Posts tagged Case
[Case Study] Lessons in High Performance Computing with Open Source
Feb 2nd
Providing adequate software and tools for researchers has always been of great importance to organizations, but has often come at a great cost. In an era of constantly evolving technology and rapidly dwindling budgets, my IT team has had to work with a large pool of researchers to provide cost-effective solutions that meet the ever-growing demand for innovation and computing power.
I am an Information Technologist for the Department of Statistics and Probability at Michigan State University. The Department is home to award-winning faculty with a wide variety of expertise in fundamental and interdisciplinary research, and over 100 graduate students from all over the world. Keeping the faculty and students ahead of their research is a constantly evolving challenge for my team and I.
Sponsor
Evolution of Statistical Software
Erik Segur is an Information Technologist for the Department of Statistics and Probability at Michigan State University.
For many years, most statistical analysis in our department was done in Matlab, S-Plus, SPSS or SAS. Even with a Higher Educational discount, most of the software required yearly renewal fees that quickly devoured our IT budget. Things started to change when the R language, which was first developed in 1993, began to gain traction in statistics communities in the early 2000s. R is an open source programming language and software environment that is used for statistical computing and data analysis. Several years ago, we began the transition at Michigan State to R; today, it is used for the majority of the research in the department–as well as being a central focus of our statistics curriculum. By switching to the free, open source version of R, our department has been able to cut thousands of dollars each year in software costs and have focused more on fueling and expanding research.
Lesson #1: The Shortcomings of Open Source
As more people began to use R and the analysis became increasingly complex, researchers began to face a large problem: time. Research was taking several months to complete in terms of processing jobs. Often, there is a need to run the calculations several times to ensure accuracy; waiting three months for one to complete was simply not feasible. It was taking R this long to process the jobs because the iterations were computed in serial, one right after another, using only one processor core at a time.
Bo Cowgill from Google once said “The best thing about R is that it was developed by statisticians. The worst thing about R …is that it was developed by statisticians.”
Until the spring of 2010, R was a 32-bit application and could only access a limited amount of memory. The maximum amount of memory that could be accessed by R was only 3GB. When dealing with large datasets researchers were quickly running out of memory as well as discovering they needed a solution to deal with large data efficiently.
Bo Cowgill from Google once said “The best thing about R is that it was developed by statisticians. The worst thing about R …is that it was developed by statisticians.” Even though R was–and still is–constantly evolving, the department needed a solution that could keep up with hardware technology and compute calculations in an efficient, scalable manner.
Lesson #2: Find Commercial Enhancements for Open Source
Our search for a more effective version of R ultimately brought us to a product called Revolution R Enterprise by Revolution Analytics, which provides commercial support and software for open source R. It takes advantage of multiple processor cores by using optimized assembly code and efficient multi-threaded algorithms that use all of the processor cores simultaneously. Although this addressed a lot of the issues of open source R, professors were only using Revolution R on their desktops. The next question was, how we could combine the power of our servers to dramatically decrease our computation times?
Lesson #3: Expanding to Infinity and Beyond
Open Source R is a memory-bound language. This means that all of the data, matrices, lists etc. need to be stored in memory. Issues quickly arose when data sets became several gigabytes large and were too big to fit into memory. This required implementing parallel external memory algorithms and data structures to handle the data. These challenges were tackled by Revolution Analytics as they developed the R language for a High Performance Computing (HPC) environment.
There are often great pieces of software created through open source, but they generally lack key features needed for an enterprise environment. Combined with commercial backing and expertise, these projects can be further developed and expanded to meet the needs of large-scale enterprise environments.
In 2010, Revolution Analytics offered Revolution R Enterprise free for academic users and shifted the focus of their enterprise software to big data, large scale multiprocessor computing and multi-core functionality. Revolution Analytics was going to tackle everything the department needed. The evolution was complete: open source R went from an inefficient single core program to a HPC environment.
Once the department could schedule R jobs in an HPC environment, the demand began to drastically increase. The HPC cluster is now scheduling more than four times the amount of jobs that were scheduled in previous semesters, from 200 jobs over a year ago to over 800 jobs this past semester. Jobs that were taking over three months to complete on open source R were completed in less than a few days with Revolution R. Computational jobs are now run multiple times with significantly higher levels of accuracy than ever before.
Conclusion
There are often great pieces of software created through open source, but they generally lack key features needed for an enterprise environment. Combined with commercial backing and expertise, these projects can be further developed and expanded to meet the needs of large-scale enterprise environments. IT departments can provide enhanced solutions to their users that adapt to the expanding world of cloud and High Performance Computing environments–all while minimizing the impact on a shrinking budget.
Photo courtesy of Shutterstock.
View full post on ReadWriteWeb
Case Over Who Owns Ex-Employee’s Twitter Followers Moves Forward
Feb 1st
San Francisco-based U.S. Magistrate Judge Maria-Elena James will allow a case by a company arguing that a Twitter list created by an ex-employee is its property to proceed.
