Pricing Models and Pricing Schemes of IaaS Providers: A Comparison Study Mohan Murthy M K Sanjay H A Ashwini J P Global Services Department of. Department of Curam Software International, Information Science & Engg Information Science & Engg An IBM Company Nitte Meenakshi Nitte Meenakshi Bangalore, India Institute of Technology Institute of Technology maakem@gmail.com Bangalore, India Bangalore, India sanju.smg@gmail.com ashwini.janagal@gmail.com ABSTRACT Cloud Computing is one of the most significant advances in information technology since the rise of the global Internet network. The cost effectiveness of the cloud computing is one of the reason for its popularity. The cost can be further reduced by the customer by carefully selecting the cloud service provider, according to his/her requirement. This work is an attempt to list the different pricing models and pricing schemes used by some of the popular IaaS providers and also to understand how the varies for the same requirement in different IaaS providers. This survey helps the customers to understand the different pricing models and to choose the better IaaS provider depending on his/her requirement. Categories and Subject Descriptors C.2.4 [Computer-Communication Networks]: Distributed Systems - Distributed applications. General Terms Management, Measurement, Economics Keywords Cloud Computing, IaaS, Pricing Models, Storage Space 1. INTRODUCTION Cloud computing model is becoming popular these days. Cloud computing provides computation, software, data access, and storage services that do not require end-user knowledge of the physical location and configuration of the system that delivers the services [1]. As Clouds are designed to provide services to external users, providers need to be compensated for sharing their resources and capabilities [2]. Permission to make digital or hard copies of part or all of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. ICACCI '12, August 03-05 2012, CHENNAI, India Copyright 2012 ACM 978-1-4503-1196-0/12/08 $10.00. Based on the service provided, in cloud computing three service models are defined Software as a Service (SaaS) SaaS sometimes referred to as "on-demand software," is a software delivery model in which software and its associated data are hosted centrally and are typically accessed by users using a thin client, normally using a web browser over the Internet [3]. The software services are provisioned on a pay-as-you-go basis to overcome the limitation of the traditional software sales model [4]. Ex - Salesforce. Infrastructure as a Service (IaaS) Infrastructure as a Service (IaaS)is the delivery of computer hardware (servers, network, storage, and data center space) as a service [5]. Ex Amazon, RackSpace. Platform as a Service (PaaS) With Platform as a Service (PaaS), the provider delivers more than infrastructure. PaaS offers an integrated environment to design, develop, test, deploy and support custom applications [5]. Ex Windows Azure, Google App Engine. Cloud computing model has several advantages. Cost effectiveness is one of them. In this model user will pay for what he/she has used. In cloud computing world Infrastructure as a Service (IaaS) is the delivery of computer hardware (servers, network, storage, and data center space) as a service. The IaaS customer rents computing resources instead of buying and installing them in their own data center. The service is typically paid for on a usage basis. The service may include dynamic scaling so that if the customer winds up needing more resources than expected, he can get them immediately (probably up to a given limit) [5]. IaaS providers use different pricing schemes for different requirements. For example in case of Amazon EC2 three types of purchasing options are available. They are On-Demand Instances - On-Demand Instances let users to pay for computing capacity as per their usage with no long-term commitments [6]. Reserved Instances - Reserved Instances give user the option to make a low, one-time payment for each instance user wants to reserve and in turn receive a significant discount on the hourly usage charge for that instance [6].
Spot Instances - Spot Instances enable user to bid for unused Amazon EC2 capacity. Instances are charged the Spot Price, which is set by Amazon EC2 and fluctuates periodically depending on the supply of and demand for Spot Instance capacity [6]. We can observe that pricing scheme is different for these options. The will be different from one IaaS provider to the other for the same requirement and also IaaS providers use variety of pricing models. By carefully selecting the IaaS provider user can reduce the cost. Price Initial fee time Normal Pricing Reduced pricing with initial fee The aim of the paper is to list the different pricing models used by the IaaS providers and also to compare the pricing of different IaaS providers for the same user requirement. Here we mainly focus on the storage space and computational resources (CPU, RAM and Hard disk space) provided by the IaaS providers. In this work we attempted to identify different pricing models used by the IaaS providers by analyzing their pricing and also