Applied Mechanics and Materials Online: 2013-08-30 ISSN: 1662-7482, Vol. 379, pp 235-239 doi:10.4028/www.scientific.net/amm.379.235 2013 Trans Tech Publications, Switzerland Assessing efficiency of cloud-based services by the method of linear programming S.V. Razumnikov 1, a 1 652055, Russia, Kemerovo region, Yurga, Leningradskaya str.26 Yurga Technological Institute branch of Tomsk Polytechnic University a demolove7@inbox.ru Keywords: Cloud-based services, assessment of efficiency, linear programming, model, information technologies. Abstract. This paper summarizes cloud computing services presented at the market nowadays, describes their features, and introduces an economic and mathematical modeling algorithm for assessing the output of cloud-based services using the linear optimization method. Introduction We cannot think of a single person living in the modern world without using information technologies. For the last few years, information technologies have been intensively introduced to all spheres of life and business in Russia. Enterprises spend significant parts of their budgets on IT. It is commonly agreed that old-style data centers are now endangered species. For the last 2 3 years, cloud-based technologies (services) keep gaining popularity, although they are still in the formative stage and quite new to Russia and especially to corporate information systems (IS). Benefits introduced by cloud computing are huge; however, they require a thorough risk assessment. An enterprise lured by quick profits and reduced costs mentioned whenever migration to cloud computing is brought up may lose a lot of money if it does not consider the potential risks. Today, more and more IT managers choose cloud-based technologies. The core idea of cloud computing is enabling access to highly standardized pools of convenient services and software which altogether allow for highly efficient use of resources. Lack of serious research works devoted to evaluation of risks and efficiency of cloud computing prevents many firms from migrating to the cloud model. At the present, a number of assessment methodologies are offered, which are primarily aimed at investment projects. Only a small part of them addresses information technologies, and when it comes specifically to cloud-based services, there are no methodologies at all. That is why analysis of risks and efficiency of cloud-based services is an important problem. As a result, there is a pressing need for developing a method for assessing efficiency and risks of cloud-based services in corporate ISs. The concept and characteristics of cloud computing The term cloud computing was derived from the cloud symbol commonly used for denoting remote computers on network diagrams. The concept cloud computing is rather simple, though uncommon. The consumer s computer is responsible only for user interaction, while the main computational resources are provided by servers on the network (which are in fact the cloud). The simplest and the most common example is file hosting services such as Dropbox. Here, we double the benefit: firstly, there is no need for additional space on local disks of the user s computer (although it is not a problem anymore nowadays); secondly, the user can access his or her files from any computer connected to the Internet. Obviously, file storage is far not the only application of cloud technologies. A cloud may host not only data, but also applications. As an example, Google Docs and Zoho Docs use such an approach. The cloud model offers developers many advantages [1]. First, it makes software piracy practically impossible, since software resides only on the servers. Second, software becomes very easy to sell: it is not bound to any physical media and, moreover, not even handed to the consumer. This also All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of Trans Tech Publications, www.ttp.net. (ID: 130.203.136.75, Pennsylvania State University, University Park, USA-17/05/16,01:38:30)
236 Innovation Materials and Manufacturing Technologies, Economic Aspects in Enterprises makes technical support extremely easy, because client-side failures are very rare and are mostly eliminated by simply clearing the web browser cache. Of course, there are downsides, too: cloud-based applications require the software vendor to supply huge computational capabilities. To hit the market, the vendor needs to ensure failsafe operation of all services and an acceptable level of privacy and protection against attacks. Many people are opposed to cloud computing services. They emphasize the fact that a user of a cloud service does not actually control his or her own data. Even if the service is secure and hacker-proof, law enforcement authorities can access user data on legal grounds without informing the owner. That would be much less possible if only local software was used. And if the service is not secure enough, even more dangers arise. By hacking it, an intruder gains control over data of many users and possibly their computers, too [3]. Types of cloud-based services Any cloud-based service involves storage of data on a server in the Internet. What differs is kind and purpose of that data. Plain file hosting does not interest users any more. Today, one of the most promising cloud-based models is software delivered as a service. Software as a service (SaaS) is a software delivery model in which software is hosted on a vendor s server and users accesses it via the Internet (mostly, using a web browser). All