PhoneDog LLC, which reviews mobile phones and other tech products, is claiming that former employee Noah Kravitz owes it $340,000, or $2.50 for each Twitter follower he kept by switching the name of his Twitter account after he stopped working for PhoneDog. James denied a motion to dismiss by Kravitz on Monday in a case that is being closely watched by companies that have employees develop social media platforms as part of their jobs.
Sponsor
The case, however, may not be the precendent-setting lawsuit employers and employees are hoping for. As reported last month, the case is fact specific, meaning it will be judged on its individual merits and similar situations at other companies may have different facts and, therefore, different outcomes.
“This case is another example of the application of relatively old legal rules applied to new technology,” Bill Nolan, an attorney with Barnes & Thornburg LLP, said in January. “It’s the 2011 version of the salesperson taking the Rolodex when he/she leaves the company.”
Kravitz started the @Phonedog_Noah Twitter account while he was tweeting and writing for the online publication. When he left the company in October 2010 he changed the account’s handle to @noahkravitz and retained the more than 17,000 followers he had amassed while working for PhoneDog.
“The Court’s decision yesterday [Monday] in effect means that PhoneDog has met the minimum requirements to survive a motion seeking to throw PhoneDog’s claims out of court, but it was not a decision as to whether or not PhoneDog is entitled to the relief it seeks,” Cary Kletter, Kravit’z attorney, said in a staement. “Ultimately PhoneDog will be unable to prove its allegations against Mr. Kravitz, and Mr. Kravitz will prevail.”
View full post on ReadWriteWeb
Northcutt Releases SEO Case Study on Ahosting, Inc. – PR Web (press release)
Jan 26th
|
Northcutt Releases SEO Case Study on Ahosting, Inc.
PR Web (press release) Northcutt, a competitive SEO and content marketing firm based in Chicago, Illinois, has completed a new six month case study on Ahosting, Inc. The Chicago SEO firm reports visitor gains of 207% and an increase of 228% in search engine traffic following … |
View full post on SEO – Google News
Yahoo Hits $610 Million Jackpot in Lottery Scam Case
Dec 10th
Yahoo has claimed victory in a $610 million lawsuit against a group of individuals it accused of operating a massive spam and fraud campaign.
A U.S. district court in New York issued a ruling in favor of the company, ordering the defendants to p…
View full post on Search Engine Watch – Latest
Google’s Performance Presents Case for PPC Opportunities
Dec 7th
Industry research firm comScore released their search engine market share report for the month of October. The report continues to illustrate what consumers already know. Google is dominating search engine market share. With an overwhelming 65.6 percent ownership of the market over second-place Yahoo! search (15.2 percent, respectively) 12 billion searches were conducted on Google [...]
Follow SEJ on Twitter @sejournal
View full post on Search Engine Journal
VMware’s Case for Reducing Cloud Management to Its Basic Tasks
Dec 2nd
Yesterday, during their rollout conference for WebLogic Server 12c, engineers from Oracle introduced developers to a concept they called Virtual Assembly Builder. Their idea was this: When you take a Java application, with all its myriad underpinnings and dependencies, and move it into a virtualized environment such as a cloud, certain libraries that the app needed when running on a physical server will need to be substituted with VM-aware components for optimum performance. The philosophy there is, you want the virtual machine to closely approximate the characteristics of the physical one, including how it works and how it’s managed.
When I ran that concept past a pair of VMware marketing managers this afternoon, they could not have disagreed any more pointedly with Oracle’s basic philosophy. During a demonstration of VMware technologies including vFabric, they explained why products like AppInsight, released last month, present a more simplified dashboard for administrators of applications deployed through cloud-based IaaS services.
Sponsor
“We’re currently seeing this bifurcation between what we call cloud providers and cloud consumers,” begins Steve Henning, VMware’s head of applications product marketing. “You can think of the cloud providers as the traditional infrastructure operations teams, managing the deployment and performance of infrastructure forever.” Organizations have become accustomed to using these teams for managing applications in terms of the resources they consume within their physical data center infrastructure, he says.
Just the facts, ma’am
Translating that context into the cloud, he argues, doesn’t make sense. Public cloud infrastructure managers should have their own tools (he suggests vSphere Operations Manager) for managing resources on their own behalf. The only information the owner of an application should require is not how the components of the cloud are performing, asserts VMware, but rather how the application is performing.
“The cloud gives the opportunity to be able to empower applications owners to be able to manage applications at the business level, independently of the underlying infrastructure,” the product manager goes on. “What we’re trying to do is bring together a set of tools that can allow the application owners to be able to efficiently and effectively deploy their applications, [and] manage their performance in a closed-loop way that’s independent of the infrastructure.”
Companies that are customers of public clouds are expecting service levels that approximate the numbers hammered out in their SLAs, or the tiers that providers often “color” with bronze, silver, and gold. So VMware says, these customers need their IaaS-based apps to report those numbers in the same formats. Performance problems within an application can be rendered in the context of such a report, so that customers can address those problems to the extent that they can: for instance, by adding storage capacity, compute power, or memory.
