Cloud Computing and Machine Learning

Introduction

Cloud computing is Internet-based utility computing, basically shared resources, software and information that are used by end-users hosted on virtual servers. Some people term anything beyond a companies or user’s personal firewall to be in cloud (Knorr, Gruman, n.d., para. 2).Personal computing has evolved in three phases. The first phase was where the data and application was stored on a local desktop. The second phase is where the applications reside on a local server and utility software on the desktop along with the internet to provide valuable information. The third phase is where most of the data and software will reside somewhere on the internet. This phase is identified as cloud computing. Although more than 80% of world-wide computational data is being predicted to move into clouds in the next five to ten years, there are a number of unanswered questions which will decide the speed of development in this arena (Nelson, 2009, para. 1). It is a new concept in the field of information technology still trying to get acceptance, where-in the technological services/ applications are provided by a third-party.

Cloud Computing – Nomenclature

From the days when use of internet started becoming popular, the network architecture diagrams show internet as cloud to hide the complexities from the end-user – hence the name cloud computing. The applications/ software are provided as services which reside in data centers with server farms and redundant storage. End-users can access them via web browsers in laptops or hand-held devices. Google Apps, Facebook and salesforce.com are examples of cloud computing which are becoming very popular especially with the younger generations. A very basic example of cloud computing is a web-based email service like Gmail where the complexity of storage and presentation of user data is provided by a third party, and users are leveraging them by means of web browsers(Nelson, 2009, para. 12).

Benefits of Cloud Computing

Cloud Computing is a totally different paradigm of personal as well as corporate computing which drifts away from the traditional software business model. There are several positive factors for deciding in favor of moving to cloud computing. Some of the important favorable factors are discussed below. Entry cost for a business to set itself up in cloud is much less compared to traditional setup. Cost for running Information Technology (IT) business in cloud world can be visualized as operational cost. Organizations no longer need to shell out big amounts for hardware and software upgrades (Cunningham & Wilkins, 2009, p. 3). They no longer need to worry about end of life for hardware/ software. Companies no longer need to spend huge amounts on buying new hardware for scalability. In short, infrastructure costs to get in to cloud are much lower than that for the traditional model. Pricing is generally based on usage options. Organizations do not need to maintain a workforce of IT people and can focus on strengthening their business domain knowledge.

Organizations can use only applications which are suited for their need and not pay huge amounts on licenses for using a software suite. Most of the time buying a software suite is of lesser worth from the business perspective as only few features of the suite are really used and the cost associated with upgrades and licenses are expensive. With canned applications in cloud, business can decide and pay for only the applications that they need. Using virtualization as the main technology, additional computing resources can be added dynamically without having any downtime. By monitoring peak load and server usage, an application can be dynamically switched over to a virtual machine with greater computing speed or the current virtual machine can be augmented with more computing resources . The organization using the application does not need to pay extra for the scalability feature. In virtualization, high performance servers are split into multiple machines catering to different customers. Some of the popular products are VMware and XEN. With the infrastructure being located in multiple data centers and huge server farms, users can be productive anytime from anywhere in the world using web-browsers in laptops, desktops or hand-held devices. Business executives travelling to client locations need not carry hard copies and everything with them, as with a simple click of the mouse they can reach the information gateway.

Concerns with Cloud Computing

As there is darkness after light similarly with the benefits discussed in the earlier sections there are a few concerns floating around with the acceptance and utilization of cloud computing. With the applications provided as service and being used by different customers, personalization seems to be one limitation. The characteristics are similar to a COTS product but if proper negotiations are made, product-based companies most of the time do allow customization. There is a serious concern with data being centralized and out of business control. Sensitive data will now be under the control of a third-party and according to some experts this is a compliance issue and some feel that this is probably going to be the end of confidential records management (Cunningham & Wilkins, 2009, p. 7). When it comes to the disclosure policies, some people do not know where to draw the line and knowingly or unknowingly share vital information and with data out on the third-party space that concern becomes even more critical. With everything accessible through internet, business will be totally dependent on the network and the service provider’s infrastructure. Business will come to a halt if the network/ internet are down. As illustrated by Patrick Cunningham in his article (2009), currently in the IT world when we need to troubleshoot an issue, one advantage in our favor is that the application logs and the database are within the premises of the enterprise. With the shift to cloud, this aspect will be lost and thus special support or contract binding needs will be there between the service provider and the business for e-discovery. With business sensitive internal data being maintained by the vendor the dependency on the vendor’s unique API and proprietary interfaces could create a possible lock-in with the vendor. If under some circumstance the business is dissatisfied with the vendor, moving to a new vendor means data needs to be reformatted and converted which can be time consuming and expensive.

Return on Investment

By moving to cloud computing companies can save huge amount of money. As illustrated by Raichura (2009), in couple of online articles please find below charts showing examples of savings that an organization can achieve by moving to the cloud. The first table illustrates the difference in cost of having infrastructure on premises versus having it in the cloud. The second table illustrates the cost saving considering storage, service, infrastructure and platform in the cloud.

Conclusion

After having discussed the merits and de-merits it clearly stands out that there is no one clear answer in favor or against of cloud computing. So far it seems that is going to be a hybrid solution from the corporate standpoint. Some critical, business sensitive applications will continue with the traditional business model till the concerns about privacy and legal matters are cleared from cloud computing whereas simple canned applications will become more and more popular in the cloud space. The concept and technology is here to stay but it still in its infancy and there is a long road ahead to get to maturity.

Add a Comment

Your email address will not be published. Required fields are marked *