Efficiency of Cloud Computing Data Centers - Thesis Example

? Efficiency of Cloud Computing Data Centers Enterprise data centers had provided the required infrastructure for organizations since the integration of technology in the business field. With the passage of time, the customers’ demands started increasing due to which the limited data storage and bandwidth became insufficient for the organizations. This created the need for an alternate form of computing known as cloud computing. It offers almost unlimited, dynamic allocation of resources on the demand of the client. This paper discusses the aspects in which the cloud computing data centers prove to be more efficient than the conventional data centers. It will also highlight the new techniques that are being adopted to overcome the physical issues found in the traditional data centers. 1. Introduction Large organizations have always had the capital to purchase numerous servers, software etc to support their businesses. It is the medium and small sized organizations that have struggled in this regard. Earlier, the only alternate that was available for such companies was to maintain their IT infrastructure in someone else’s premises but now things have changed. It is a thing of the past when great investments were required to launch a business or service on the internet. Cloud computing provides developers with online availability of their computing needs to save them from any heavy initial investments for the IT infrastructure. The service is paid off in the same manner as the electricity grid i.e. the payment equals to the amount of consumption of the services. Cloud computing is a concept that has been desired by developers since a long period of time. Large investments are not required by the developer to launch a service or product since all the computing capability requirements are available online. Cloud computing is known as the availability of application services, software, platforms etc that would provide the basis of any desired service on the web. The computing capability for different applications, software etc that is availed from the internet as a service, is referred to as a ‘cloud’. Armbrust (2009) explained that the private cloud is referred to as to the internal clouds of an organization that are not permitted to be used except the owner of the cloud i.e. the respective organization, whereas, cloud computing is referred to as the ‘public cloud’ since it is available for use by any individual who might need instantaneous computing capability. This innovative concept has even leaped the performance level of data centers to a new level and has reduced costs for independent developers and organizations. 2. Background Information Warr (2009) stated that cloud computing is also referred to as ‘Platform as a Service’ (PaaS) since developers use it as a platform to run their applications or even develop them from scratch without investing in the acquisition of the required operating system or server hardware. Another name that is used for this concept is ‘Infrastructure as a Service’ (IaaS); it includes the dynamic usage of CPUs, virtual servers, network, storage, systems software etc. Software as a service (SaaS) is another type of cloud computing that provides access to software to users on their need basis while the software runs on the vendor’s side of the data centers. The conventional mode of using the software requires the installation of the software on the user’s computer but this latest technology requires no such installation. Since many decades, internet users have been availing the service of uploading their documents and pictures on the internet by different services like Flickr, Snapfish etc; this concept has enabled the people to save their data ‘on the cloud’. The latest concept of cloud computing is far more usable, scalable and efficient for individual developers and organizations. 2.1 Criteria of Effective Performance Rajan (2011) provided some relevant characteristics that can be helpful to evaluate the effectiveness and performance level of an IT infrastructure. He stated that the data centers need to satisfy the following criterion; Functional Requirements, which facilitate the achievement of successful functioning of the business operations. Non-functional requirements, which ensure the quality of service (QoS) being provided to the clients. Quality of service (QoS) is maintained in typical business conditions. Quality of service is maintained in atypical business conditions. Service is available 24/7 Both of these requirements play an important part in the effectiveness and performance level of the service provider. The following sections shall consider only the non-functional aspects of the business operations. Physical conditions in which the data centers are kept are also managed to improve the efficiency in terms of power consumption, appropriate cooling techniques, data center designs etc. These aspects shall also be addressed in detail in the following sections. 