Optimized cloud instance management strategies using Amazon Ec2.
Cloud computing is an internet based computing that provides shared processing resources and data to computers and other devices on demand. It provides on-demand access to shared pool of computing resources like networks, servers, storage, applications and service which can be made available with minimal management effort. Cloud Computing facilitates users and enterprises with various capabilities to store and process their data in third-party data centres. The growth of cloud computing is due to the its high-capacity networks, service-oriented architecture, autonomic and utility computing. The advancement of cloud computing is due to its features like high performance, scalability, accessibility, high computing power and cheap cost of service.
The main technology of cloud computing is virtualization. Virtualization software separates a physical computing device into one or more virtual devices. Each of these virtual devices can easily be used and managed to perform computing tasks. Virtualization helps to speed up IT operations and reduces the cost by increasing infrastructure utilization. Cloud computing adopts Service Oriented Architecture that helps the user to break the problems into services that can be integrated to provide a solution. Cloud computing provides all of its resources as services. Cloud computing make use of well-established standards and practices gained in the Service Oriented Architecture which made the cloud services global and easy access.
The aim behind this survey is that how effectively system can optimize cloud instance management so that both the cloud providers and clients get benefits. The benefit of clients will be in service and cost level and the main benefit of service providers will be in saving wastage of resources. For this system called Automated Cloud Instance Management System. This system can automate the start and stop instances at the specified indicated times, if required at a click system can start or stop an instance at any time. The images of each instance called AMIs can be created at a single click. The custom reports can be generated based on the run time of each instances. The Run log of each instance can be stored to monitor the cloud service provider service. Billing calculation reports for each instance so that there is no loss per month for us. Each week running hour calculations reports makes our clients feel better about our company. Automated Cloud Instance Management System contains a Cost and utilization optimization mechanism for set of running Amazon EC2 instances. This system receives information on the currently used set of instances (their number, type, utilization) and proposes a new set of instances for serving the same load that minimizes cost and maximizes utilization and performance efficiency.
A. Amazon Ec2:
Amazon EC2 is an Infrastructure-as-a-Service (IaaS) that widely opens Amazon's large computing infrastructure to its users. The Amazon EC2 service is very elastic and flexible to users. This means Amazon EC2 extend or shrink its infrastructure by launching or terminating new virtual machines or so called instances.
Amazon Elastic Computing Cloud (called Amazon EC2) is one of the most important web services, providing resizable compute capability as and when required. The basic unit of Amazon EC2 is "instance". Instance represents a virtual resource with specific storage, network and computational characteristics. Each instance will be having an operating system and located physically in one of the Amazon's datacentres across the world.
The user can use any of the five instance types currently available on offer, the characteristics of which are summarized in Table II.
As Amazon EC2 does not provide job execution or resource management services, a cloud resource management system can act as middleware between user and Amazon EC2. This can reduce resource management complexity, avoids resource wastage and optimizes cost of resource utilization.
Amazon EC2 guarantees a service level agreement in which the client is compensated if the resource (instance) is not available for acquisition at least 99.95% of the time, 365 days/year.
II. System Design:
Motivation behind this project is that how effectively the end user (Customer) can make use of Cloud Computing Services. This project is meant to have a customized service from cloud providers which makes the Customer's life easy.
Cloud service providers like Amazon web services, rack space hosting, windows Azure will provide a lot of services like Computing, Database, Storage and Content, Delivery, Management Tools, Security and Identity, Analytics, Application Services etc. Some of the very common cloud services of Amazon web services are EC2, RDS, S3, IAM, COGNITO. As none of the cloud service providers like Amazon EC2 does not provide job execution or resource management services, a cloud resource management system can act as middleware between user and Amazon EC2
Consider we have a Software Product. We provide services to say, 100 companies in different countries. Obliviously we choose cloud services for this. For each company we make an instance in our cloud space, purchase an appropriate configuration system from cloud. If we require 24*7*356 service in cloud, that may lead to huge cost for companies, which normally won't prefer by most of the companies. So the companies choose an appropriate time every day to get the services, let's say, a time 8.00 am to 8.00 pm every day. If we want to do this manually is a tedious task. We want to enter the cloud service console start each and every instance, bind elastic IP for the instance. At the same time to stop instance we want to stop each instance, disassociate elastic Ips for each instance at time.