we are suggesting cloud users to select the best IaaS provider depending on their requirement. We consider some popular IaaS providers like Amazon, RackSpace, Nirvanix, CloudSigma when comparing the s and identifying the pricing models. Rest of the paper is organized as follows, in section 2 we will list some popular pricing models used by the IaaS providers. Section 3 compares the pricing of the IaaS providers for storage space requirement. In section 4 we will compare the pricing of IaaS providers for the computational resources (CPU, RAM and Hard disk space). Figure 2.2 Price Model Variation in Linear. In one more variation of the Linear model, instead of collecting the subscription charge in the beginning, if the resource utilization time crosses some predefined time duration a discount in the total billed amount can be offered to the user. This model is observed in Cloud Sigma [10]. 2.3 Step Model In the step model per unit will not be a constant. As the resource utilization increases, it decreases in a step manner as shown in Fig 2.3. For example till the resource utilization I 1 the unit will be p 1, once the utilization crosses I 1 then the unit will become p 2 as shown in the Fig. 2.3. This model can be observed in Amazon S3 [9]. In this model resource utilization vs total will be as shown in the Fig. 2.4. 2. PRICING MODELS There are many pricing models available. Some of the interesting pricing models we came across by going through different IaaS providers pricing schemes are listed below. 2.1 Linear Model In Linear model the is directly proportional to the resource consumed. As the resource utilization increases the also increases linearly as shown in the Fig. 2.1. The per unit resource utilization will be a constant. This model is observed in Nirvanix [7] and RackSpace [8]. Unit p1 I1 p2 I2 Resource Figure 2.3 Price Model Step. resource Figure 2.1 Price Model Linear. 2.2 Variations of Linear Model Slight variation of linear model is observed in Amazon EC2 [6], where if the user is interested in long term utilization of a resource then initially a subscription charge will be collected from the user, later the hourly usage charge will be reduced as shown in the Fig 2.2. resource Figure 2.4 Resource Utilization vs Total
3. COMPARISON OF IAAS PROVIDERS PRICING FOR STORAGE SPACE User can purchase storage space from IaaS providers. Cloud user should be careful in selecting IaaS provider because the pricing will be different in different IaaS provider for the same requirement. The table 3.1 tabulates pricing, for storage space which will be provided by popular IaaS providers (Amazon, RackSpace and Nirvanix). From the tabulated data we can conclude that if the user s storage space requirement is less than or equal to 1 TB, RackSpace is cheaper when compare to the other IaaS providers. If the user requirement is more than 1 TB then Amazon pricing beats all other vendors. Fig. 3.2 shows the effective per TB vs storage space capacity. In this graph we can observe that up to 50 TB effective per TB for Amazon and RackSpace are almost same. We can also observe that effective per TB is always same in RackSpace and Nirvanix but in case of the Amazon as the storage space requirement increases the effective / TB decreases. Table 3.1 Price Comparison for Storage Space. Amazon Size (In Tb) Price Data Transfer In - Price Effective Price Per TB 1 143.36 102.4 245.76 25 3215.36 2560 231.0144 50 6415.36 5120 230.7072 100 12047.36 10240 222.8736 5001 281656.32 512102.4 158.72 Rackspace 1 153.6 81.92 235.52 25 3840 2048 235.52 50 7680 4096 235.52 100 15360 8192 235.52 5001 768153.6 409681.92 235.52 Nirvanix 1 256 102.4 358.4 25 6400 2560 358.4 50 12800 5120 358.4 100 25600 10240 358.4 5001 1280256 512102.4 358.4 - All the s in US Dollars, - Prices calculated per month Figure 3.2 Effective Price/TB for Different IaaS Providers for Storage Space 4. COMPARISON OF IAAS PROVIDERS PRICING FOR COMPUTATIONAL RESOURCES We have analyzed different IaaS providers with respect to computational resources (CPU, RAM & Hard disk space). Here we are summarizing the pricing of some popular IaaS providers (Amazon EC2 and CloudSigma). 4.1 Amazon EC2 Amazon provides different virtual machine instance types. In this paper we are considering standard instance types i.e. small, large and extra-large [11]. The specifications of these instances are given in the table 4.1. Instance type Table 4.1 Specifications of Instance Types CPU in EC2 Compute Units* RAM in GB STORAGE in GB Small 1 1.7 160 Large 4 7.5 850 Extra Large 8 15 1690 * One EC2 Compute Unit provides the equivalent CPU capacity of a 1.0-1.2 GHz 2007 Opteron or 2007 Xeon processor [11] As we mentioned earlier Amazon provides on-demand, reserved and spot purchasing options. The pricing scheme of the ondemand and reserved instances, are given in the table 4.2 and table 4.3 respectively. Since the spot instances will be varying based on the supply of and demand for the spot instance capacity we are not considering them here. The pricing also depends on the region from where user has selected the instances. Here we have given the pricing of the instances for the region US East (Virginia).
