TMCnet News
A FEW IMPORTANT STEPS FOR CLOUD SERVICES(Daily Trust (Nigeria) Via Acquire Media NewsEdge) The event of a few weeks ago is good news for the providers of cloud services: the U.S. government provisionally allowed its Department of Defense (DoD) contractors to use Amazon cloud (Amazon Web Services, AWS) for storing, processing, and transmitting some unclassified government data. Even though the Nebula cloud development was funded by the U.S. government, the facility does not process any U.S. government data that is remotely sensitive. Let me recap a few of the challenges that have plagued the adoption of cloud services. You would think that everyone will be beating a path to cloud computing services, given the potential advantages the technology has for you. For example, you do not need to acquire and maintain the computer hardware that you require for your work, you have easy access to "self-upgrading" application software products in some cloud deployments, and there are ready software platforms where you can develop your own applications that you sell to people. You also have essentially inexhaustible random access and disk storage memory capacities that scale with your needs. As I suggested in a previous article in this column (7 October 2013), cloud computing in engineering and related tech areas should be god-sent, a no-brainer, the holy grail of engineering analysis, as well as the low-hanging fruit of cost-reduction in the engineering analysis business. After all, cloud computing promises seamless collaboration on engineering projects over diverse geographical regions on the globe. Your company's employees in South Korea, US, Scotland, and Nigeria can simultaneously login to the cloud and carry out an analysis on a computationally-intensive project that the employees are collaborating on, or they can undertake a "post-processing" of an already analyzed engineering problem. The various branches of your company, no matter where they are located on the globe, can share a common pool of "elastic" computing infrastructure, computer network technicians and engineers; heating, ventilation, and air-conditioning (HVAC) facilities, and so on, all of which may be located in one physical building. Furthermore, with cloud computing, the whole of your enterprise all over the world can use the same versions of the analysis software tools at any point in time. Thus, the case for cloud computing in engineering analysis is quite strong on technical and economic grounds. So, what is wrong with cloud computing? There are barriers to widespread adoption of cloud computing in many aspects of enterprise. For example, in technology, issues such as security, data privacy, intellectual property (IP) violations, the management overhead associated with the humongous databases in SQL or MySQL, the overhead of moving large data back and forth on the Internet, with the attendant latency problems and bandwidth costs, and the need to sync parallelization procedures are some of the "show stoppers. Visualization, which is usually an integral part of engineering analysis, is hard to do online, and approaches currently in use - such as those based on Microsoft Silver Light, Adobe's Flash technology, and WebGL - are not well developed for the visualization tasks in engineering analysis. Moreover, there is significant vendor lock-in in this area of Internet technology. In my last article on this topic in this column, I did not emphasize the role that inadequate bandwidth plays in the adoption of cloud computing in engineering practice. More recent research results and opinion articles since then have suggested that the issue of bandwidth is indeed of major concern to many engineering enterprises that would have otherwise adopted cloud computing. You see, engineering analysis tends to be data-intensive. For example, when you simulate phenomena that depend on time, you may be forced to transfer data from the cloud to your desktop at every time step during analysis. Real-time visualization (on your desktop) of analysis results (computed on the cloud) involving massive data is infeasible because of the bandwidth requirement. Other concerns that engineers have with cloud computing include intellectual property (IP) violation, fear of data loss, and cybersecurity. The IP and data loss issues are more relevant to commercial engineering industries, while governments and major defense contractors will in addition be very concerned with cybersecurity. An article that appears in the 17 July 2014 issue of The Wall Street Journal focuses on cloud adoption issues in the enterprise resource planning (ERP) business, suggesting a few reasons why big firms delay using the cloud. The authors, Steven Norton and Clint Brown, feel that "there is still too much risk associated with potentially unreliable Internet connections and a dependence on third parties to manage computer servers. They quoted a chief information officer (CIO) of a major company saying: "if our ERP system goes down for five days, we're out of business." Another CIO who uses SAP felt that "moving such an ERP system to the cloud would be akin to a "heart transplant." Yet another CIO said: "thinking about the complexity of moving the ERP systems to the cloud 'makes my heart hurt.'" A few weeks ago, Amazon Web Services became the first commercial cloud provider authorized to handle the Defense Department's most sensitive unclassified data. Amazon also says that it recently launched a private cloud for the U.S. Central Intelligence Agency (CIA), to service the intelligence community. Other cloud providers are also said to be picking up new business in the civilian government. Thus, the outlook for cloud services is quite encouraging. The take-away from this article is that the U.S. government has somewhat eased the nerves of some jittery potential cloud service users by allowing a portion of unclassified Depart of Defense data to be stored, processed, and transmitted on Amazon cloud. The acceptance of the technology so far by the CIA is also encouraging. |