Showing posts with label Distributed processing. Show all posts
Showing posts with label Distributed processing. Show all posts

2/27/2012

Porsche 911 concept car makes debut at Mobile World Congress

Here's another opportunity to see the Porsche 911 Carrera concept car in action. This time, it's from the floor of Mobile World Congress, where Derek Kuhn, QNX's VP of marketing, takes us on a tour of the car's coolest features.

Highlights include one-touch phone pairing, BlackBerry PlayBook integration, distributed media playback, high-def handsfree calls, and a dynamically reconfigurable digital instrument cluster. Roll the tape...


 

3/12/2010

30 years of QNX: Holy terabits, Batman!

If you’ve been following this blog series on the history of QNX, you’ll know that I’m still making my way through the 1980s. So I hope you don’t mind if I temporarily leap ahead 20 years — I was going to post this story when we got to 2004, but something happened last week that made me change my mind.

Confused yet? Well, brace yourself. Because this story really begins in 1998.

In 1998 QNX Software Systems released a self-booting demo that squeezed the QNX operating system, a windowing system, a web browser, a web server, some games, and lots of other goodies onto a single 1.44M floppy. Millions of people took this demo for a test drive, including a developer from Cisco who was impressed by the operating system’s elegance and efficiency.

So did he immediately tell his colleagues at Cisco to look into this cool OS from QNX? Nope, not a chance. Because at the time, the QNX OS worked only on x86 systems.

Little did he know that QNX engineers were working on a spanking new version of the QNX OS — a version that could run on multiple processor architectures, including those used in Cisco’s networking equipment.

Fortunately, he expressed his frustration to an acquaintance at QNX Software Systems, who immediately told him about the new OS. One thing led to another, and on May 18, 1999, QNX Software Systems announced it had been chosen as Cisco’s preferred realtime operating system vendor.

After that, things went quiet — until 2004. That’s when Cisco unveiled the CRS-1, the highest-capacity Internet router ever developed. How high? Enough to handle up to 92 trillion bits per second, which is equivalent to all the Internet traffic of a single mid-sized country. Or, as my colleague Andy Gryc likes to put it, enough to squirt out 18,000 CDs per second.

As you no doubt suspect, the CRS-1 is controlled by the QNX operating system. More specifically, the router’s IOS XR software, which manages hundreds of parallel processors, is based on a straight, unmodified QNX Neutrino microkernel and QNX transparent distributed processing.

So why am I talking about this now? Because Cisco has just unveiled the new CRS-3, which has triple the capacity of the CRS-1 — 322 trillion bits per second.

How much is that? Enough, according to Cisco, to download the printed collection of the Library of Congress in just over one second. (One assumes that the collection would have to be digitized first.)

Considering that my web browser has a hard time downloading a single web page in one second, I’m impressed. Holy terabits, indeed.

8/06/2009

QNX chosen to power GE’s flagship Mark VIe control platform

It controls gas turbines, steam turbines, hydro plants, and nuclear plants, including the largest power and water facility on the planet. It runs in hundreds of installations worldwide — from Cairo to Quebec to Los Angeles — delivering electric power to homes, hospitals, factories, and businesses. It’s called the GE Mark VIe, and GE Energy promotes it as the company's "most advanced control platform."

The heart of the GE Mark VIe is a single-board controller based on the QNX Neutrino RTOS and the Freescale PowerQUICC 8349 processor. Designed for high-uptime operation, the Mark VIe supports:
  • a variety of redundancy options, including dual or triple controllers, dual or triple power sources, and dual or triple networks


  • precision diagnostics to minimize mean-time-to-repair


  • on-line repair to maximize the mean-time-between-forced-outages

Moreover, each QNX-based controller can communicate with up to three network switches. This allows each controller to monitor multiple redundant inputs and use a voting system to determine which input values are correct.

The Mark VIe can handle rugged environments, with a rated temperature range of 0 to 65°C (32 to 149 °F). It doesn’t require cooling fans, even at maximum temperature.

According to GE, the Mark VIe addresses a wide range of applications, from small governors to plant DCS applications. These include heavy-duty gas-turbine retrofits, aeroderivative engine-control retrofits, steam-turbine retrofits, and hydro control.

I can’t begin to do justice to the Mark VIe’s many features. So to learn more about it, click here. To download the data sheet, click here. And to learn about some of the applications that use the Mark VIe, check out these links:

4/11/2008

Rx for oil-rig blowouts

Got 15 seconds? Then check out this video of a oil-rig blowout — the action starts at exactly 00:15:



Fortunately, no one got hurt in the incident. And just as fortunately, real-time control devices called blowout preventers (BOPs) are making blowouts a thing of the past. This week, Control Engineering magazine profiled a BOP based on the QNX Neutrino RTOS and the McObject high availability database. It’s a pretty cool system, but more important, it keeps oil rigs from bursting into incendiary versions of Old Faithful.

Oil rigs are surrounded by saltwater. Lots of it. Salt plays havoc with electronics, so this is one system where redundant, standby controllers come in handy. To achieve this redundancy, the BOP relies on QNX Neutrino transparent distributed processing (TDP), which can merge any number of real-time control systems into a seamless, peer-to-peer network. Together with the McObject database, QNX TDP ensures that the primary and standby controllers share the same, up-to-date view of the system.

This application is a shoe-in for my list of all-time life-saving QNX-based systems.

2/04/2008

Getting a leg up on robotic design

I was tempted to buy a Roomba robovac last week, until my wife warned me of the pandemonium that would ensue when our schnauzers got a hold of it. So while I can’t entertain you with a cute video of my dogs attacking a robotic vacuum cleaner, how about the next best thing: a giant robotic leg powered by QNX.

This week, Carnegie Mellon’s student newspaper posted a story on Jonathan Hurst, a grad student who plans to develop a six-legged robot that can walk, run, and even climb stairs. Already, Hurst has designed an innovative leg that uses fiberglass springs to emulate natural running movements — a departure from the rotating gear motors and pneumatic actuators of traditional robots.

For his project, Hurst is using software from RHex, a QNX-based robot that is the “first documented autonomous legged machine to have exhibited general mobility... over general terrain.” RHex also serves as the basis for AQUA, a project dedicated to building amphibious robots that work underwater. Click here to see the RHex and AQUA robots in action.

If you search the web, you’ll find lots of other robotic projects based on QNX, including the famous Cog and Kismet robots from MIT. In most cases, these projects use QNX because of its realtime capabilities. However, the researchers at MIT needed to solve an additional problem: implementing efficient interprocess communications among Cog’s 32 processors. They found the solution in QNX transparent distributed processing (TDP), which allows an application to access remote software and hardware resources without special software coding. Recently, QNX published the source code for TDP on Foundry27.