Information technology sector could be able to experience huge economic benefits and flexibility in resources because of cloud computing. Today many of organizations use cloud computing and moved their storage and processing to it. Since cloud services are popular these days and many organizations have been shifted to cloud computing it is important to evaluate cloud performance and identify its issues. Cloud computing services are provided by third party provider who is the owner of the infrastructure. Users can send, receive as well store the data on the internet through cloud computing. So users can experience the parallel computing by using any number of virtual machines. Even though there are several advantages in cloud computing we can identify several performance issues in cloud computing.
Keywords: Cloud service, Performance
Cloud computing is a very interesting concept and many companies and organizations are moving their applications and utilities in to data centers. So through internet it can be accessed from any place. John Mac Carty is the first person who suggested that one day computers can be used as a public utility. 1 2 3 4 5 Public, private, community and hybrid are four deployment cloud models. 5Platform as a service (PaaS), Infrastructure as a service (IaaS) and Software as a service (SaaS) are three types of cloud services models. We can identify so many factors that can affect cloud computing performance and resources due to the expantion of the cloud computing environment and increase the number of users who are using cloud environment. The focus of the paper is to explain the performance issues and challenges in cloud computing. Other sections of this paper are organized as follows. In section 2 I will discuss about the background details. The discussion is included in section 3. Aim and objectives, scope and intended outcome are described respectively.
The “Intergalactic computer network” concept firstly suggested in 1960’s and cloud computing has developed since then. IBM introduced “Virtual machine” which makes virtualization in computing as well as communication in 1970’s. Telecommunication organizations could able to offer network connections which were virtualized with lower cost and a great quality because internet became much common in 1990’s.Cloud computing term was expressed 6 as a new computing paradigm in late 1990’s. Amazon was a major contributor who developed cloud computing from “grid” and “distributed” computing to “utility” computing at the time they introduced Amazon web service. Amazon discovered they can experience their facilities in efficient manner by using this new cloud computing concept. Eric Schmidt, CEO of Google developed Google Docs to give opportunity to its users to experience the cloud computing. The co-founder of Apple, Steve Jobs launched “Digital Hub” MacWorld 2001.Today cloud computing plays a major role in our digital lifestyle.
Many sectors are experiencing the benefits of cloud computing and some of them are explained in below.
• Reliability – companies which provide the cloud computing services are reliable such as Google, Amazon and Microsoft.
• Scalability – users have to pay what they use and no need for infrastructure.
• Stability – Now cloud computing offers a stable service.
• Backup – companies which provides cloud computing service gives automatic backup service to users. So they no need to suffer from data losing which stored in cloud.
• Effort and time saving – No need to spend lot of time in searching storage solutions and retrieving the data from external storage devices as well as writing data to storage.
• Security – One person’s data cannot be access or read by another person because it requires password, id etc. in order to access the cloud.
• Accessibility – Mobile devices like tablets and smart phones can access the data which is stored in cloud.
• Saves money – Since users can store any type of data in cloud they no need to purchase storage devices.
Some of cloud computing applications are described as follows.
• Universities use cloud computing concept to course delivery, collaboration and better usage of software as well as hardware.
• Multimedia organizations supply data to customers using “cloud based content distribution networks”.
• Many people use cloud computing for store data like videos and music, messaging and social networking.
• Business organizations also use cloud computing in order to reduce the cost, manage workforce and experience the higher flexibility.
Cloud computing service delivery models and performance
Three types of delivery models are described as follows.
• Software as a service – users can use application from the cloud without installing software on that person’s computer or machine. And also user no need to worry about handling the storage, servers as well as network. One of the best examples for this model is Gmail. 1 2 3 4
SaaS performance is understood by users while it defined as application scalability, throughput and response time of transactions.
• Platform as a service – This model gives a platform to make application for developers. Users no need to handle operating systems, servers and network. Google App Engine is an example for this model. 1 2
PaaS performance is not directly identified by users and explained as middleware scalability, throughput and response time of transactions.
• Infrastructure as a service – users have to manage software and operating systems but no need to worry about managing the cloud infrastructure. Since this is connected to a virtual engine 7people who use this type can use a virtual machine to access. 8 Example for this model is “Amazon Web Services’ Elastic Computer Cloud”.
