subscribe to this page
by William Bulkeley
The Wall Street Journal
October 8, 2007
This article announces an interesting new partnership between Google and IBM. The partnership will form a cloud for data storage and processing by computer science and programming students and professors. The cluster of computers forming the cloud network will consist initially consist of 400 computers, with plans to expand to 4,000. The cloud will be led by the University of Washington in Seattle, but it will be accessible by students and faculty at 5 other universities including: Carnegie Mellon University, Stanford University, MIT, University of California at Berkeley, and the University of Maryland. The purpose of the project is to improve computer science and computer programming academic curriculum. Until recently, curriculum in both fields has been primarily focused on "teaching students how to program a single server and not giving them opportunities to learn about parallel programming." [pg. 2] IBM and Google intend to create clustered computer environments in academic centers so that students can be more familiar with their (and a growing number of other companies) style of infrastructure in preparing them for future professional experiences. Google and IBM will initially each contribute $20M-$25M to get the project off the ground.
Is operating in the cloud really the way of the future? Obviously, Google and IBM (along with Amazon, and several others) would say "of course!" This article raises questions about whether new technology is adopted because it is necessarily more efficient, cheaper, etc., or because it is what professionals in their respective fields are taught and feel comfortable using? This collaboration between IBM and Google will be interesting to follow and learn more about as time progresses.
"The Information Factories"
by George Gilder
Wired Magazine
October 2006
Google is building a new server farm along the Columbia River in a small town in Oregon called The Dalles. This server farm will be the site for Google's new 30-acre campus and it will be the largest and most powerful server farm Google has built to date. This farm has been located in The Dalles because the town is home to a dam with a 1.8 Gigawatt power station and the next generation of servers revolves around the issue of power. In order for Google to operate its servers the limiting factor is access and the cost of energy. Google currently has about two dozen server farms located around the world and in total, these farms house an estimated 450,000 servers. Yes, this is a description of the network of hard wear that makes up Google. For you tech folk, the Google servers are estimated to have a capacity of 200 petabytes of hard disk space and 4 petabytes of RAM. This would be enough to copy everything on the Net dozens of times. To put the energy consumption issue in perspective, last year, the servers of the major internet search engines consumed just shy of 5 gigawatts of energy. 5 gigawatts of energy is enough to power the Las Vegas Metropolitan area operating at full force on the hottest day of the year.
Why is the article relevant to a discussion about cloud computing? Cloud computing is about using the power of remote computers systems and servers to operate rather than using the hardware and data storage of your personal computer. The server farms, like the new Google farm being discussed in this article, represent the infrastructure behind cloud computing. In order to use Gmail, view and upload videos to YouTube, etc., we rely on these server farms. As server farms grow larger to house more and more data (as a result of cloud computing) and energy costs increase, what will happen to cloud computing? Could the cost of energy to run server farms eventually lead to the end of cloud computing? Not if there is an alternative energy source...Not surprisingly, Google's foundation, Google.org is heavily invested in research pertaining to alternative energy sources.
