At existing cloud computing environment, virtual machine allocation mechanisms have been actively applied to resources management for processing dynamic workload in large-scale data center. In this paper, we address the problem of how many virtual machines should be allocated to a server, so that a given percentile of the execution time of a job is bounded by a predefined value. We first consider the case of “dedicated CPU”, whereby we calculate the CPU allocated to a single VM. Subsequently, we extend the analysis to the case where the CPU is shared equally by different groups of VMs. In this case, we calculate how many VMs need to be allocated in each group.

Software-Defined Networking (SDN) allows network capabilities and services to be managed through a central control point. Moving Target Defense (MTD) on the other hand, introduces a constantly adapting environment in order to delay or prevent attacks on a system. MTD is a use case where SDN can be leveraged in order to provide attack surface obfuscation. In this paper, we investigate how SDN can be used in some network-based MTD techniques. We first describe the advantages and disadvantages of these techniques, the potential countermeasures attackers could take to circumvent them, and the overhead of implementing MTD using SDN. Subsequently, we study the performance of the SDN-based MTD methods using Cisco's One Platform Kit and we show that they significantly increase the attacker's overheads.

Applications such as voice and video require network paths that satisfy several different quality of service (QoS) metrics, such as delay, jitter, packet loss rate and availability. The calculation of paths under multiple QoS metrics, such as the above four metrics, is a difficult problem since these metrics are in general incompatible. The authors propose a simple method for combining the above four QoS metrics into a single composite QoS metric which can be used as a link cost in Dijkstra's algorithm in order to calculate a path. The authors evaluated the proposed method in a multi-domain routing environment where domain reachability information is available through a service oriented architecture paradigm, and they show that it outperforms two commonly used methods. The results are also applicable to routing within a single domain.

An efficient analytical method is presented for the calculation of blocking probabilities in a tandem queuing network with simultaneous resource possession. This queuing network model is motivated from the need to model optical burst switching networks, where the size of the data bursts varies and the link distance between two adjacent network elements also varies depending on the network's topology. A fast single-node decomposition algorithm is developed to compute the blocking probabilities in the network. The algorithm extends the popular link-decomposition method from teletraffic theory by allowing dynamic simultaneous link possession. Simulation is used to validate the accuracy of the algorithm.

Resource allocation is a fundamental but challenging problem due to the complexity of cluster computing systems. In enterprise service computing, resource allocation is often associated with a service level agreement (SLA) which is a set of quality of services and a price agreed between a customer and a service provider. The SLA plays an important role in an e-business application. A service provider uses a set of computer resources to support e-business applications subject to an SLA. In this paper, we present an approach for computer resource allocation in such an environment that minimizes the total cost of computer resources used by a service provider for an e-business application while satisfying the quality of service (QoS) defined in an SLA. These QoS metrics include percentile response time, cluster utilization, packet loss rate and cluster availability. Simulation results show the applicability of the approach and validate its accuracy.

The composition of services has been a useful approach to integrating business applications within and across organizational boundaries. In this approach, individual services are federated into composite services which are able to execute a given task subject to a service level agreement (SLA). An SLA is a contract agreed between a customer and a service provider who define a set of several quality of services (QoS). An SLA violation penalty is a way to ensure the credibility of an advertised SLA by a service provider. In this paper, we consider a set of computer resources used by a service broker who represents service providers to host enterprise applications for differentiated customer services subject to an SLA and its violation penalty. We present a novel framework for a QoS-constrained resource provisioning problem, and propose a capacity planning approach to optimizing computer resources for all service sites owned by service providers subject to multiple QoS metrics defined in the SLA and their violation penalties. Simulation results show that the proposed approach is efficient for reliable resource planning in service composition.

We consider a set of computer resources used by a service provider to host enterprise applications subject to service level agreements. We present an approach for resource optimization in such an environment that minimizes the total cost of computer resources used by a service provider for an enterprise application while satisfying the QoS metric that the response time for executing service requests is statistically bounded. That is, gamma% of the time the response time is less than a pre-defined value. This QoS metric is more realistic than the mean response time typically used in the literature. Numerical results show the applicability of the approach and validate its accuracy.

