Towards Optimized Fine-Grained Pricing of IaaS Cloud Platform

     Towards Optimized Fine-Grained Pricing of

                        IaaS Cloud Platform

Abstract:

In Cloud many pricing schemes in IaaS platform are already proposed with pay-as-you-go and subscription/spot market policy to guarantee service level agreement, it is still inevitable to suffer from wasteful payment because of coarse grained pricing scheme. In this paper, we investigate an optimized fine-grained and fair pricing scheme. Two tough issues are addressed: (1) the profits of resource providers and customers often contradict mutually; (2) VM-maintenance overhead like startup cost is often too huge to be neglected. Not only can we derive an optimal price in the acceptable price range that satisfies both customers and providers simultaneously, but we also find a best-fit billing cycle to maximize social welfare (i.e., the sum of the cost reductions for all customers and the revenue gained by the provider). We carefully evaluate the proposed optimized fine-grained pricing scheme with two large-scale real-world production traces (one from Grid Workload Archive and the other from Google data center). We compare the new scheme to classic coarse-grained hourly pricing scheme in experiments and find that customers and providers can both benefit from our new approach. The maximum social welfare can be increased up to 72:98% and 48:15% with respect to DAS-2 trace and Google trace respectively.

Architecture Diagram:

              

Existing System:

In general, there are two serious issues in deploying and provisioning virtual machine (VM) instances  over IaaS environment, refined resource allocation  and precise pricing for resource renting . Refined resource allocation is usually implemented by deploying VM instances and customizing their resources on demand, which impacts the performance of VMs to complete customers’ workload. Precise pricing is also known as Pay-as-you-go, which involves multiple

types of resources like CPU, memory, and I/O devices . Pricing is a critical component of the

cloud computing because it directly affects providers’ revenue and customers’ budget .

Proposed System:

An appropriate pricing scheme which can make both providers and customers satisfied is becoming a major concern in IaaS environment. In Amazon EC2, for example, the smallest pricing time unit of an on-demand instance is one hour . Such a coarse-grained hourly pricing is likely to be economically inefficient for short-job users. For instance, users have to pay for full hour cost even their jobs just consumed the resources with a small portion (such as 15 minutes) of the one-hour period. Such a phenomenon is called partial usage waste, which appears very often as cloud jobs are quite short in general . Based on the recent characterization of Cloud environment versus Grid systems, cloud jobs are usually much shorter (such as dozens of minutes) than Grid jobs (such as dozens of hours or days). This will induce serious partial usage waste issu. The current hourly pricing scheme probably induce idle charged resources especially for short jobs, while the fine-grained pricing scheme not only makes users pay less but also makes provider gain more due to the optimization of unit price for the same service time and more users served.

Implementation  Modules:

1.     Resource pricing scheme

2.     Fine Grained Model

3.     Coarse Grained Model

4.     Cloud Time Allocating

Resource pricing scheme:

The existing classic IaaS cloud pricing schemes, and then analyze the partial usage waste issue, and finally formulate our optimized fine-grained pricing model by taking VM maintenance overhead into consideration. Our main objective is to satisfy both customers and providers, and reach maximum social welfare meanwhile. In recent times, the pricing schemes broadly adopted in IaaS cloud market  can be categorized into three types: pay-as-you-go offer, subscription option and spot market. Under the pay as- you-go scheme, users pay a fixed rate for cloud resource consumption per billing cycle (e.g., an hour) with no commitment. On-Demand Instances are often used to run short-jobs or handle periodic traffic spikes. In the subscription scheme, users need to pay an upfront fee to reserve resources for a certain period of time (e.g., a month) and in turn receive a significant price discount. The billing cycles in the subscription scheme are relatively long compared to the pay-as you- go scheme, and can be one day, one month, or even several years . Therefore, it is suitable for long-running jobs (like scientific computing.