“Google CEO: ‘Cloud Computing’ Is Key to Patient-Owned PHR’s”
by Don Long
Medical Device Week
March 3, 2008
This article reports on the 2008 Health Information and Management Systems Society’s (HIMSS) annual conference, at which Eric Schmidt (Google’s CEO) was a presenter. PHR stands for “personal health record,” a term familiar to many in the health care industries, but not so familiar to those outside the industry. A PHR is essentially a log or journal of an individuals health care treatment, evaluation and laboratory results. Today, many individuals do not have PHRs, but rather, their doctors and other health care providers individually maintain medical records on their behalves. PHRs in an electronic format has been referred to as EMRs (electronic medical record), and, “[t]he federal government has been the pioneer in the field of developing standards for universal, interoperable and portable EMR that are often linked to the development of universal health care coverage.” [pg. 1] Of course the underlying assumption with regard to EMRs is, “that the government would be the one to have, and the one to control, this information.” [pg. 2] One of the primary themes of discussed at the conference was the importance of who controls and has access to an individual’s PHR/EMR. The distinction drawn between the two is that ideally, a PHR would also be in electronic (digital) format, but rather than having the government have access too and control individuals medical records (as would be the case for EMR), each individual would control their health care information. Hence, the idea is to make PHRs “patient-centric,” as opposed to having the government or health care providers in control of individuals’ medical records. [pg. 2]
This is where cloud computing comes into the picture. Schmidt was at the HIMSS conference to give a presentation on Google Health. Google Health is a technology that Schmidt says is still in the development phase. What is Google Health? According to Schmidt, Google Health is “a system for enabling the creation of PHR, based on ‘cloud computing’ – offering healthcare data that would be completely portable and privacy-protected.” [pg. 4] The idea is that the system will be “consumer-focused – users can access their data and control who sees it. The data follows the consumer, wherever they go. Interoperability is important…[the system]...would not be tethered to a particular health system.” [pg. 4]
This article raises a few important policy issues. First, there is the issue of creating the informational system necessary to consider providing universalized health care and how individuals medical records should be controlled, stored, and collected. Second, rather obviously stemming from the prior matter, is the issue of privacy. Hypothetically, the idea behind using cloud computing (through Google Health) to create and manage individuals’ PHR is so that the individuals have control over their medical records and have the ability to grant a particular health care provider or the government access to their records upon their approval. If privacy is such an important issue, we should certainly be asking whether we want Google serve as the gateway for establishing and maintaining our medical records and history.
“Computing In The Clouds”
by Aaron Weiss
The Guide to Computing Literature, Networker Magazine
December 2007
In 1943, IBM Chairman Thomas Watson said, “I think there’s a world market for maybe five computer.” [pg. 18] The personal computing industry that began in the 1970’s and current popularity of cloud computing prove that Watson’s statement could not have been more wrong. Weiss defines cloud computing generally as the ability to distribute computer processes over a large number of small computers/servers in order to maximize the efficient use of resources. The idea being if one were to do an internet search through Google, for example, that Google could distribute the work of doing the actual search over a large number of computers rather than one large (and powerful) computer doing the search and returning the results to the user. The relevant question, in this case, is how does Google most efficiently distribute the task of fulfilling the search to many individual computers/servers in order to decrease the time it takes to conduct the search and then return the results to the user.
The article also provides working definitions of SaaS and utility computing in order to understand how they relate or should be considered as part of the larger cloud computing phenomenon. The most important and influential SaaS established to date is the creation of web-based email. While many individuals, organizations and companies do not entirely depend on web-based email service, the trend is quickly moving in that direction. Weiss refers to SaaS, as in the case of web-based email, as merely a revival of an older concept known as “thin client” computing. In the realm of cloud computing, the most relevant concern that emerges is privacy because operating in the cloud and allowing a third party to store and/or process your digital information requires a high level of trust.
This article is relevant because it sheds light on the fact that cloud computing is a popular, “buzzword almost designed to be vague[.]” [pg. 25] One reaction to this piece is to feel that it is not possible to provide a complete definition for the terminology ‘cloud computing.’ Nevertheless, a more appropriate conclusion might be to think of cloud computing as a trend that “draws on many existing technologies and architectures.” [pg. 25]
danah boyd is a doctoral candidate in the School of Information at the University of California-Berkeley and a Fellow at the Harvard University Law School Berkman Center for Internet and Society. apophenia is her blog.
In this post, she relates a story from a friend of hers, who centralized virtually all of his online life (and therefore, his life) on Google products, only to have his account erased by Google because of a fraudulent phishing attack. The story raises concerns about storing all your data in one place and with one company.
Article from Harper's Magazine about how much Google's server farms are polluting a river in the Pacific Northwest.
Google Docs--but at what cost?