|
|
| Recent Changes Printable View Page History Edit Page | ||||||||||||||||||||||||||||||||||||
|
Content Last Modified on June 24, 2010, at 07:16 AM CST
SAS+SOA+Wireless+? ...posted 062105 by rdp Still hoping and waiting for inputs. Something will turn up? Super Computing ...posted 062205 by rdp By Stephen Shankland URL: http://news.zdnet.com/2100-9584_22-5443764.html After two and a half years at the top of a list of the world's fastest supercomputers, NEC's Earth Simulator has finally been dethroned: IBM's Blue Gene/L officially is the new king of the hill. The Blue Gene/L succession, while expected, reflects IBM's sustained push in recent years to expand its expertise from business computers into high-performance technical computing. On the newest incarnation of the Top500 list, updated twice a year, IBM has 216 systems, which account for 49.4 percent of list members' collective performance. The ascendance also highlights the rise of Linux, the open-source operating system that Blue Gene/L runs and that IBM helped to champion. The new No. 2 system, Silicon Graphics' Columbia, also uses Linux. Blue Gene/L performed 70.7 trillion calculations per second, or teraflops, nearly twice the 35.9 teraflops of Earth Simulator. And as expected, Columbia was clocked at 51.9 teraflops. IBM's new MareNostrum, at 20.5 teraflops, arrived in fourth place. The Top500 list is compiled by Hans Meuer of the University of Mannheim in Germany, Erich Strohmaier and Horst Simon of Lawrence Berkeley National Laboratory, and Jack Dongarra of the University of Tennessee. It was released Monday at the SC2004 supercomputing conference in Pittsburgh. Computers are scored by the Linpack benchmark, a convenient measurement but even list organizers recognize it is an incomplete measurement of real-world ability. Top-end supercomputers use thousands of processors, draw power measured in megawatts and occupy entire floors of buildings. But less and less often are they based on exotic or specialized computing technology. A total of 236 of the 500 systems use the same Intel Xeon processors that show up in the majority of off-the-shelf servers. That's up from 226 in the June list. Systems using Intel Itanium processors, another mainstream alternative and the chip used in Columbia, increased from 61 to 87 over the same period. Thirty systems use Advanced Micro Devices' Opteron--another common server chip--compared to 32 in June. IBM's Power family of processors are used in 62 systems, down from 75 in June. Supercomputers assembled from these more ordinary components are connected into massive clusters that pass messages and share each others' memory over high-speed networks. Clusters now account for 296 of the 500 systems on the list. IBM and HP dominate the computer manufacturers on the list. IBM slipped from 224 systems in June to 216 systems, while HP grew from 140 to 173. In the entire supercomputing market, Big Blue is getting bigger and nearing HP's top rank. From 2002 to 2003, IBM's share of the market grew from 28.2 percent to 30.2 percent with sales of $1.62 billion, according to IDC. HP, with $1.79 billion in revenue for 2003, shrank slightly from 33.6 percent to 33.5 percent. A few years ago, a "terascale" computer--one that could perform more than a trillion calculations per second--was a remarkable achievement. Indeed, in 1993, the total performance of all 500 machines was 1.12 teraflops. Now, 398 of the systems on the list have crossed the teraflop threshold, and the total performance has reached 1.13 petaflops, or quadrillion calculations per second. The slowest machine on the list was 851 gigaflops, or 0.851 teraflops. Japan long has had a place on the list, but now other Asian countries are rising. Thirty of the top machines were in Japan and 57 in other Asian countries, including 17 in China. Managing Your Information Content (part of ILM) ...posted 062105 by rdp The new face of business intelligence: dashboards and scorecards By Wayne Eckerson 20 Jun 2005 | SearchCIO.com In the last two years, business intelligence (BI) has finally reached critical mass. It's recognized by business leaders as a vital tool for helping organizations achieve strategic objectives, optimize performance and improve decisions and plans. As a result, executives are increasing BI budgets to provide employees with timely and relevant information. A major reason BI has become so attractive is the emergence of dashboards and scorecards. These tools allow companies to digest large volumes of information and deliver information in an intuitive format, making it easy to identify and respond to critical, time-sensitive events, as well as explore issues and trends without getting hopelessly lost in reams of data or reports. Earlier generation of query, reporting and analysis tools stymied widespread user adoption, and many BI tools have proven to be too complex for the average user. It turns out that most BI tools were designed for power users who spend most of their time sifting and crunching data, not casual users who use information as an auxiliary tool for doing their jobs. Thanks to dashboards and scorecards, these problems are dwindling. Dashboards and scorecards provide a layered interface that conforms to the way users work rather than forces them to conform to the way the tools work. Like peeling an onion, users move through successive layers of information in a carefully guided and systematic manner, uncovering critical information on an as-needed basis. A study by The Data Warehousing Institute shows that 51% of organizations already use a dashboard or scorecard and another 17% are currently developing one. The study also shows that among organizations that already have a dashboard or scorecard, almost one-third use it as their primary application for reporting and analyzing data. (See chart.) Does your organization use a dashboard or scorecard? Yes: 51% No: 32.5% Under development: 16.5% Base: 473 business intelligence professionals. Source: Wayne Eckerson, "Strategies for Developing Analytic Applications" (TDWI Report Series, The Data Warehousing Institute, 2005.) Dashboards and scorecards are visual display mechanisms within a performance management system. Many people use the terms interchangeably, but the primary difference is that dashboards are used to monitor the performance of operational processes while scorecards are used to chart the progress of tactical and strategic goals. (See chart -- "Comparing features.") In the end, it doesn't matter whether you use the term dashboard or scorecard as long as the tool helps organizations get the information they need. Both dashboards and scorecards must display critical performance information on a single screen so users can monitor them at a glance and then drill down to view more information as needed. For this reason, I use the term "performance dashboards" to describe any application that uses a dashboard or scorecard interface. Performance dashboards conform to the way users work, make it easy to digest critical information at a glance, and then explore and find the root causes of problems. As a result, performance dashboards resonate with users, especially executives and managers who have been searching for an intuitive way to measure, monitor and manage their organizations and groups.