Table 4.2 Pricing Details of On Demand Instance Types. Instance type Linux/Unix usage Windows usage Small $*0.085 per hour $0.12 per hour Large $0.32 per hour $0.48 per hour Extra Large $0.64 per hour $0.96 per hour * US Dollars Table 4.3 Pricing Details of Reserved Instance Types. Instance Type Initial Fee 1 Year Term 3 Years Term Linux/Unix Usage Per Hour Price Windows Usage Per Hour Price Small $227.5 $350 $0.03 $0.05 Large $910 $1400 $0.12 $0.20 Extra Large $1820 $2800 $0.24 $0.40 As the resource utilization duration increases the total will be less in case of the reserved instances as shown in the Fig. 4.4. So if the customer has the requirement for long term usage of the resource then he/she can select reserved instance type. For short term usage he/she can select the on demand instance. Figure 4.5 Price Comparison of Unix and Linux On Demand Standard Small Instance Type 4.2 Cloud Sigma In CloudSigma it is possible to combine CPU, RAM and storage to form a server in any combination within the minimum and maximum overall resource sizes. There is no real concept of specific server sizes on the CloudSigma platform [10]. The range of the computational resources and pricing details are given in the table 4.6. This pricing is excluding the Operating System license charges. In CloudSigma a separate license fees should be paid for operating system on monthly basis. The is 40.5 USD/month for Microsoft Windows Server 2003 Standard (32-bit or 64-bit), Microsoft Windows Server 2008 Standard (32-bit) and Microsoft Windows Server 2008 R2 Standard (64-bit) [10]. Table 4.6 Range and Pricing Details of Computational Resources in CloudSigma Computing resources Minimum Range Maximum Unit Price per unit in USD CPU 0.5 20 GHz 0.0225 RAM 0.5 32 GB 0.0293 Storage 1 1024 GB 0.1800 Figure 4.4 Price Comparison of Reserved and On Demand Standard Small Instance Type We can observe in the tabulated data (table 4.2 and 4.3) that the for linux/ unix is less when compare to windows. If the customer s applications are compatible with linux/ unix operating system then he/she can go for the instance which is having the linux/ unix operating system. Price comparison for unix and windows VMs of on demand small instance type is shown in Fig. 4.5. CloudSigma offers discount on the total billed amount. The discount scheme is given in the table 4.7. Table 4.7 Discount Scheme in CloudSigma % Discount Duration in months 1 3 6 12 24 36 0 3 10 25 35 45 Because of this discount in CloudSigma as the resource utilization duration increases per unit resource utilization decreases as shown in the Fig. 4.8.
Figure 4.8 Price per Unit for Computational Resources vs Utilization Duration in CloudSigma 4.3 Amazon EC2 vs CloudSigma We will take Amazon s EC2 standard small instance type (on demand and reserved) for comparing with CloudSigma s. We are assuming one EC2 compute unit as 1.1 GHz. The table 4.9 gives pricing details of Amazon EC2 and CloudSigma. From the tabulated data we can conclude that for short term usage Amazon EC2 s on demand small instance is better in. But as the resource utilization duration crosses 9 months then Amazon EC2 s reserved instance is better. The advantage with the CloudSigma is the cloud server is configurable in terms of RAM, CPU and Storage space but with the Amazon EC2 it is of fixed size. User has to buy the nearest instance type of the Virtual Machine in Amazon EC2. If the Virtual Machine capacity is more than the application requirement then user will be paying the extra money unnecessarily. If the virtual machine capacity is less, then user s application performance will be degraded. So care must be taken while selecting the Virtual Machine instance type. Table 4.6 Price Comparison of Amazon EC2 and Cloud Sigma Dura tion in mont hs Amazon pricing in USD On dema nd Reserve d 1 yr term Reserv ed 3 yrs term Cloudsig ma pricing in USD 3 183.6 292.3 414.8 240.0261 6 367.2 357.1 479.6 445.4093 12 734.4 486.7 609.2 742.3488 18 1101.6 843.8 738.8 1113.523 24 1468.8 973.4 868.4 1286.738 30 1836 1330.5 998 1608.422 36 2203.2 1460.1 1127.6 1633.167 Fig. 4.11 shows the comparison of Amazon EC2 s standard small instance (on demand and reserved) with the pricing of CloudSigma s computational resources which are equivalent to Amazon EC2 s standard small instance. Figure 4.11 Price per Month for Computational Resources 5. CONCLUSIONS From the analysis of pricing schemes of different vendors we can conclude that varies from vendor to vendor for the same requirement. A vendor will be using different pricing models. User should select the appropriate vendor and the pricing model according to his requirement to reduce the. Even though is an important factor when selecting an IaaS provider there are many other factors like customer support, availability of the service, penalty (if the service level objective is not met) etc. User should consider all these factors before choosing an IaaS provider. In future we will be concentrating more on pricing models and also we are planning to build a software system which selects the IaaS resources as per the application requirements dynamically and also selects better IaaS provider by considering QOS parameters & cost effectiveness. 6. REFERENCES [1] http://en.wikipedia.org/wiki/cloud_computing (retrieved on 4th Sep 2011) [2] R. Buyya, C. S. Yeo, S. Venugopal, J. Broberg, and I. Brandic, Cloud Computing and Emerging IT Platforms: Vision, Hype, and Reality for Delivering Computing as the 5th Utility, Future Generation Computer Systems, 25(6), (pp. 599-616), Elsevier Science, Amsterdam, The Netherlands. [3] http://en.wikipedia.org/wiki/software_as_a_service [4] Linlin Wu, Saurabh Kumar Garg and RajkumarBuyya, SLA-based Resource Allocation for Software as a Service Provider (SaaS) in Cloud Computing Environments, 2011 11th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing [5] Cloud Computing for Dummies - Judith Hurwitz, Robin Bloor, Marcia Kaufman, Fern Halper [6] http://aws.amazon.com/ec2/pricing/ [7] http://www.nirvanix.com/how-to-buy/self-servicepricing.aspx [8] http://www.rackspace.com/cloud/cloud_hosting_products/ files/pricing/ [9] http://aws.amazon.com/s3/ [10] http://www.cloudsigma.com/en/pricing/-schedules [11] http://aws.amazon.com/ec2/instance-types/