Predicting the service performance of Internet applications is important in service selection, especially for video services. In order to design a predictor for forecasting video service performance in third-party application, two famous service providers in China, Iqiyi and Letv, are monitored and analyzed. The study highlights that the measured performance in the observation period is time-series data, and it has strong autocorrelation, which means it is predictable. In order to combine the temporal information and map the measured data to a proper feature space, the authors propose a predictor based on a Conditional Restricted Boltzmann Machine (CRBM), which can capture the potential temporal relationship of the historical information. Meanwhile, the measured data of different sources are combined to enhance the training process, which can enlarge the training size and avoid the over-fit problem. Experiments show that combining the measured results from different resolutions for a video can raise prediction performance, and the CRBM algorithm shows better prediction ability and more stable performance than the baseline algorithms.

The paper presents a novel approach to estimate the performance of MPI collective communications. Our objective is to help researchers to make appropriate decisions on their message-passing applications. For each collective communication, we attempt to apply LogGP and P-LogP standard point-to-point models. The resulted models are compared with the empirical data in order to identify the most suitable for performance characterization of collective operations. For the communications on large clusters with large size messages, the network contention problem can significantly affect the performance. Hence, to reduce the relative gap between the prediction and the measured runtime, the contention issue is also modeled, by a queuing theory analysis method, and taken in account with the total performance estimation. The experiments performed on a cluster which consists of 64 processors interconnected by Gigabit Ethernet network show encouraging results. For any collective operation, given a number of processors and a range of message sizes, there is at least one model that predicts the performance precisely. We could achieve a gap between the predicted and the measured run-time around 15%. Thus, by handling the contention problem, we could reduce around 80% of the relative gap.

Resource performance prediction is known to be useful in resource scheduling in the Grid. The disk I/O workload is another important factor that influences the performance of the CPU and the network which are commonly used in resource scheduling. In the case of disk I/O workload time-series, the adaptation of a prediction algorithm to new time-series should be rapid. Further, the prediction should ensure that the prediction error is minimum in the heterogeneous environment. The storage workload (i.e., the disk I/O load) is a dynamic variable. A prediction parameter based on the characteristics of the current workload must be prepared for prediction purposes. In this paper, we propose and implement the OPHB (On-Line Parameter History Bank). This is a method that stabilizes the incoming disk I/O workload time-series fairly quickly with the help of accurately determined ESM (Exponential Smoothing Method) parameters. The parameters are drawn from a history database. In the case of forecasting with ESM, a smoothing parameter must be specified in advance. If the parameter is statically estimated from observed data found in previous executions, the forecasts would be inaccurate because they do not capture the actual I/O behavior. The smoothing parameter has to be adjusted in accordance with the shape of the new disk I/O workload. The ESM algorithms utilise one of the accumulated parameter histories chronicled by OPHB's Deposit operation. When a new time-series is started, an appropriate parameter value is looked up in the Bank by OPHB's Lookup operation. This is used for the time-series. This process is fully adaptive. We evaluate the proposed method with SES (Single Exponential Smoothing) and ARRSES (Auto-Responsive SES) methods.

Efficient task scheduling is critical for achieving high performance in grid computing systems. Existing task scheduling algorithms for grid environments usually assume that the performance prediction for both tasks and resources is perfectly accurate. In practice, however, it is very difficult to achieve such an accurate prediction in a heterogeneous and dynamic grid environment. Therefore, the performance of a task scheduling algorithm may be significantly influenced by prediction inaccuracy. In this paper, we study the influence of inaccurate predictions on task scheduling in the contexts of task selection and processor selection, which are two critical phases in task scheduling algorithms. We develop formulas for the misprediction degree, which is defined as the probability that the predicted values for the performances of tasks and processors reveal different orders from their real values. Based on these formulas, we also investigate the effect of several key parameters on the misprediction degree. Finally, we conduct extensive simulation for the sensitivities of some existing task scheduling algorithms to the prediction errors.