Source: Wayne Eckerson, "Performance Dashboards: Measuring, Monitoring, and Managing Your Business" John Wiley & Sons, October 2005.) Performance dashboards deliver the following three layers of information that let users navigate from summary-level views of critical data to detailed views in a structured manner: Summary view: The top layer provides a customized view of a few critical metrics -- or key performance indicators -- tailored to each user's role in the organization and the processes they manage. This layer charts individual or group performance against targets and thresholds defined by managers in annual planning sessions or annual trends. Color-coding and symbols alert workers when performance is above or below targets and needs attention. In essence, this layer is where users monitor information. The dashboard, scorecard or portal interface essentially becomes a graphical exception report. Multidimensional view: Often, the first step in responding to an alert is to gather additional information. To do this, a user simply clicks on the metric or chart to drill into the next layer in the dashboard or scorecard, which lets users explore issues and trends in a free-form manner. This layer provides multidimensional analysis tools that enable users to navigate through the data by dimensions (e.g., customer, geography and time) and hierarchies (e.g., country, region, city), usually in a prescribed manner defined upfront by IT administrators. More colloquially, these point-and-click tools let users "slice and dice," "drill down or up," or "pivot" the data to view exceptions and trends from any perspective they want. Transactional view: The bottom layer lets users view transaction data, such as invoices, shipments or trades, which may be stored in a data warehouse or the operational system that captured it. Users often need such data to understand the root cause of a problem, such as missing or incomplete orders or a salesperson who has been sick. This layer either connects users to predefined operational reports or dynamically generates a SQL query to pull data out the operational system or data warehouse where it stored. As the new face of BI, performance dashboards are transforming BI from a departmental activity spearheaded by power users to an enterprise resource leveraged by all workers. Performance dashboards are fulfilling the long-standing promise of BI to help organizations work smarter, optimize performance and achieve strategic objectives. Wayne Eckerson is the director of research and services for The Data Warehousing Institute, a worldwide association of business intelligence and data warehousing professionals. He has covered data warehousing and business intelligence since 1995 and is the author of the book, Performance Dashboards: Measuring, Monitoring, and Managing Your Business, which will be published by John Wiley & Sons in October 2005. He can be reached at weckerson@tdwi.org. Open Source Java ...posted 060206 by rdpearson This is a Presentation suggestion from Robert Pearson > This question came up on another list and I was curious what some of I agree with Eric and the other positive replies to your post. Eric Boutilier's Sun Weblog at: "Q: Where's my Java? A: From now on, it's Just There." [e2eiod comment...]
http://www.intelligententerprise.com/channels/appmanagement/showArticle.jhtml?articleID=188101077
Very interesting... I agree with Chris' statement for: "NewI/O Project Justification A web browser is basically an Internet Document Browser (or IDB). It was not designed to run programs across the Internet. For example, it does not keep track of program state (Read the NewI\O Blog for technical discussions about the current state of the art). Existing systems appear to me to be kludges to get the document browser to do something for which it was not designed. What we need is a browser designed from the ground up to run programs across the Internet. What we need is an Internet Application Browser (IAB). We need something with the remote display capabilities like X-Windows, but with additional multi-media capabilities, wrapped in something with the ease of use of a web browser. There is an opposing view to NewI/O posted at Eric Boutilier's Weblog: "Ajax/Java and Databases I must say, Google calendar is quite impressive. Maybe the Ajax phenomenon will finally solve my seemingly never-ending search for The One True Calendar: A calendar that not only am I motivated to use (radically fast, minimal-click item entry) but one that I can also get the rest of my family to use. Just think! we'd post and merge all our events, large and small, into one gloriously perfect, ubiquitously accessible, yet secure, MetaBoutilierCalendar.net. (16-year-old got a big Spanish test tomorrow? Bing! goes the alarm on my computer screen at 8:00 the night before prompting me to shout, "Joe, time to `BBL' your IM buddies so we can review the subjunctive!") Anyway, what's this got to do with a UNIX/FOSS blog? Well being a former systems engineer, I also can't help wondering about the databases that are running behind the scenes of all these nifty new "Ajaxian" applications. Which is why these two recent articles really caught my eye:
(By the way, I have no idea what database Google calendar uses.) " Comment: Alternative to NewI/O Eric, nice to see you are as enthusiastic as I am about Google Calendar! Regarding Ajax, databases and my post... well, I don't know for sure what Google Calendar uses, but in the case of most Ajax apps out there, behind all the client-side magic sits a classic CGI, or servlet, or what have you using a regular DB. I guess the kind of DB connectivity I wrote about is still in its infancy--and all the more exciting for that. Andrea Posted by Andrea Campi (81.208.74.189) on May 10, 2006 at 06:39 PM CDT If wishes were horses... |
||||||||||||||||||||||||||||||||||||
| Recent Changes Printable View Page History Edit Page | ||||||||||||||||||||||||||||||||||||