In this paper, we study a new concept of traffic grooming in wavelength-division multiplexing (WDM) ring networks that aims at eliminating both the bandwidth underutilization and the scalability concerns that are typical of all-optical wavelength-routed ring networks. Our objective is to reduce the network cost while preserving the benefits of alloptical WDM ring networks. In order to assess the efficiency of our proposal, all underlying network costs are compared. These costs include that of the transceivers required at node level, as well as the number of wavelengths. Our results show that the proposed aggregation technique can significantly improve the resource utilization while reducing the network cost.

With the number of e-Business applications dramatically increasing, service level agreement (SLA) plays an important part in Web services. A SLA is a combination of several quality of services (QoS), such as security, performance, and availability, agreed between a customer and a service provider. Most existing research addresses only one QoS metric, and in the case of the response time, the average time to process and complete a job is typically used. In this paper, we study trustworthiness, percentile response time and availability. We consider all these qualities for a trust-based resource allocation problem which typically arises in Web services applications. We formulate the trust-based resource allocation problem as an optimization problem under SLA constraints, and we solve it using an efficient numerical procedure

In enterprise computing, customer requests often need to be distinguished, with different request characteristics and customer’s different service requirements. In this paper, we consider a set of computer resources used by a service provider to host enterprise applications for differentiated customer services subject to a service level agreement. We present an approach for resource optimization in such an environment that minimizes the total cost of computer resources used by a service provider for such an application while satisfying the QoS metric that the response time for executing differentiated service requests is statistically bounded. That is, each ã(r)% of the time the response time is less than a pre-defined value for class r customers. This QoS metric is more realistic than the mean response time typically used in the literature. Numerical results show the applicability of the approach and validate its accuracy.

In this paper, we propose and evaluate a new concept of traffic aggregation in wavelength-division multiplexing (WDM) optical networks that aims at eliminating both the bandwidth underutilization and the scalability concerns that are typical of all-optical wavelength-routed networks. Our objective is to reduce the network cost while preserving the benefits of all-optical wavelength-routed networks. In order to assess the efficiency of our proposal, all underlying network costs are compared. These costs include that of the transceivers and cross-connect ports required at node level, as well as the number of wavelengths. Our results show that the proposed aggregation technique can significantly reduce the network cost

We consider an edge optical burst switching (OBS) node with or without converters, and with no buffering. The OBS node serves a number of users, each connected to the switch over a fiber link that supports multiple wavelengths. Each wavelength is associated with a 3-state Markovian burst arrival process. The arrival process permits short and long bursts to be modeled. We model the edge OBS node as a closed nonproduct-form queueing network, with multiple heterogeneous classes, and we develop a suite of approximate decomposition algorithms to analyze it. Our approximate algorithms have a good accuracy, and they provide insight into the effect of various system parameters on the performance of the edge OBS node.

We present an approximate analytical method to compute efficiently the call-blocking probabilities in wavelength-routing networks with multiple classes of both unicast and multicast calls. Our approach involves the following steps. We start with an approximate solution to a linear single-class unicast network which we developed earlier. Next, all classes of calls on a particular route are aggregated to give an equivalent single-class model. We then extend the path decomposition algorithms that we have developed for single-class networks to handle mesh networks with multiple classes of calls. We show how to use these path decomposition algorithms to decompose large networks with multicast paths into smaller subsystems with only linear paths, which, in turn, are solved by the product-form approximation algorithm. We also consider a state-dependent Poisson arrival process for multicast calls which is more accurate in capturing the behavior of these calls.

We study a class of circuit-switched wavelength-routing networks with fixed or alternate routing and with random wavelength allocation. We present an iterative path decomposition algorithm to evaluate accurately and efficiently the blocking performance of such networks with and without wavelength converters. Our iterative algorithm analyzes the original network by decomposing it into single-path subsystems. These subsystems are analyzed in isolation, and the individual results are appropriately combined to obtain a solution for the overall network. To analyze individual subsystems, we first construct an exact Markov process that captures the behavior of a path in terms of wavelength use. We also obtain an approximate Markov process which has a closed-form solution that can be computed efficiently for short paths. We then develop an iterative algorithm to analyze approximately arbitrarily long paths. The path decomposition approach naturally captures the correlation of both link loads and link blocking events. Our algorithm represents a simple and computationally efficient solution to the difficult problem of computing call-blocking probabilities in wavelength-routing networks. We also demonstrate how our analytical techniques can be applied to gain insight into the problem of converter placement in wavelength-routing networks.