This paper discusses performance prediction of clock generation PLLs using a ring oscillator based VCO (RingVCO) and an LC oscillator based VCO (LCVCO). For clock generation, we generally design PLLs using RingVCOs because of their superiority in tunable frequency range, chip area and power consumption, in spite of their poor noise characteristics. In the future, it is predicted that operating frequency will rapidly increase and supply voltage will dramatically decrease. Besides, rigid noise performances will be required. In this condition, it is not clear neither how performances of both PLLs will change nor the performance differences between both PLLs will change. This paper predicts and compares future performances of PLLs using a RingVCO and an LCVCO with a qualitative evaluation by an analytical approach and with design experiments based on predicted process parameters. Our discussion reveals that the relative performance difference between both PLLs will be unchanged. As technology advances, power dissipation and chip area of both PLLs favorably decrease, while, noise characteristics of both PLLs degrade, which indicates low noise PLL circuit design will be more important.

This paper describes the design and implementation of a testbed for predicting master/slave (M/S) programs written using Message Passing Interface (MPI) programs. The testbed, named M/S Emulator (MSE), aims at assisting developers in evaluating the performance of M/S programs and dynamic load-balancing strategies on clusters of PCs. In order to realize this, MSE predicts the communication time by using a realistic parallel computational model, an extension of the LogGPS model. This extended model improves the prediction accuracy on a large number of processors, because it captures the master's bottleneck: the overhead required for retrieving arrival messages from the slaves. Current MSE also employs a best effort emulation method for predicting the calculation time. In our experiments, MSE demonstrated an accurate prediction on clusters, especially on a larger number of nodes. Therefore, we believe that our extended model enables us to analyze the scalability of the M/S program performance.

In this paper, we apply the Semi-markov Memory and Cache coherence Interference (SMCI) model, which we had proposed for invalidating based cache coherent parallel computers, to an updating based protocol. The model proposed here, the SMCI/Dragon model, can predict performance of cache coherent parallel computers with the Dragon protocol as well as the original SMCI model for the Synapse protocol. Conventional analytic models by stochastic processes to describe parallel computers have the problem of numerical explosion in the number of states necessary as the system size increases. We have already shown that the SMCI model achieved both the small number of states to describe parallel computers with the Synapse protocol and the inexpensive computation cost to predict their performance. In this paper, we demonstrate generality of the SMCI model by applying it to the another cache coherence protocol, Dragon, which has opposite characteristics than Synapse. We show the number of states required by constructing the SMCI/Dragon model is only 21 which is as small as SMCI/Synapse, and the computation cost is also the order of microseconds. Using the SMCI/Dragon model, we investigate several comparative experiments with widely known simulation results. We found that there is only a 5. 4% differences between the simulation and the SMCI/Dragon model.

Predicting the performance of high speed wide area ATM networks (WANs) is a difficult task. Evaluating the performance of these systems by means of mathematical models is not yet feasible. As a result, the creation of simulation models is usually the only means of predicting and evaluating the performance of such systems. In this paper, we use measurements to validate simulation models of TCP/IP over high speed ATM wide area networks. Validation of simulations with measurements is not common; however, it is needed so that simulation models can be used with confidence to accurately characterize the performance of ATM WANs. In addition, the appropriate level of complexity of the simulation models needs to be determined. The results show that under appropriate conditions simulation models can accurately predict the performance of complex high speed ATM wide area networks. This work also shows that the user perceived performance is dependent on host processing demands.

This paper presents the performance prediction method on sentence recognition system which uses a finite state word automaton. When each word is uttered separately, the relationship between word recognition score and sentence recognition score can be approximated using the number of word sequences at a minimum distance from each sentence in the task. But it is not clear that how we get this number when the finite state word automaton is used as linguistic information. Therefore, we propose the algorithm to calculate this number in polynomial time. Then we carry out the prediction using this method and the simulation to compare with the prediction on the task of Japanese text editor commands. And it is shown that our method approximates the lower limit of sentence recognition score.