documents created by the user are also stored at the vendor s server. If the user needs to edit an existing local document, it first needs to be uploaded to the server. Later on, the user can access all of his or her documents from any computer. Often, cloud-based software go beyond mimicking functions of ordinary applications and introduce their own features. First of all, these include file sharing: documents residing on a server readily support assigning a list of users allowed to view a them. Cloud-based applications consist of two parts: the client (loaded to the user s computer and stored in the web browser cache) and the server part. The client may implement a significant part of the functionality (document editing and viewing) or just display results of operations executed on the server; however, the client part cannot run independently. To function properly, the client may require additional plug-ins such as Java Runtime Environment or Adobe Flash Player; recently, the JavaScript language is widely used. If most of applications we use are hosted in a cloud, why not migrate user interface elements even the whole desktop into a cloud, too? Such a solution is implemented in the Desktop as a Service (DaaS) concept. A cloud-based desktop, just like other cloud-based applications, is accessed via a web browser and represents a web page with cloud applications represented either as external links or as embedded elements. By setting such a desktop as the browser s home page, a user can work entirely in the cloud, not even bothering to use locally installed software [4]. The key to providing cloud services to a wide variety of devices is to implement two-way awareness between the client and the cloud. Obviously, capabilities of client devices may differ, and the cloud is not always available to clients. That is why a uniform service delivery model is not appropriate. Inconsistency between the service delivery model and the device s capabilities can negatively affect employee productivity and the system s functionality and security and invalidate investments to development of cloud-based services [1]. Problem statement IT implementation projects are in fact investment projects, but their financial output is less evident, and the risks are higher. An IT project has a significantly wider scope because it addresses not only initial financial investments, but also post-deployment stages such as maintenance, support, updating, training, etc. All of these require extra resources and efforts. At this point, cloud-based services are more convenient since they do not require complex or expensive software and hardware; however, they imply higher risks and less evident efficiency. All popular methodologies for assessing efficiency of IT implementation are based on simulation models and do not provide any optimization algorithms. However, in some cases, a linear programming problem can be formalized and solved where one of the financial rates will be used as
Applied Mechanics and Materials Vol. 379 237 the objective function, and the rest will be constrained. Moreover, all of these methodologies are static in the sense that they pre-evaluate investment costs, the implementation program, etc., unlike dynamic models which allow correction of the rates at the t-th step depending on the rate values at the previous steps [5]. Modeling assessment of efficiency of cloud-based services Let us consider an example of usage of linear programming for assessing efficiency of various cloud-based services used for creating advertisements. We will analyze creation of 3 types of advertisements: a presentation, a video advertisement, and an audio advertisement. To create the advertisements, we need to do the following: 1. Create and/or edit photographs (slides). 2. Create a soundtrack, cut and mix music. 3. Create and a process a video clip. 4. Mix down all the components. 5. Store the files in a cloud. 6. Share the files. These tasks can be accomplished using the following cloud services. 1. AnyMaking: photo and picture editing. 2. PicJoke: creation of photo effects (no Photoshop skills required). 3. Mp3freemusic: music search and download service. 4. Mp3cut: service for creating ringtones and splitting music. 5. Jaycut: cloud-based video editor. 6. Minus: cloud-based file repository. 7. Share: quick transfer of files of any size. Source data for the problem of effective use of cloud-based services is presented in Table 1. Table 1. Source Data for the Problem of Effective Use of Cloud-based Services Time Costs for Processing One Advertisement (a ij ) Total Cloud-based Services Presentation Video Audio Working advertisement advertisement Hours (b i ) AnyMaking 2 2-30 PicJoke 1 1-25 Mp3freemusic 2 3 3 30 Mp3cut 2 4 3 35 Jaycut 2 3-20 Minus 5 5 5 50 Share 1 1 1 15 Profit from the sale of one type of the advertisement (efficiency)(c j ) 23 30 20 The table lists the services and estimated time needed to complete each work stage using a cloud service. It also specifies total working hours for each service which must not exceed the constraints and profit resulting from the sale of a particular kind of advertisement. Let us evaluate the effect of usage of the cloud-based services; specifically, we will calculate how many advertisements we can create using these services within a limited period of time to maximize the profit. Let us introduce the unknown variables: x 1: quantity of presentations; x 2 : quantity of promo video advertisements; x 3 : quantity of promo audio advertisements.