LaaS performance is explained as infrastructure performance, scalability, capacity, reliability.
Performance determinants characteristics of upper service layers are depend on underlying layers characteristics. For an example scalability of SaaS depend on laaS scalability.
Cloud Computing Deployment Models
Deployment models which described in here are based on user requirement.
• Private cloud – This type is maintained by the private companies but Third party also can do the management. This type can be hosted externally or internally in the organization. Eucalyptus is an example for the private cloud. 9
• Community cloud – share the infrastructure with different companies which have common goals. These types of clouds are based on agreement between companies or organizations. Because of the shared cloud, can experience economic benefits. Facebook is an example.
• Public cloud – Service provider gives services to users with more economic benefits. Google App Engine is one example for this public cloud.
• Hybrid cloud – This model is made up from using two or more deployment models like community, public or private. Possible to manage resources internally or externally. 1
Performance of Cloud Computing
Cloud services performance was exposed as it uses more in information and technology sector. There can be different response time. Response time can be less or large. If the user makes many requests the performance of cloud computing may reduce. Performance issues can encounter because of huge distribution of applications which are cloud based. There are several ways we can increase the cloud performance. We can enhance the cloud performance by encrypting part of files. So it provides ability to parallel uploading. It increases the processing speed as well as transfer speed. By grouping the data in to several large portions and distributing those to different clouds, we can improve the cloud performance. So all the data which are in cloud will not be stole or affected even a hacker attack. Just only few data parts may be affected. Cloud performance may also depend on the downloading as well as uploading time. Group the data in to several parts and then encrypt those parts means the uploading time. Decrypt the data sectors and make them as a one file means the download time. Both download tasks and upload tasks happens same time. We can increase the performance of cloud by rising the caching of web pages that are frequently use in clouds.
There are several ways we can test the cloud performance. Stress test over cloud, latency test on cloud, capacity test on cloud, browser performance test on cloud are some examples. Performance test in cloud is different from other applications. The goal of performance test in cloud is to measure the throughput, latency and other factors.
Factors affective on performance
Today the “performance” cannot be considered as just a classical concept but also it consist many more concepts like scalability, efficiency, reliability and etc. There are several factors which affect to the performance of cloud computing due to the increment of cloud service users and expand of cloud environment. Some of them are expressed below.
• Storage capacity – This is very important factor when evaluate the performance and physical memory may lead to performance increment or decrement. 10
• Buffer capacity – Requests which the server cannot fulfill are stored in a temporary memory location. So it may cause to reject requests if there is a lower buffer capacity. Because of that performance can be affected by buffer capacity.
• Disk capacity – cloud performance may affected by the size of disk. 10
• Amount of users – Large number of users in data centers may cause to reduce the cloud performance.
• Availability – cloud performance may result to increase if the services are available at any time and have efficient access to cloud.
• Network bandwidth – Performance may result to reduce if bandwidth is low. 11
• Location – Distant between user and the data center effective on the cloud performance.
• Recovery – The data in the clouds may be lost due to different reasons. So the time needed to retrieve the data as well as amount of data which need to recover may affect the performance of cloud.
• Service level agreement- SlA can act as a major contributor in performance increment. Because it concerns about larger area of performance and availability. SLA makes IT managers to think about three main areas. They are cost, continuity and data protection.
Security, workload, scalability, processor power, usability, latency, Fault tolerance are some other factors which can be affective on cloud computing performance.
Performance issues of cloud computing
Issues in Cloud computing performance are correlated with network, security methods, system, communication channel and service implementation.
We can identify many problems which can be effective on performance of cloud computing. Cloud computing have several performance issues because of network sharing and resources sharing. The performance may reduce if there is just only one physical swarm and all closest virtual machines are accessing to it and deactivating as well as activating at the same time. And also there may have huge load on cloud assets. Parallelized computing is one factor which caused to reduce the cloud computing performance. Parallelized computing adds much more overhead on performance. So we should not always do it. If the cloud computing network user sends more requests cloud has to bear much more burden. So it causes to reduce the cloud performance. 12 Because this model has just only one input and one output. And this causes to slower down the process as well as reduces the performance. User must check whether the system is capable to run the requested application. The cloud computing performance will decrease if the system is also not capable. 13
Some of popular organizations such as Amazon, Google and Microsoft suffered from cloud computing failures in the past. Following example explain performance issues in cloud services.
