loads and network sizes

In spite of the larger performance gain such as higher throughput and almost zero packet retransmissions compared to TCP Reno, TCP Vegas still has a few obstacles for it to be deployed in the Internet. Studies have shown unfair treatment to Vegas connections when they compete with Reno connections. Other issues identified with TCP Vegas are problems of rerouting, persistent congestion, and discrepancy in flow rate tied with starting times and link bandwidth. We reinvestigate these issues rind propose modifications to the congestion avoidance mechanism of the TCP Vegas, with the slow-start and congestion recovery algorithms of Vegas remaining untouched. Unlike the solutions proposed in the recent past to deal with some of these issues, our solution it neither dependent on any critical parameter values nor on the buffer management scheme at the routers (e.g., RED). Our experiments show that the modified TCP Vegas (Vegas-A) it able to obtain a fairer share of the network bandwidth when competing with other TCP flows. We also show that Vegas-A can tackle rerouting issues and rectify Vegas's bias against higher bandwidth flows. At the same time, our experiments prove that Vegas-A preserves the properties of Vegas that have made it a noteworthy protocol.

A medium access control (MAC) technique for multi-hop wireless networks with quality of service (QoS) assurance is introduced The proposed protocol deals with most of the hidden node problems in a multi-hop environment and also provides service differentiation for various types of traffic: real-time constant bit rate (CBR) traffic, real-time variable bit rate (VBR) traffic, and non-real-time datagram traffic, by guaranteeing bounded delay for real-time traffic, at the same time avoiding lockout of datagram traffic. It uses a distributed mechanism of scheduling and reserving the channel access for real-time sessions. Transmission of datagram or starting packet of a new realtime session is granted only after checking the reservation such that they do not affect the scheduled ongoing real-time sessions. The protocol has been implemented and tested using the VINT Network Simulator 2 (NS2). The performance evaluation and comparison confirm the improved QoS features and multimedia support compared to various existing QoS enabled MAC protocols.

Standard TCP suffers in the wireless environment due to channel errors and frequent temporary disconnections of the wireless link. To solve both problems, we propose an end-to-end scheme, proactive-WTCP, in this paper. It only modifies the mobile end. Link layer retransmission scheme is used to deal with the wireless packet corruption. Mobile host withholds the duplicate acknowledgements for a dynamically calculated period in order to allow link layer recovery. Mobile host deals with the temporary disconnections by forcing the TCP sender into the persist mode before disconnection. Acknowledgements sent upon reconnection makes TCP sender exit the persist mode and resume the transmission. We evaluate proactive-WTCP in wireless LAN and GPRS network scenarios using throughput and transfer delay as the performance metrics for comparison. The simulation results show the significant improvements over TCP New Reno and Snoop with multicasting techniques.

In the past years, a variety of unicast and multicast routing protocols for mobile ad hoc wireless networks (MANETs) have been developed. Zone routing protocol (ZRP) is one of these unicast routing protocols, which is a hybrid protocol that proactively maintains routing information for a local neighborhood (routing zone), while reactively acquiring routes to destinations beyond the routing zone. In this paper, we adapt ZRP for multicast routing purpose, and call it multicast zone routing protocol (MZRP). MZRP is a shared tree multicast routing protocol, which proactively maintains the multicast tree membership for node's local routing zone at each node while establishes multicast tree on-demand. It is scalable to large number of multicast senders and groups. The use of IP tunnel mechanism during data packets transmission improves the overall performance. The evaluation and comparison with ODMRP are performed on the ns-2 simulator.

This paper reports a study on the performance enhancements of two extensions to the standard TCP implementation - Selective Acknowledgement (SACK) and Header Checksum (HACK) - over satellite links that are characterized by high latency and high bit error rate. We also examine the effectiveness of TCP Trunk, an edge-to-edge aggregation and congestion control mechanism, over the satellite link. Our study on the effect of varying the TCP window size over a long latency link for New Reno, SACK, HACK and TCP Trunk implementations show that increasing window size does improve the performance, but only up to a certain value of the window size, and a further increase actually reduces the performance. Other interesting observations from our experimental study are: SACK enabled TCP Trunk across the satellite link edge routers can improve the throughput regardless of the end host TCP implementation; disabling the link layer CRC and instead implementing the HACK extension to the TCP (and of course HACK+ SACK) can improve the throughput further.