We analyze a multilayered queueing network that models a client-server system where clients and servers communicate via synchronous and asynchronous messages. The servers are organized in groups such that they form a multilayered hierarchical structure. The queueing network is approximately analyzed using a decomposition algorithm. Numerical tests show that the approximation algorithm has a good accuracy.

We study a class of circuit-switched wavelength routing networks with and without wavelength converters, and we present the first part of a new analytical framework to accurately and efficiently evaluate the blocking performance of such networks. The model allows non-uniform traffic, it accounts for the correlation among the loads at all links in a path, and it can be used when the location of converters is fixed but arbitrary. We first construct an exact Markov process that captures the behaviour of a path in terms of wavelength use. We also obtain an approximate Markov process which has a closed-form solution that can be efficiently computed for short paths. We then develop an iterative algorithm to analyze approximately arbitrarily long paths. The algorithm decomposes a path into shorter segments which are then studied in isolation using the corresponding approximate Markov process. The individual solutions are appropriately combined to obtain a solution for the original path. Finally, we demonstrate how the analytical techniques can be used to gain insight into the problem of converter placement in wavelength routing networks.

We consider broadcast WDM networks operating with schedules that mask the transceiver tuning latency. We develop and analyze a queueing model of the network in order to obtain the queue-length distribution and the packet loss probability at the transmitting and receiving side of the nodes. The analysis is carried out assuming finite buffer sizes, non-uniform destination probabilities and two-state MMBP traffic sources; the latter naturally capture the notion of burstiness and correlation, two important characteristics of traffic in high-speed networks. We present results which establish that the performance of the network is a complex function of a number of system parameters, including the load balancing and scheduling algorithms, the number of available channels, and the buffer capacity. We also show that the behavior of the network in terms of packet loss probability as these parameters are varied cannot be predicted without an accurate analysis. Our work makes it possible to study the interactions among the system parameters, and to predict, explain and fine tune the performance of the network.

We consider the problem of supporting multipoint communication at the media access control (MAC) layer of broadcast-and-select WDM networks. We first show that bandwidth consumption and channel utilization arise as two conflicting objectives in the design of scheduling algorithms for multicast traffic in this environment. We then present a new technique for the transmission of multicast packets, based on the concept of a virtual receiver, a set of physical receivers which behave identically in terms of tuning. We also show that the number κ of virtual receivers naturally captures the tradeoff between channel utilization and bandwidth consumption. Our approach decouples the problem of determining the virtual receivers from the problem of scheduling packet transmissions, making it possible to employ existing scheduling algorithms that have been shown to successfully hide the effects of tuning latency. Consequently, we focus on the problem of optimally selecting the virtual receivers, and we prove that it is NP-complete. Finally, we present four heuristics of varying degrees of complexity for obtaining virtual receivers that provide a good balance between the two conflicting objectives.

In this paper, we present an approximate analysis of a generic shared-medium ATM switch with input and output queueing. Input traffic is assumed to be bursty and is modelled by an interrupted Bernoulli process (IBP). Three different bus service policies are analyzed: time division multiplexing (TDM), cyclic, and random. The output links may have constant or geometric service time. The analysis is based on the notion of decomposition whereby the switch is decomposed into smaller sub-systems. First, each input queue is analyzed in isolation after we modify its service process. Subsequently, the shared medium is analyzed as a separate sub-system utilizing the output process of each input queue. Finally, each output queue is analyzed in isolation. The results from the individual sub-systems are combined together through an iterative scheme. This method permits realistic system characteristics such as limited buffer size, asymmetric load conditions, and nonuniform destinations to be taken into consideration in the analysis. The model's accuracy is verified through simulation.