Fine Grained Model:

We propose a novel optimized fine grained pricing scheme to solve the above issues. The objective of our work is two-fold, with regard to the classic coarse-grained pricing scheme and inevitable VM-maintenance overhead. On one hand, we aim to derive an acceptable pricing range for both customers and providers, and also derive an optimal price that satisfies both sides with maximized total utility. On the other hand, we hope to find a best-fit billing cycle to maximize the social welfare related to both sides. There are three key terms in our fine-grained pricing scheme: resource bundle, time granularity and unit price. The resource bundle serves as a kind of container to execute task workloads based on user demands. The time granularity is defined as the minimum length in pricing the rented resources. The unit price specifies how much the user needs to pay per time granularity for the resource consumption. We give an example to illustrate the above terms. In Amazon EC2’s pay-as a small on-demand VM instance (Unix/Linux usage). In this example, the VM instance which bundles CPU, RAM, data storage and bandwidth together is the resource bundle. The time granularity is one hour and the unit price is equal to $0:044. As another example, in Cloud Sigma, the resource bundle is not an instance but just some type of resource like CPU or RAM. Cloud Sigma does not bundle them together but allows customers to finely tune the combination of resources on demand exactly. Our work focuses on the time granularity and the unit price, aiming to implement an optimized fine-grained pricing scheme with regard to VM-maintenance overhead like VM boot-up cost. For simplicity reasons, the cloud resource bundle mentioned in our paper is referred to as VM instance similar to EC2 on-demand instance.

Coarse Grained Model:

The objective of our work is two-fold, with regard to the classic coarse-grained pricing scheme and inevitable VM-maintenance overhead. On one hand, we aim to derive an acceptable pricing range for both customers and providers, and also derive an optimal price that satisfies both sides with maximized total utility. On the other hand, we hope to find a best-fit billing cycle to maximize the social welfare related to both sides. There are three key terms in our fine-grained pricing scheme: resource bundle, time granularity and unit  price. The resource bundle serves as a kind of container to execute task workloads based on user demands. The time granularity is defined as the minimum length in pricing the rented resources. The unit price specifies how much the user needs to pay per time granularity for the resource consumption. We give an example to illustrate the above terms. In Amazon EC2’s pay-as you-go option, users need to pay $0:0441 per hour for   a small on-demand VM instance (Unix/Linux usage). In this example, the VM instance which bundles CPU, RAM, data storage and bandwidth together is the resource bundle. The time granularity is one hour and  the unit price is equal to $0:044. As another example, in Cloud Sigma , the resource bundle is not an instance but just some type of resource like CPU or RAM. Cloud Sigma does not bundle them together but allows customers to finely tune the combination of resources on demand exactly. Our work focuses on the time granularity and the unit price, aiming to implement an optimized fine-grained pricing scheme with regard to VM-maintenance overhead like VM boot-up cost.

Cloud Time Allocating:

The designed optimized fine-grained pricing scheme should also satisfy providers, yet providers may suffer higher overhead due to finer pricing rates. In the classic coarse-grained pricing scheme, the provider will suffer To to manage VM instance every one hour service. In the optimized fine-grained pricing scheme, the provider may suffer higher overhead due to more frequent context switch among VM instances. We use an example to highlight the difference between the two pricing schemes in Fig. 3. Obviously, the provider will suffer higher loss of payment in the second pricing scheme with finer granularity due to more frequent VM overheads appearing in the whole service time.

                          

Algorithm:

                    

             

Configuration:-

H/W System Configuration:-

         Processor                     -    Pentium –III

             Speed                                -    1.1 Ghz

RAM                                -    256  MB(min)

Hard Disk                         -   20 GB

Floppy Drive                    -    1.44 MB

Key Board                        -    Standard Windows Keyboard

Mouse                               -    Two or Three Button Mouse

Monitor                             -    SVGA

 

 S/W System Configuration:-

v   Operating System            :Windows95/98/2000/XP

v   Application  Server          :   Tomcat5.0/6.X                                             

v   Front End                         :   HTML, Java, Jsp

v    Scripts                             :   JavaScript.

v   Server side Script             :   Java Server Pages.

v   Database                           :   Mysql 5.0

v   Database Connectivity     :   JDBC.