Grid resources pricing: A novel financial option based quality of service-profit quasi-static equilibrium model

Abstract

Use of grid resources has been free so far and a trend is developing to charge the users. The challenges that characterize a grid resource pricing model include the dynamic ability of the model to provide a high satisfaction guarantee measured as Quality of Service (QoS) -from users perspectives, profitability constraints -from the grid operator perspectives, and the ability to orchestrate grid resources for their availability on-demand. In this study, we design, develop, and simulate a grid resources pricing model that balances these constraints. We employ financial option theory and treat the grid resources as real assets to capture the realistic value of the grid compute commodities. We then price the grid resources by solving the finance model. We discuss the results on pricing of compute cycles based on the actual data of grid usage pattern obtained from the WestGrid and the SHARCNET. We extend and generalize our study to any computational grid. Background/IntroductionOver the years, there has been an upsurge in the development and applications of resources-intensive computations such as the Search for Extraterrestrial Intelligence at home (SETI@Home). The consequence is the emergence of a massively parallel resource-intensive computational facility -the computational grid (Grid computing). Research efforts have focused mostly on security related issues [15] and middleware/infrastructure-based issues [16]. Since grid resources usage has been free, there are only few efforts reported in the literature that develop models to price grid resources.Grid resources pricing is a challenging task when viewed as a generic pricing problem. In a grid system for example, the resources exist as non-storable commodities and are distributed across wide geographical regions, ownership is by dissimilar organization whose rights and polices are diverse. To price the grid resources, we assume that the grid compute commodities (gcc) are real assets. The gccs include CPU cycles, memory, network bandwidths, throughput, computing power, disks, processor, and various measurements, instrumentation tools. Since gccs are transient at t 0 , t 1 , · · · , t n , their availability (α) for usage fluctuates between "now" (t n|n=0 ) and "later" (t n > 0| t n =1 ) (option to wait) which accounts for uncertainty (ũ) where gcc : α → [0, · · · , 1]. The flexibility attributes of the gccs afford the applicability of real option defined in a fuzzy domain [0, · · · , 1] (also called membership function) [6].Real option theory provides a mechanism for valuating investment opportunities in situations where the cash flows are not deterministic. Real options can be applied when quantifying the accrued gains from prototyping a product as well as in managerial flexibility. Several of these gains can be hard to quantify [12] and as a result, a comparison of alternative opportunities becomes even harder if the managerial flexibility vary across the opportunities. Given that certain factors such as environmental and economics are inevitable, traditional val...

Integrating a financial option based model with GridSim for pricing Grid resources

“…The top layer consists of the grid scenario, user requirements, I/O interface, and application configuration. In Figure 2, we show our pricing architecture (which we integrate into the GridSim toolkit in Figure 3) to price the grid resources using financial option based model [1]. The integration of our financial option-based pricing model to the GridSim toolkit is at the top levels of the GridSim toolkit architecture.…”

“…In [1], we designed a pricing architecture -abstract representation that comprises of grid services layer and the price optimization layer. In this architecture, the grid services layer houses the middleware while our pricing model integrates on the top layer of the middleware.…”

“…To hold a real option means to have a certain possibility for a given time to either choose for, or against investment decision. In [14], we designed a financial option model and proposed a justification for price adjustments in using a global base prices and a control factor in p f [1]. Further study [2] showed that without unnecessarily charging for grid resource usage, we can satisfy both the user in terms of the Quality of Service (QoS) and the providers' incentives (profitability).…”

“…Pricing these compute commodities is challenging because of the specific characteristics of the grid resources [1]: heterogeneity of resources (geographically dispersed ownership and time zones affects availability of resources) and volatility of resources since they exist as compute cycles). These two characteristics (among others) accounts for a high level of flexibilities (uncertainty and fluctuation) of resources.…”

Committee-based evaluation and selection of Grid resources for QoS improvement

“…Many papers have been published to address to the QoS based resource reservation scheduling and negotiation [6,7] issue. Market based scheduling and negotiation algorithms [8] including the commodity model [9,10] and the auction model [11] appraise the quality of a resource as its attribution in the commerce based negotiation and mapping process. All these work focus on guaranteeing the QoS by ensuring every task is allocated with appropriate execution duration on a corresponding resource and little attention has been put on whether or not this resource can finish the task in time.…”