The bias of TCP's congestion avoidance mechanism against connections with long round trip times (RTT) is a known fact. Many alternative congestion avoidance policies have been proposed to improve the fairness. Though the proposed policies attempt to address and resolve the fairness issue, we show that they tend to be harmful to connections that traverse either slow links, like 56 Kbps modem links, or long thin networks (LTN), like cellular links. We specifically consider a very common scenario where the last-hop link connecting the end user (i.e., client) may be a slow link or LTN. In this case, the TCP sender (i.e., server) is usually unaware of the network path, and when it is equipped with such policies, it increases probing into the network in a quest for the non-existent bandwidth. We conduct simulation studies to evaluate the impact of the proposed policies on connections that traverse either slow links or LTNs. We notice that the proposed policies cause increased buffer overflows at the last-hop router, thereby degrading the performance of the connection. We study the impact of increased buffer sizes at the last-hop router and also the effect of advertising a limited receive window. We show that the impact of the policies, on connections traversing slow links or LTNs, can be reduced by selectively disabling the policies.

In future wireless networks, different traffic classes will exhibit a large variety of characteristics and QoS requirements, such as transmission rate, maximum tolerable bit error rate and timeout specifications. However, currently there is no standard way of guaranteeing QoS in wireless access networks like wireless LAN based on IEEE 802.11. In this paper, we propose medium access control protocol enhancements and a distributed scheduler for QoS guarantees and fairness over IEEE 802.11 WLAN. The proposed scheme uses distributed scheduling at both the access point and the user terminals to schedule the transmission of packets according to their delay requirements. The algorithms used for both scheduling are the same. A flexible and fair resource allocation method among the traffic classes and the user terminals is provided by this scheme. Its performance has been evaluated using the UCB/LBNL/VINT Network Simulator and an implementation in the Linux kernel.

Markers, one of the building blocks of a traffic conditioner, play a major role for resource allocation in a DiffServ network. Recently, there has been a considerable research interest in designing intelligent markers, tailored for TCP traffic. The TCP dynamics make the design of a marker difficult in many respects. We list the issues related to designing a TCP-friendly marker and propose an intelligent marker, namely, memory-based marker (MBM) to address those issues. We illustrate the benefits of the proposed marker over some existing markers in a DiffServ network. The marker was implemented in NS simulator and extensive simulations were done to study the behaviour of MBM. Our results show significant improvement in TCP performance, especially in achieving fairness among priority flows with distinct round trip times, windows, and target rates. The marker is capable of protecting TCP flows in cases of congestion caused by unruly UDP flows. We also investigate the impact of coexisting assured service UDP flows on the assurance to the TCP flows. The major benefits of MBM are its simplicity, least sensitivity to parameters and transparency to the end hosts.

Wireless networks have become increasingly common and an increasing number of devices are communicating with each other over lossy links. Unfortunately, TCP performs poorly over lossy links as it is unable to differentiate the loss due to packet corruption from that due to congestion. We present an extension to TCP which enables TCP to distinguish packet corruption from congestion in lossy environments resulting in improved performance. We refer to this extension as the HeAder ChecKsum option (HACK). We implemented our algorithm in the Linux kernel and performed various tests to determine its effectiveness. Our results have shown that HACK performs substantially better than both SACK and NewReno in cases where burst corruptions are frequent. We also found that HACK can co-exist very nicely with SACK and performs even better with SACK enabled.

Markers, one of the building blocks of traffic conditioners play a major role for resource allocation in a DiffServ network. Recently, there has been a considerable research interest in designing intelligent markers, tailored for TCP traffic. The transient congestion levels in the network can affect the TCP throughput adversely and have great impact on the effectiveness of the marker. To address this issue, we propose an intelligent marker, which relies on an ECN-like feedback control mechanism between ingress and egress traffic conditioners. We present the control architecture and the marker algorithm, and illustrate the benefits of the proposed marker in a DiffServ network. The marker was implemented in the NS simulator and extensive simulations were done to study its effectiveness in the assurance of throughput, especially under the case that AS flows require different target rates, and co-exist with the unruly UDP flows. The ECN-like feedback mechanism that we use for the marker is also suitable in an MPLS domain for providing signaling to the ingress router of a congested label switched path (LSP) to mitigate congestion via dynamic traffic engineering techniques.