238 Innovation Materials and Manufacturing Technologies, Economic Aspects in Enterprises The general form of the economic and mathematic model (Formula 1): i F(x) = c j * x max; (1) The form applied to our example (Formula 2): F(x)= 23x + 30x2 + 20x3 1 max; Let us compose the system of constraints on the objective function in the general form (3): (2) a11x1 + a12x2 +... + a xi b, ij 1 a21x1 + a22x2 +... + aijxi b2,... aijx1 + aijx2 +... + a xi bi. ij Applied to our example (4): (3) 2x1 + 2x2 30, x1+ x2 25, 2x1 + 3x2 + 3x3 30, 2x1 + 4x2 + 3x3 35, 2x1 + 3x2 20, 5x1 + 5x2 + 5x3 50, x1 + x2 + x3 15. After solving the problem using simplex tables [2], we get the following basis: х 1 =1, х 2 =6, х 3 =3. The optimal value of the objective function (profit) equals 263 monetary units. Let us verify the solution using the MS Excel Solver feature (Figure 1). (4) Fig. 1. Solving the Problem in MS Excel The obtained values indicate that, using the mentioned cloud services for a restricted period of time, we will create one presentation, 6 video advertisements, and 3 audio advertisements, in which case we will gain profit of 263 monetary units.
Applied Mechanics and Materials Vol. 379 239 Summary Assessment of an investment project on its own does not ensure full analysis of efficiency when cloud-based services are used or information technologies are introduced. Modern software packages for financial analysis or business planning do not cover all specific features of IT projects. That is why, when efficiency must be assessed, we can consider them only as the first step of the cost and profit evaluation for IT implementation projects. Russian packages Project Expert, Alt- Invest, TEO-Invest, etc. do not support automated calculation of nonrecurring and operating costs for economical and financial assessment of projects. All these facts dictate the necessity of further research works on development and refinement of methods for assessing economical efficiency of investment IT projects (cloud-based services) and creation of an information system which would allow to evaluate information technologies in particular and/or cloud-based services and address all disadvantages of the mentioned modeling systems. Namely, apart from the basic economic rates, the model and methods must feature the following: a) employ optimization algorithms; b) take impact of specific risks into account; c) provide graphic and analytic tools for comparing various IT projects; d) support reviews by experts (using heuristic methods), evaluate their competence and compare the results; e) use dynamic programming for evaluating allocation of investments and efficiency of cloud services. References [1] V.D. Kalachanov, L.I. Kobko. Economic Efficiency of Introduction of Information Technologies. Moscow: MAI Publishing House, 2006. p. 177. [2] V.I. Reyzlin. Numeric Optimization Methods: Tutorial. Tomsk: Publishing House of Tomsk Polytechnic University, 2011. p 105. [3] T. Valentinova. What Cloud Services Really Are. Wardwareportal.ru, 03/09/2009. http://www.hwp.ru/articles/chto_v_deystvitelnosti_predstavlyayut_soboy_oblachnie_servisi/ [4] A. Moskalenko. Cloud and Mobility. What Can Save Russia s IT Market, InLine Group, 01/24/2013. http://www.inlinegroup.ru/events/press-releases/5635.php [5] Delivering Cloud-based Services in a Bring-Your-Own Environment. IT Best Practices. Cloud Computing and IT Consumerization, June 2012. http://www.intel.com/content/ dam/www/public/us/en/documents/best-practices/delivering-cloud-based-services-in-a-bringyour-own-environment.pdf.
Innovation Materials and Manufacturing Technologies, Economic Aspects in Enterprises 10.4028/www.scientific.net/AMM.379 Assessing Efficiency of Cloud-Based Services by the Method of Linear Programming 10.4028/www.scientific.net/AMM.379.235