The above case is only a single scenario. There are n number of situations where the cloud services, even though it is highly beneficial to end users, is a tedious task to maintain according to the need of customers, for product based companies.
What if we have an Automated Cloud Instance Scheduling Service Management System which controls all the services provided by the cloud service provider? This system can automate the start and stop instances at the specified indicated times, if required at a click we can start or stop an instance at any time. The images of each instance called AMIs can be created at single click. The custom reports can be generated based on the run time of each instances. The Run log of each instance can be stored to monitor the cloud service provider service. Billing calculation reports for each instance so that there is no loss per month for product service provider. Each week running hour calculations reports makes our clients feel better about our company. Creating various Security Groups for cloud service, creating new elastic ip for allocating to new instances, Route53 configurations etc are other added services through this system.
The Proposed System aims at following aspects:
1. To make available the cloud web services in an efficient way so that customer can reduce the cost of usage of cloud service.
2. Ease of usage of required cloud service can be increased so that the customer have no burden in utilizing various services. Self-control over all the actions performing in cloud services which is very important when security issues are considered.
3. Creating customized cloud service reports, which makes the cloud service usage in an effective manner.
4. Automatically create backups and snapshots of your EC2 instances, RDS databases and deletes old backups.
5. With an easy to use interface, managing your actions is simple and effective.
6. System was built from the ground-up for high-availability and resiliency. Your actions will run on-time, that's the guarantee.
7. Start and stop your instances so they run during business-hours only.
8. Decrease your EC2 instance type on weekends or at night when less power is needed.
9. Keeping stale and unused resources around costs you money. AutomatedCloud Instance Management System can delete those unused resources for you.
10. Create Amazon AMIs for the instances manually and automated trigger
A. Initial Setup:
Before the automated cloud instance management system configuration, an Amazon EC2 console account is to be created. After creating a console account, create a user group in our account. An access key and a password gets generated from the Amazon EC2 console. This access key and password is used for the authentication purpose of all API calls to Amazon EC2.If any of the access key or password gets wrong then we won't be able to communicate through Amazon EC2 API calls. The cloud instances for the cloud are created in the Amazon EC2 console account according to the required configuration. For each clients each account is created in Amazon EC2 console. Usually we choose m3.large cloud instance type. For each instances an elastic IP is generated and mapped for the instance.
After the above action, a start instance action and a stop instance action is registered in the automated cloud instance management system. The time to start the instance every day will be specified in start instance action and the time to stop the instance will be specified in stop instance action.
B. Schedulers and Force Start/Stop Instances:
A scheduler is set in the system which checks in every minute whether any of the registered instance action needs to be triggered at that time. If the instance action is a start instance action, then an API call is triggered to start the instance corresponding to the instance id and to associate the corresponding registered elastic ip to the instance. Similarly if the instance action is a stop instance action, then an API call is triggered to stop the instance corresponding to the instance id and to disassociate the elastic ip. Thus the client instance would not be run an extra minute than required by the client.
If a client is required to get start or stop at a particular point of time, there is a provision in the system to make force start/stop. All that is needed to do is to select the instance actions that are required to start/stop, select the action to be performed and do a bulk update action.
The schedulers and force start/stop made it very flexible to manage the cloud instances. The product provider company employees need not login to amazon console to start the instance and bind corresponding public IP to it. Instead of this, enter Automated Cloud Instance Management System and make this process at a click. So managing cloud services made it easier and confortable by using Automated Cloud Instance Management System..
C. Automatic Snapshots and Images:
Automated Cloud Instance Management System can be used to automatically create snapshots and images for you. It can create AMI images of your EC2 instances. It can even examine your EC2 instances and create snapshots of each attached EBS volume automatically.
Automated Cloud Instance Management System has the provision to take AMI of single instances by selecting the cloud instance. The AMI is stored in Amazon cloud. It is a backup of the cloud instance. If in case some weird things happened and database got crashed, then within seconds the instance can be regenerated from the corresponding AMI.