2.2 Traditional ‘On-Premise’ Data Centers Before the evolution of cloud computing, organizations relied on conventional form of data centers which were owned and maintained by the organizations themselves. The ownership and thus maintenance of the data centers proved to be a major source of expenditure for the business owners. 2.2.1 Level of Scalability There was no provision of scalability in the traditional enterprise data centers. They constituted of limited capacity and bandwidth. The eventual increase in the size of the data and computing needs required further investments in the infrastructure. Although, the increased number of servers or other requirements might only be useful for a limited period of time but had to be purchased to fulfill the needs of the business, for example, an online shopping website might face increased traffic during the month of December due to the holidays shopping. The increment in servers and data centers would prove to be a source of greater revenues and quality service during the days of heavy traffic but served as an overhead for the organization for the rest of the year. The ‘on-premise’ data centers proved to be very costly due to their nonalignment with the requirement of scalability. The on-premise data centers proved to satisfy the typical business conditions but fell short during the atypical ones unless greater data centers were provided. At times, the sudden rush of requests also crashed the servers due to which the service could be shutdown and interrupted for the clients. Cloud News Desk (2008) described the conventional data centers and explained that organizations had full control over their data but were even liable for all disasters themselves. The infrastructure and data could not be shared with anyone and had limited number of members with its respective access rights. 2.2.2 Maintenance and Update Requirements Enterprise owned IT infrastructure had to be updated with the latest patches and versions to continue providing the latest functions and operations. This often meant greater investments since all new versions are not free of cost. A competent maintenance team was required to manage the systems and deploy effective security measures for its protection. Multiple management tools were adapted by the organization for the smooth functioning of different applications and services. The management and maintenance of such an extensive infrastructure created complex job descriptions for the employees. It is due to the constant maintenance and required updates that the management greatly focused on the performance level of the IT infrastructure rather than on their respective business. 2.2.3 Level of Mobility The conventional data centers did not offer mobility since the data physically resided in the premises of the organization. The data was not available on the web which restricted its usage and accessibility. 3. The Advent of Cloud Computing Cloud computing is the end result of the evolution of concepts of virtualization and utility computing. Cloud computing has provided such computing services that have never been witnessed in the history of technology. It has given more power to the people and has cut down their costs to a great extent. Google, Amazon and Microsoft are known to be the chief providers of cloud computing services since they have given their customers reliable and effective avenues to improve their business. Amazon is considered to be the first provider of a cloud platform- Amazon Elastic Computer Cloud (EC2). According to Miller (2010), it is a web service that allows the user to utilize numerous computing units on need basis. Google and Microsoft have also provided multi-faceted web services for cloud computing namely, Google AppEngine, Windows Azure (Microsoft Corporation, 2011), respectively. The data centers that are coupled with cloud computing services tend to provide better performance than the traditional enterprise data centers. According to a survey by Gartner Inc. (2011), cloud computing is the top ranked technology of the year 2011. The sample of this survey consisted of more than 2,000 CIOs. The aspects where the cloud computing data centers outperform the conventional approach are explained below: 3.1 Level of Scalability The most attractive aspect about cloud computing is that it offers scalability like never before. Cloud computing facilitates the dynamic provision of resources as per the demand of the user. The required resources can range from servers to application platforms, software etc. It is a self-service therefore the resources can be acquired as soon as the need arises (without any delays). An even more attractive aspect of this scalability is the approach that is adapted by the service providers for the payment of cloud computing services- metering. The payment for cloud computing equals the number of resources that have been utilized alongside their consumption time period. This form of billing is often considered analogous to the electricity grid; pay as much as being utilized. In this manner, organizations can acquire greater number of computing power in the times of heavy traffic and release the resources when the need ceases to