Presents an approximation algorithm for the analysis of a buffered banyan ATM switch which allows complex (input and output buffering) switching elements, bursty traffic, non-uniform destination distribution and permits the analysis of large scale switches.>

A single-hop OSI structured network with multiple layers of sliding window flow control and packet fragmentation between layers is analyzed. An approximation algorithm is presented to hierarchically reduce the network to a single queue whose performance characteristics represent the original network. Network transmission characteristics are restricted to Erlang distributions. Validation against exact and simulation results showed that the approximation algorithm has a satisfactory error level.>

The authors present an analytical model of a buffered banyan asynchronous transfer mode (ATM) switch which allows complex switching elements, bursty traffic, and nonuniform destination distributions and permits the analysis of large-scale switches. The ATM switch is analyzed by decomposing it into individual switching elements. Each switching element is then analyzed numerically in isolation assuming that its arrival and service processes are known. The parameters of the arrival and service processes of the switching elements are obtained using an iterative scheme. The results obtained are approximate and validation tests have shown that they have good accuracy. Using this model, the cell loss, throughput, and the mean time to traverse the switch were obtained for different traffic parameters and buffer sizes within a switching element.>

The authors consider a multiclass open queueing network with class-dependent window flow control; that is, the total number of jobs of each class that may be present in the network cannot exceed a given value. A job that arrives at the network during the time that the current number of jobs of the same class is equal to the population constraint, is forced to wait in an external queue. A method is presented for obtaining an approximate solution of such a queueing network. The method is based on the use of an equivalent closed queuing network model, which is analyzed using an approximate product-form solution technique. The performance parameters of the original open queueing network are easily derived from the equivalent closed queueing network. Numerical results show that this method is fairly accurate.>

An iterative approximation algorithm for analyzing a closed queueing system with a K-sibling fork/join queue is presented. The iterative procedure is based on a combination of nearly complete decomposability and the Gauss-Seidel method. The approximation procedure gives good results for the mean response time and the system throughput. The iterative procedure converges to the exact solution in the case of the closed 3-sibling fork/join queue.>

An open queuing network for analyzing multilayered window flow control mechanisms consisting of different subnetworks is presented. The number of customers in each subnetwork is controlled by a semaphore queue. The queuing network is analyzed approximately using decomposition and aggregation. The approximation was validated against exact numerical and simulation data, and it was found to have an acceptable relative error. The queuing model is easy to implement, and it can be included in a software package. Using this queuing network, a case study involving the modeling and analysis of the ISO X25 flow control mechanism is presented.>

An approximation method for obtaining the throughput of cyclic queueing networks with blocking as a function of the number of customers in it is presented. The approximation method was developed for two different blocking mechanisms. It was also extended to the case of the central server model with blocking. Validation tests show that the algorithm is fairly accurate.>

A closed cyclic queueing system with multiple classes is analyzed numerically. The queueing system was originally motivated by a need to model circuit switching traffic in a data multiplexing scheme which provides both circuit and switching capabilities. The queueing system consists of two nodes, a primary and a secondary node. Customers in the primary node may require a random number of servers simultaneously. The secondary node simply provides the means of modeling the arrival process to the primary node. An efficient numerical procedure is used to obtain performance measures such as throughput, queue-length distribution, and utilization.

An approximation procedure is developed for the analysis of tandem configurations consisting of single server finite queues linked in series. External arrivals occur at the first queue which may be either finite or infinite. Departures from the queueing network may only occur from the last queue. All service times and interarrivai times are assumed to be exponentially distributed. The approximation algorithm gives results in the form of the marginal probability distribution of the number of units in each queue of the tandem configuration. Other performance measures, such as mean queue-length and throughput, can be readily obtained. The approximation procedure was validated using exact and simulation data. The approximate results seem to have an acceptable error level.

The exponential open queue network model studied here consists of n symmetrical queues in parallel served by independent first-level servers in tandem with a second-level server. Blocking of the flow of units through a first-level server occurs each time the server completes a service. The server remains blocked until its blocking unit completes its service at the second-level server. An approximate expression of the probability distribution of the number of blocked first-level servers conditioned upon a service completion of a first-level server is obtained. This expression compares well with simulation data. Based on this distribution, an approximate expression of the queue-length probability distribution is derived assuming a processor-sharing type of service. The exact condition for stability of the queue network is also derived. Some potential applications are discussed, and a quantitative evaluation of the model is given through a case study.

In a disc I/O system, the execution time of a read or write request depends partially upon other requests being executed at the same time. This concept is developed in this paper, and it is then used together with some standard results from queueing theory to produce a simple nonlinear regression model for predicting the response time of a disc I/O system. The model is of wide applicability although it was developed originally for a specific system.