After a product release, usually the product based company took backup of the previous state, that is the state just before the update took place. This is to ensure that the service should not get affected if some weird circumstances happened during the system updates. In order to handle this situation Automated Cloud Instance Management System gives the provision to take bulk AMI backups of cloud instances in a single click. All the user need to do is that to select the region corresponding to the cloud instance. Then the AMI of all cloud instances in that region is automatically called and AMI is generated and stored in Amazon cloud space.
As time goes on, you may want to delete old snapshots and images. System can delete old snapshots and images based on matching criteria and age. It can also ensure that a minimum number of snapshots and images are preserved
D. Billing Calculations:
Amazon EC2 service changes in timely basis. Running your AWS instances 24-hours a day, 7 days a week is not always what you want. Many times, you simply want to have your instances running for only part of the day. They charge for the amount of time the cloud instance runs, and the cloud configuration we selected for cloud instance.The Automated Cloud Instance Management System gets a real picture of number of running time for each instance in weekly basis and the amount charged by amazon. Thus the product service provider can give a clear report to client.
E. Performance Analysis:
By make use of Amazon cloud watch service the performance of each cloud instance can be analysed. The performance analysis and traffic of each instance gets a clear picture for the product service provider to upgrade or degrade the cloud instance configuration. In Automated Cloud Instance Management System there is a provision to upgrade or degrade the cloud instance configuration instantly. At low traffic times, the cloud instance can be degraded to low instance configuration
This work makes available cloud web services in an efficient way so that customer can reduce the cost of usage of cloud service. The Ease of usage of required cloud service is another motive on which we work for so that the customer has no burden in utilizing various services. With Automated Cloud Instance Management System we create customized cloud service reports, which make the cloud service usage in an effective manner. Our system triggers automatic backups and snapshots of your EC2 instances, RDS databases and deletes old backup. System was built from the ground-up for high-availability and resiliency. Your actions will run on-time, that's the guarantee. We focuses on easy to use interface, managing cloud actions is simple and effective. We made it possible to have self-control over all the actions performing in cloud services which are very important when security issues are considered.
[1.] Amazon Web Services--AWS: aws.amazon.com
[2.] Buyya, R., et al., 2009. Cloud computing and emerging IT plat-forms: Vision, hype, and reality for delivering computing as the 5th utility. Futur. Gener. Comput. Syst., 25: 599-616.
[3.] Ward, J.S., A. Barker, 2012. Semantic Based Data Collection for Large Scale Cloud Systems. DIDC, pp: 13-22.
[4.] Kung, H.T., C.-K. Lin, D. Vlah, 2011. CloudSense: Continu-ous fine-grain cloud monitoring with compressive sensing. USENIX HotCloud.
[5.] Shao, J., H. Wei, Q. Wang, H. Mei, 2010. A Runtime Model Based Monitoring Approach for Cloud. IEEE CLOUD, pp: 313-320.
[6.] Simon Ostermann, Alexandru Iosup, Nezih Yigitbasi, Radu Prodan, Thomas Fahringer, and Dick Epema, :An Early Performance Analysis of Cloud Computing Services for Scientific Computing
[7.] Joseph Doyle, Vasileios Giotsas, Mohammad Ashraful Anam and Yiannis Andreopoulos: Cloud Instance Management and Resource Prediction For Computation-as-a-Service Platforms
[8.] Amazon Case Studies: aws.amazon.com/solutions/case-studies/
[9.] Amazon CloudWatch: aws.amazon.com/cloudwatch
[10.] Amazon Elastic Compute Cloud: aws.amazon.com/ec2
[11.] FCincu, M., 2014. Scheduling highly available applications on cloud environments. Futur. Gener. Comput. Syst., 32: 138-153.
[12.] Kokkinos, P., T.A. Varvarigou, A. Kretsis, P. Soumplis, E.A. Varvarigos, : SuMO: Analysis and Optimization of Amazon EC2 Instances
[13.] Ferrer, A.J., et al., 2012. OPTIMIS: A holistic approach to cloud service provisioning. Futur. Gener. Comput. Syst. 28(1): 66-77.