exist. King (2008) pointed out that specifically, this on-demand increment in the infrastructure volume benefits the medium and small sized organizations to a great extent since they can deliver good quality service at a minimal cost. Cloud computing service providers like Amazon, Google and Microsoft invest great capital to buy the resources, rather than the user. With reference to cloud computing, Tim (2010) stated an analogy: “why buy the taxi when you can use it”. These services can be used from anywhere in the world via the web. The level of scalability is almost unlimited in cloud computing therefore it will never cease to cater to any demand of the user. This aspect fulfills the criterion of the efficiency since companies can offer uninterrupted service without reduced instances of crash or overload of request. Service providers can offer their service effectively in both, typical and atypical business conditions, by acquiring as much computing power as required. 3.2 Level of Mobility Cloud computing offers mobility (with respect to data and intellectual assets of the organization) since it resides on the shared servers of the cloud computing service providers. The data can be accessed from anywhere; this enhances its mobility and accessibility. Laplante (2011) pointed out some advantages that emerge from such mobility; offices can be displaced from one place to the other with minimal hassle since there is no physical movement of IT infrastructure and thus setting up at the new one. Partnerships can be facilitated by the platforms of cloud computing since the sharing of data with other parties can be achieved with convenience. 3.3 Maintenance and Update Requirements All the resources that are offered in the cloud computing platforms are up-to-date therefore require no updates or patches. The service providers ensure that they offer their customers with the most updated versions of the applications, software etc to earn customer loyalty and greater market share in the competitive market. No extensive IT maintenance team is required to supervise and monitor the data centers since it is not physically present within the vicinity of the organization. Security of the data and intellectual assets is one of the major concerns for any organization. Organizations tend to invest major portions of their capital in the deployment of effective security measures. The management can focus on the business processes, quality of the service rather than on the maintenance and security of the data centers. 4. Physical Issues of Traditional Data Centers Along with the digital threats of security, server crashes, denial of service attacks etc, some physical issues have also prevailed since the integration of computers in the business field. Some of the physical issues that have been witnessed in the traditional data centers have been related to the IT equipments, such as, excess power consumption, quality of the air in the data center and temperature of the environment. All of these threats increase the costs for the organization and pose a probable threat of harming the equipments. Several techniques have been adapted in the past to overcome the physical issues like the placement of sensors at different places in the data centers. The computational overload in any small area of the data center could cause the increase in the power consumption and thus cause that small area to become more heated that the other areas. Cowan and Gaskins (2006) stated it is due to this reason that many sensors were needed to accomplish the task since poor humidity levels or high temperature could be present in any small area of the data center without the knowledge of the management. Another limitation was that the increment of the IT infrastructure involved the expansion of the premises of the data centers. 5. Resolutions of Physical Issues in Cloud Computing Data Centers The most prominent cloud computing service providers, like Microsoft and Google, have provided their customers with energy efficient and massive data centers that can cater to their (almost) unlimited computing power needs. According to Johnston (2010); Microsoft launched two huge data centers in 2009. One was opened in Chicago that was spread over 700,000 sq feet of space and the other in Des Moines, Iowa. New techniques and technologies have been adapted by them to contain energy and equipment costs. ToolBox.com (2008) described PUE in the following manner; Power Usage Effectiveness (PUE) is defined as the ratio of the accumulated energy consumption (server plus cooling) and energy consumption of the server that is utilized for some useful purpose. The constant endeavors of Microsoft to lower the energy consumption have earned them great results. The graph showing the decrement in their PUE is present in Appendix A. One of the main aims of modern data centers is to improve the PUE as much as possible. Microsoft used containerized servers in their new data centers; these containerized servers are 20 to 40 foot long freight containers that give the company with more computing power in a smaller space. These servers are