Maintaining the abstraction
A forthcoming version of vCenter Operations Manager, Henning tells us, will give cloud infrastructure managers an application-centric view of performance, alongside its existing VM-centric view. That way, application performance teams (should IaaS providers have them) may be dispatched to address issues at the applications level, on behalf of their customers. This has been a highly requested feature.
“If I’m an applications owner, and I’m deploying on a public cloud, why would I want to manage the performance of the infrastructure that is being given to me by that provider?” asks Henning. “I am expecting a level of service from them, and I should be able to get a measurement of that service, but I’m certainly not going to sit there and try to diagnose infrastructure problems. If I’m not getting the level of service that I’m expecting, I’ll go to somebody else who can provide me that level of service.”
By contrast, an application owner wants to see data points like: the relative speed of transactions, the throughput level of middleware, the operations per second for specific segments of Java or other code. “These are things that I can manage myself, independently of the infrastructure. And as long as I’m getting the green light that it’s performing well, I don’t need to manage the CPU and utilization on a platform I don’t even own.”
“A company that comes with an approach that cloud and/or virtualization is another layer, is actually missing the ability to unlock the value that cloud computing brings,” adds Shahar Erez, VMware’s group product manager for data center intelligence. “Cloud computing is all about abstracting the layers that you don’t need to understand in order to achieve productivity and agility in the layers you own… Not knowing allows me to focus on providing more business value in the layer that I can make impact on.”
Shifting abstraction from people to services
Erez gave us an example of an IaaS customer who complained that whenever he needs to scale his Web site, he has to place a request of the IaaS provider’s infrastructure team, which winds its way through the help desk, and eventually acquires the necessary approval to provision a new VM with a new deployment of Apache Tomcat. The whole process could take a full three weeks. That’s not the fault of cloud architecture, though; with the right tools and with less arbitrary bureaucracy, the provisioning process could instead take minutes.
How does the owner of the application know for certain that the service level reported by his IaaS provider is accurate, and not just some Perl script with a green light on it? VMware’s Henning tells us that the app owner’s vFabric tools will be able to spot performance problems wherever they are, and identify independently (not just through some certificate from the service provider) whether the problem is addressable on his end. “If it’s not, but these guys are tellin’ me I’ve got a green light, I’m gonna go move somewhere else. I’m not going to deploy agents with my application that are going to measure the performance of the infrastructure. If you tell me it’s green, and it’s really red based on everything I can see in my application performance management suite, that’s looking at it at the business level, I’ll go somewhere else.”
View full post on ReadWriteWeb
Two Examples of How One Line of Code Could Kill Your SEO [Case Studies] – Search Engine Journal
Nov 29th
Irish Hotel Drops Autocomplete Defamation Case Against Google
Nov 24th
An Irish hotel was suing Google for an autocomplete suggestion that implied the hotel was in financial trouble. That case has now been withdrawn, although no details of any settlement have been released.
Hotel Defamation Case Withdrawn
The …
View full post on Search Engine Watch – Latest
How Search Engine Watch Grew Google News Traffic by 300% [Case Study]
Nov 22nd
One of our stated goals in the Search Engine Watch redesign was to increase traffic from Google News in order to establish ourselves as an authoritative publication on the search marketing industry. The results of our project are below:
Averag…
View full post on Search Engine Watch – Latest
With Square’s Card Case, Pay By Saying Your Name
Nov 3rd
Today Square updated its Card Case digital wallet app. If you leave Square’s “auto open” tabs feature on and walk to your favorite merchant, you can place your order just by saying your name. Merchants who have signed up for Square will be able to see users who are within 100 meters of a location. Just say “put it on [insert name here],” and the payment will be completed. Then you’ll receive a push notification on your phone with the charge amount.
Sponsor
“We want to make payments fade away,” said Megan Quinn, director of products for Square. “People don’t appreciate that; they enjoy making a purchase and feeling like a regular at places they shop.” 
In his presentation, Jack Dorsey described mobile purchasing as something that should feel “magical,” like a one-click Internet purchase.
Merchant cards can be synced with a business’ Twitter stream. (After all, this is Jack Dorsey’s other start-up.) By integrating business’ Twitter streams into the app, users can see if there are any additional deals or discounts. Users can also tweet at a business with any questions, comments or feedback.
Over 20,000 merchants are catalogued in this directory since it opened to the public in August. Currently, 800,000 merchants have activated Square.
In August, Square updated its app, no longer requiring a signature on purchases under $25. Square launched Square Register, a tool for payment processing and for merchants to organize menus, sales and receipts, in May of this year.
Square is not an NFC. It’s also not a dongle, which turns your smartphone into a credit card reader. This new feature places Square in competition with dongle providers Verifone, Intuit, Erply, ROAMPay, TRUSTe and PayAnywhere.
View full post on ReadWriteWeb
