Group items tagged

Welcome to the first episode of Cloud Cover!  Join Ryan and Steve as they cover the Windows Azure platform, digging into features, discussing the latest news and announcements, and sharing tips and tricks. Follow and interact with us at @cloudcovershow In this episode: Learn about the Service Management API and how to use PowerShell cmdlets to manage your cloud services. Find out how to get started quickly on the Windows Azure platform. Other topics include: SQL Azure updates! Windows Azure Drives (XDrive).  Hear about some cool new Windows Azure storage management tools. Azure Reader architecture.

When Sven Groot was in town a while ago we dropped by Erik Meijer's world and got a look at what he and team have been and still are working on (thus there is no out-of-date property of this fun and insightful interview that is off-the-cuff as it gets: deep Channel 9 ). It's great that we were able to put a real live Niner into fire in one of Erik's team meetings. There is a great deal to learn here. Thank you, Sven, for being a real sport! Great stuff in here. Tune in!

A deluge of high-energy physics data is headed toward servers in Geneva, Switzerland, later this month. That's because the European Organization for Nuclear Research (CERN) now says it plans to restart its Large Hadron Collider (LHC) soon for a run that could last as long as two years at a collision energy of seven TeV (tera–electron volts, 3.5 TeV per beam). As CERN ramps up the world's most powerful particle accelerator to operate well beyond its previous best performance, the lab's computer systems must likewise be tuned so they can properly capture and analyze all of this new output.

The MATLAB team at MathWorks tested performance scaling of the backslash ("\") matrix division operator to solve for x in the equation A*x = b. In their testing, matrix A occupies far more memory (290 GB) than is available in a single high-end desktop machine—typically a quad core processor with 4-8 GB of RAM, supplying approximately 20 Gigaflops. Therefore, they spread the calculation across machines. In order to solve linear systems of equations they need to be able to access all of the elements of the array even when the array is spread across multiple machines. This problem requires significant amounts of network communication, memory access, and CPU power. They scaled up to a cluster in EC2, giving them the ability to work with larger arrays and to perform calculations at up to 1.3 Teraflops, a 60X improvement. They were able to do this without making any changes to the application code. Here's a graph showing the near-linear scalability of an EC2 cluster across a range of matrix sizes with corresponding increases in cluster size for MATLAB's parallel backslash operator:

What made me fall of my chair is the methodology/architecture part of this statement. It’s hard enough (but doable) to use RDF to map philosophically similar APIs. It’s a non-starter to use it to bridge architectural and methodological differences. I have spent a fair amount of time looking at Semantic Web technologies in the context of modeling IT systems (see the “semantic tech” category of this blog). While I think they would be a great foundation I don’t see them ever coming anywhere near what Reuven describes.