expected to be one of the main components of the modern data center designs since they are preconfigured with servers, required infrastructure to provide it with the cooling and power facilities etc. Johnston (2010) explained that their goal has been to maximize the computing capability of every containerized server while lowering the cost as much as possible. These servers can be wheeled inside the premises of the data centers thereby utilizing any empty spaces in the data center and avoid the physical expansion of the premises to a certain extent. Johnston (2009) explained that the density of these modern design servers is known to exceed 10 times as compared to the traditional data centers. Google has also announced the usage of such data center designs in their upcoming cloud computing data center facilities. According to Wheeland (2009); one of their data centers (that was discussed in Google Data Center Efficiency Summit in April 2009) was located on an area of 75,000 sq feet. The usage of such innovative data center designs decreased their PUE to 1.25. The environment of the data centers bears great importance in the computing performance and maintainability of the IT equipments. However, these aspects are mostly ignored by companies and not believed to improve the efficiency in any level. Microsoft issued a report on the new approach of maintaining their data centers and stated that they cleaned the roof and painted it white. Bhandarkar (2010) explained another approach in the report; it involved the dislocation of the walls in the region of the external air conditioning units; this facilitated a better supply of air to the equipment. He also stated that these additional changes contributed to almost 25% lowering of the PUE in two years. The cooling techniques have also been changed in the new data centers. Johnston (2010) stated that evaporative (water) cooling seems to work better than the rest of the cooling techniques adapted in data centers. This technique performs even better when the temperature is high and humidity is low. Environmental Leader (2010) stated that evaporative cooling, alongside other techniques, provided up to 80% saving in energy consumption in Voonami’s new data centers. Whereas, the areas where the weather is cold, seem to benefit from the air that is present outside the data centers (also referred to as free-cooling). The ice cold towers that might be present outside the premises of the data centers tend to take the load of the cooling mechanism. 6. Conclusion The conventional enterprise data centers have been found to possess several limitations and constraints that make their adoption difficult. Enterprise data centers constitute a limited storage capacity and much capital is needed to increase the scale of the data centers. Cloud computing data centers offer the level of scalability that facilitates almost unlimited expansion of the IT infrastructure. The cost of the cloud computing service is dependent on the usage of the requested resources therefore proves to cut costs by significant numbers. The adoption of this technology provides freedom in terms of maintenance and regular updates of the data center software and platforms. Cloud computing data centers promote mobility since they can be accessed from anywhere on the web. Therefore, it can be concluded that cloud computing data centers prove to be effective in typical and atypical business conditions but fall short in the security area. The innovative techniques to improve the physical conditions of the data centers have benefitted the organizations, as well as the society since lesser energy is being consumed from the resources of the planet. Containerized servers, innovative data center designs, evaporative cooling, free-cooling are some of the techniques that have played their role in the reduction of PUE. 4.1 Recommendations After the extensive study on the respective subject, it has been witnessed that the security aspects of cloud computing have not been addressed to a great extent. Sternstein (2010) and Boulton (2010) explained some cloud computing security standards and pointed out that it is still a developing domain since the security standards have only been formulated in the mid of the last year. According to a recent study quoted by Help Net Security (2011), 40% of the executives do not prefer the adoption of cloud computing in their organizations due to the security and privacy concerns related to it. The study had a sample of 834 executives who were chosen from 21 countries. An extensive survey of this nature revealed an alarming number regarding their inhibitions in adopting cloud computing. Mostly all the executives agreed that the technology provides better performance and efficiency than enterprise data centers but hesitated with its deployment due to privacy and security issues. The need of the hour to make the technology more accepted among the masses is to introduce greater security measures in the concept and ensure that all the service providers are following the cloud computing security standards. Some approaches have been proposed