[14.] Niloofar Khanghahi and Reza Ravanmehr, : Cloud Computing Performance Evaluation : Issues And Challenges.
[15.] Marcos Dias de Assungao, Alexandre di Costanzo, Rajkumar Buyya, A cost-benefit analysis of using cloud computing to extend the capacity of clusters
[16.] Gideon Juve, Ewa Deelman, Karan Vahi, Gaurang Mehta: Data Sharing Options for Scientific Workflows on Amazon EC2
[17.] Andrew, J. Younge, Gregor von Laszewski, Lizhe Wang, Sonia Lopez-Alarcon, Warren Carithers: Early Experiences in Cloud Computing for Scientific Applications
[18.] Nuno Santos, Krishna P. Gummadi,Rodrigo Rodrigues: Towards Trusted Cloud Computingss
[19.] Alexandra Iosup, Member, IEEE, Simon Ostermann,Nezih Yigitbasi, Member, IEEE,
[20.] Radu Prodan, Member, IEEE, Thomas Fahringer, Member, IEEE, and Dick Epema, Member, IEEE :
[21.] Performance Analysis of Cloud Computing Services for Many-Tasks Scientific Computing
(1) Febin J, (2) Suma Sira Jacob, (3) Kezi Selva Vijila,
(1) Department Of Computer Science Christian College of Engineering and Technology, Oddanchatram, Dindigul
(2) Department Of Computer Science Christian College of Engineering and Technology, Oddanchatram, Dindigul.
(3) Department Of Electronics and Communication Christian College of Engineering and Technology, Oddanchatram, Dindigul.
Received 18 January 2017; Accepted 22 March 2017; Available online 28 March 2017
Address For Correspondence: Febin J, Department Of Computer Science Christian College of Engineering and Technology, Oddanchatram, Dindigul.
Table I: a selection of cloud service providers. Vm stands for virtual machine, s for storage. Service Examples type VM,S Amazon (EC2 and S3), Mosso (+CloudFS) VM GoGrid, Joyent, infrastructures based on Condor Glide-in/Globus VWS/Eucalyptus S Nirvanix, Akamai, Mozy non-IaaS 3Tera, Google AppEngine, Sun Network Table II: The amazon ec2 instance types. The ecu is the cpu performance unit defined by amazon Name ECUs RAM Archi I/O Disk Cost (Cores) [GB] [bit] Perf. [GB] [$/h] m1.small 1 (1) 1.7 32 Med 160 0.1 m1.large 4 (2) 7.5 64 High 850 0.4 m1.xlarge 8 (4) 15.0 64 High 1690 0.8 c1.medium 5 12! 1.7 32 Med 350 0.2 c1.xlarge 20 (8) 7.0 64 High 1690 0.8 Table III: Amazon Ec2 Instance Types In Numbers 14 main types ml.small', 'ml.medium', 'ml.large', of machines ml.xlarge', 'tl.micro', 'm2.xlarge', m2.2xlarge', 'm2.4xlarge', 'cl.medium', cl.xlarge', 'cc1.4xlarge', 'cc2.8xlarge', cg1.4xlarge', 'hi1.4xlarge' 7 different US East (Northern Virginia), US West regions-- (Oregon), US West (Northern California), datacenters EU (Ireland), Asia Pacific (Singapore), Asia Pacific (Tokyo), South America (Sao Paulo) 2 Operating Windows, Linux Systems 3 RI utilization low, medium, high types 2 RI year terms 1 or three years
|Printer friendly Cite/link Email Feedback|
|Author:||Febin, J.; Jacob, Suma Sira; Vijila, Kezi Selva|
|Publication:||Advances in Natural and Applied Sciences|
|Date:||Mar 1, 2017|
|Previous Article:||The effect of gamma ray irradiation on seed viability and plant growth of Aceh's local rice (Oryza sativa L.).|
|Next Article:||A vibrant comparative analysis of LDA and SVM classification.|