• Amazon Elastic Computer Cloud (EC2) and Microsoft cloud Disaster
Amazon Elastic Cloud Computing had to face cloud disasters in the past and there were many downtimes as well. Because of this downtime in EC2 many of SMEs suffered a lot. Different types of information technology organizations complained about their losses due to the cloud services issues and many of them became idle. There was a problem with Amazon and Microsoft because of bad weather condition and at the end it became as a cloud failure. Servers also affected because of thunder lighting. So due to this issue small organizations as well as large organizations had to face a huge loss. In that time servers didn’t work properly and couldn’t access for two days.
So many events like power failures, gradation in software and huge load on server may result to cloud service failures. Recovery time depend on the number of servers which are down in data center. It is must to check whether the cloud services are properly functioning after the recovery. Service level agreement is an important factor in that case. The losses which faced because of cloud services problems have to settle as the conditions expressed in SLA.
Solutions for enhance the performance
There are several ways that can increase the cloud computing performance. One way is parallel processing. Many users can access the data which are in cloud simultaneously from parallel processing. By using hardware that have huge configuration can increase the cloud performance. More users may be able to access to the data at same time if increase the memory size. It also helps to increase the performance. By managing the resources properly we can increase the cloud computing performance. Many users can access the data at the same time if we use multiple input output model. User must use firewall in their network to protect from hackers’ attacks. So the users system may secure. 9 We must ensure the compatibility problems in operating systems in order to increase the performance.
The main intention of this paper is to discuss about cloud computing and explain performance issues in cloud computing. Cloud computing can be described as relationship between computers through the internet. History of cloud services, cloud service models, benefits and applications of cloud computing were discussed in paper. However the main focus of this paper is to discuss the performance of clouds and issues related to cloud performance. So I explained about the performance of cloud computing, factors affective for the performance, performance issues and some ways to enhance the performance. Performance can be considered as a major attribute. The performance of cloud depends on the transfer time, downloading time, uploading time processing speed and etc. several performance issues can identify in cloud since it is a shared system.
1 Abah Joshua & Francisca N.Ogwueleka, “Cloud Computing with Related Enabling Technologies,” International Journal of Cloud Computing and Services Science (IJ-CLOSER), vol. 2, no. 1, pp. 40-49, 2013.
2 Borko Furht & Armando Escalante, Handbook of Cloud Computing, 2010.
3 NIST Advisory Working Group, “NIST Cloud Computing Standards Roadmap,” NIST Special Publication 500?291, 2011.
4 M.Malathi, “Cloud Computing Concepts,” IEEE, 2011.
5 Peter Mell & Timothy Grance, “The NIST Definition of Cloud Computing,” NIST Special Publication 800-145, 2011.
6 R.K.Chellappa, “Intermediaries in Cloud computing:A new computing paradigm,” Dallas, october26-29, 1997.
7 B.R.Kandukuri,R.Paturi V,A.Rakshit, “Cloud security issues,” in IEEE international Conference on services computing, 2009.
8 Malte Schwarzkopf Derk G.Murray Steven, “The seven deadly sins of cloud computing,” University of Cambridge Laboratory.
9 Ashraf zia,Muhammad naeem khan Abdul, “Identify key challenges in performance issues in cloud computing,” Wali khan University KPK Pakistan.
10 Ioannis A. Moschakis and Helen D. Karatza, “Performance and cost evaluation of gang scheduling in a cloud computing system with job migration and starvation handling,” IEEE, 2011.
11 K. Xiong, “Service Performance and analysis in cloud computing,” IEEE, 2009.
12 Simon Ostermann and Alexandru Iosup, “A Performance analysis of EC2 cloud computing services for scientific computing,” Institute for computer science,social informatics and telecommunications Engineering(LNICST), 2010.
13 Keith R.Jackson and Krishna Muriki, “Performance analysis of high performance computing applications on the Amazon Web services cloud,” IEEE International conference on cloud computing technology and science, 2010.