HPC in the Cloud Required Reading Assignment
Just who does this editor think she is, giving me a homework assignment when I have all of these complex issues to address in the short space of a weekday?
Okay, so no, you’re right, I really have no business demanding that you immediately hit “close all tabs” to read this article from a few great minds at the University of Bristol and University of St. Andrews in the U.K. but I will demand it anyway.
Do you have it open in front of you already?
Oh, good! Such remarkable powers of persuasion I have…
Chances are, if you take a moment to skim through the subheadings and bullet points, you’re going to notice that just about every pertinent question, objection, and source of cloud-related praise or “pshaw” is addressed here, even if it is done so fleetingly. Actually, if I had not been so thoroughly engrossed in the work of Nicholas Carr this week in particular, I could have skipped his book and headed straight for this article to give me the abbreviated version. While it may not provide the level of historical context or scope Carr’s manifesto, Rewiring the World: From Edison to Google serves, it is nonetheless one of the best mid-level primers on the intersection between enterprise and the cloud—except it just happens to boast the beautiful addition of HPC into the subject.
I realize that this article is not going to be for everyone; not all of us have large-scale enterprises to keep in the air and few of us are taking a lone-wolf approach to managing existing HPC technologies in our own organizations. However, if you are directly involved in large-scale data-intensive operations and are either using or considering cloud options, this article, which is called Research Challenges for Enterprise Cloud Computing, delivers discussion on just about every imaginable issue that’s likely already been the subject of lengthy meetings between every member of the organization—not just IT and management. And that’s because either adopting or migrating to cloud computing in large enterprises is kind of a big deal, yes?
What I appreciate most about Research Challenges for Enterprise Cloud Computing is the detailed literature review on the topic in addition to its objective handling of the diverse and numerous questions that everyone in enterprise cloud and HPC seem to be asking all at once. These questions include (and sure as heck are not limited to): Is this a secure way to manage my data and processes? Will migration to cloud be a total nightmare or just a bad dream? Is there a clandestine force of superpolice who are going to help maintain standards in the cloud? What is going to happen to our organization when the cloud rolls in?
While the answers to these questions are not always delivered on a silver platter, the sheer range of opinions, mostly scholarly, is enough to help you answer your own questions given the particulars of your organization and its needs. It’s an exercise in reading between the lines to find what you want.
In the end, however, despite the fact that the authors provide you with a sufficient number of tools to begin working on your own solutions to the riddles you’ve created with your endless speculation about clouds and HPC, there are still just as many questions. The difference is that these remaining variables are less functional and more ethereal. As conversations about the cloud continue over the coming years, especially in the realm of HPC, these slippery questions will hopefully begin to give way to more thorough questioning of the critical but peripheral topics.
I had a timed multiple-choice quiz all ready for you to make sure you scoured through this week’s reading assignment but figured that you’ve already had your share of testing if you’ve either just implemented clouds or are examining the possibility.
You’re off the hook for now.
Did you peruse this article? What issues were of the most interest to you? What didn’t the authors touch on, if anything?
P.S. The authors of the article in question, Ali Khajeh-Hosseini (University of St. Andrews), Ian Sommerville (University of St. Andrews), and Ilango Sriram (University of Bristol) are all definitely worth Googling.