to enhance the security of cloud computing services: It has been found that cloud computing service providers have already implemented the encryption of the clients’ data for greater security. The default encryption scheme might be used by many clients under the assumption that it would provide them extreme protection from decrypting attempts. However, if an intruder cracks the code of the encrypted data then the intellectual assets of all the organizations can be put at great risk. Security tokens should be used for the generation of the personalized encryption code; these tokens should be possessed by few of the chosen representatives of the organization who have the authority to access the data on the clouds. These tokens would be programmed to display a new code at every cycle but all the tokens would be programmed to display the identical code since it is required to be accessed by all the representatives. The interval for the alteration of the code can be programmed as 24 hours since any less alteration interval might prove to be a hindrance in the performance level. Thus, the code will be changed every day and data can be encrypted with a new scheme every day. This will give the companies greater level of control over the security of their intellectual assets and win their trust over the innovative technology. Biometric authentication systems are one of the most reliable forms of authentication in the modern world. They promote convenience as well as uniqueness. A biometric approach should be adapted by cloud computing providers that would enhance the identification process and introduce more security. The fingerprints of the authorized personnel should be registered in the enrollment phase of the authentication and later used as a template for matching purposes in the verification phase. Such authentication will be useful when the user wishes to access the file server, database of the organization etc. 4.2 Future Work The security standards of cloud computing shall be analyzed in detail for their effectiveness and the rate at which they are being adapted by the respective providers. Data encryption mechanisms shall also be studied that have been adapted by different organizations to improve the cloud computing services. Appendix A Figure 1: Users and providers of cloud computing (Armbrust, 2009) Figure 2: Top 10 technology priorities in 2011 (Gartner, Inc., 2011) Figure 3: Comparison of cloud computing data centers and enterprise data centers Source: http://blogs.zdnet.com/Hinchcliffe Figure 4: Lowered PUE of Microsoft from 2004 to 2007 (Bhandarkar, 2010) Appendix B Armbrust, M. (2009). Above the Clouds: A Berkeley View of Cloud Computing, Technical Report, University of California Boulton, C. (2010), Google Apps for Government Meets Federal Security Standard, E- Week.com, July Issue. Bhandarkar, D., (2010), A Holistic Approach to Energy Efficiency in Datacenters, Microsoft Corporation. Cloud News Desk, (2008, November 3), Six Benefits of Cloud Computing, Web 2.0 Journal Cowan, C., Gaskins, C., (2006), Monitoring Physical Threats in the Data Center, American Power Consumption. Environmental Leader, (2010), Utah Data Center Uses Evaporative Cooling Along Way to 80% Energy Savings, Retrieved from: http://www.environmentalleader.com/2010/02/26/utah- data-center-uses-evaporative-cooling-along-way-to-80-energy-savings/ Gartner, Inc., (2011, January 21), Gartner Executive Programs Worldwide Survey of More Than 2,000 CIOs Identifies Cloud Computing as Top Technology Priority for CIOs in 2011, Business Week. Help Net Security, (2011), 40% of executives not planning to adopt the cloud, Retrieved from: http://www.net-security.org/secworld.php?id=10463 Johnston, S., J., (2010, Aug 4), Microsoft Details Efficient Data Center Designs, Datamation, Aug 4. Johnston, S., J., (2009, June 30), Microsoft's Mega Datacenters Coming Online, InternetNews. King, R. (2008), Cloud Computing: Small Companies Take Flight, Business Week, August Issue 2008. Laplante, T., (2011), Mobility – Cloud Computing’s Partner in Crime, Central Penn Business Journal Gadget Cube, February Issue. Miller, M., J., (2010, July 19), The Changing Cloud Platforms: Amazon, Google, Microsoft, and More, PCMag. Microsoft Corporation, (2011), Cloud solutions, Retrieved from: http://www.microsoft.com/en- us/cloud/cloudpowersolutions.aspx Rajan, S. (2011), Application Performance Management - Data Center vs. Cloud Computing, Ulitzer, January Issue. Sternstein, A., (2010), White House set to complete security standards for cloud computing services next year, Next Gov, Retrieved from: http://www.nextgov.com/nextgov/ng_20101202_1174.php?oref=rss Tim, M. (2010), Cloud computing explained, GoGrid Cloud Hosting Services, Retrieved from: http://pyramid.gogrid.com/ ToolBox.com, (2008), Power Usage Effectiveness (PUE), Retrieved from: http://it.toolbox.com/wiki/index.php/Power_Usage_Effectiveness_%28PUE%29 Warr, W., A., (2009), Cloud computing, Retrieved from: http://www.qsarworld.com/files/Cloud- computing.pdf Wheeland, M., (2009), Google Data Center's Massive Energy Appetite, Green Biz Group. Read More