This paper examines the attributes of TCP tunnels which are TCP circuits that carry IP packets and benefit from the congestion control mechanism of TCP/IP. The deployment of TCP tunnels reduces the many flows situation on the Internet to that of a few flows. TCP tunnels eliminate unnecessary packet loss in the core routers of the congested backbones which waste precious bandwidth leading to congestion collapse due to unresponsive UDP flows. We also highlight that the use of TCP tunnels can, in principle, help prevent certain forms of congestion collapse described by Floyd & Fall (see IEEE/ACM Transactions on Networking, vol.7, no.4, p.458-72, 1999). Using a testbed often Intel PCs running the Linux operating system and traffic generators simulating user applications, we explore: the benefits which TCP tunnels confer upon its payload of user IP traffic; the impact on the congestion within network backbones, and the protection that tunnels offer with respect to the various competing classes of traffic in terms of bandwidth allocation and reduced retransmissions. The deployment of TCP tunnels on the Internet and the issues involved are also discussed and we conclude that with the RFC2309 recommendation of using random early drop (RED) as the default packet-drop policy in Internet routers, coupled with the implementation of a pure tunnel environment on backbone networks makes the deployment of TCP tunnels a feasible endeavour worthy of further investigation.

Multi-criteria QoS routing has been regarded by researchers as a vital mechanism to support multimedia communications. Unfortunately, finding an optimal path meeting the multi-criteria is indeed a very difficult problem to solve. This paper proposes a new QoS routing algorithm which can meet multi-criteria routing constraints, is simple to implement, and is robust. In brief, our approach is a bounded flood routing with multiple constraint checking and quick pruning at routers/switches. The quick pruning mechanism releases reserved resources at nodes not belonging to the final path, as early as possible during flooding. Bounding the flooding area within a reasonable scope reduces the routing overhead. Analysis and simulation study demonstrate that the proposed algorithm is superior to a similar flood routing scheme in terms of blocking probability, overhead and connection set-up time.

In this paper, we study multicasting in the self-routing multistage interconnection networks (MINs) for asynchronous transfer mode (ATM) switch architectures. Many B-ISDN applications require multicast connections in addition to conventional point-to-point connections. This paper presents a novel approach to support multicast connection, on the basis of a restricted address encoding scheme which constructs a short fixed-size multicast header and a recursive scheme that recycles a multicast packet one or more times through the network to send it to the desired destinations. The proposed two-phase multicast algorithm provides deadlock-free multiple multicast connections in MIN-based ATM switches. The emphasis is on analyzing the performance of an unbuffered MIN-based switch using the multicast algorithm in terms of network throughput. The proposed algorithm can be easily applied to buffered MIN-based ATM switches.

Considers an N/spl times/N nonblocking, space division, input queuing ATM cell switch, and a class of Markovian models for cell arrivals on each of its inputs. The traffic at each input comprises geometrically distributed bursts of cells, each burst destined for a particular output. The inputs differ in the burstiness of the offered traffic, with burstiness being characterized in terms of the average burst length. We analyze burst delays where some inputs receive traffic with low burstiness and others receive traffic with higher burstiness. Three policies for head-of-the-line contention resolution are studied: two static priority policies [shorter-expected-burst-length-first (SEBF), longer-expected-burst-length-first (LEBF)] and random selection (RS). Direct queuing analysis is used to obtain approximations for asymptotic high and low priority mean burst delays with the priority policies. Simulation is used for obtaining mean burst delays for finite N and for the random selection policy. As the traffic burstiness increases, the asymptotic analysis can serve as a good approximation only for large switch sizes. Qualitative performance comparisons based on the asymptotic analysis are, however, found to continue to hold for finite switch sizes. It is found that the SEBF policy yields the best delay performance over a wide range of loads, while RS lies in between. SEBF drastically reduces the delay of the less bursty traffic while only slightly increasing the delay of the more bursty traffic. LEBF causes severe degradation in the delay of less bursty traffic, while only marginally improving the delays of the more bursty traffic. RS can be an adequate compromise if there is no prior knowledge of input traffic burstiness.

In this paper we consider an N x N non-blocking, space division ATM switch with input cell queueing. At each input, the cell arrival process comprises geometrically distributed bursts of consecutive cells for the various outputs. Motivated by the fact that some input links may be connected to metropolitan area networks, and others directly to B-ISDN terminals, we study the situation where there are two classes of inputs with different values of mean burst length. We show that when inputs contend for an output, giving priority to an input with smaller expected burst length yields a saturation throughput larger than if the reverse priority is given. Further, giving priority to less bursty traffic can give better throughput than if all the inputs were occupied by this less bursty traffic. We derive the asymptotic (as N --> infinity) saturation throughputs for each priority class.