News Archive (1999-2012) | 2013-current at LinuxGizmos | Current Tech News Portal |    About   

Red Hat invests in chip maker and launches real time processing initiative

Jan 29, 2001 — by LinuxDevices Staff — from the LinuxDevices Archive — views

Fremont, CA — (press release excerpt) — Cradle Technologies, Inc. today announced strategic investments from CIBC Capital Partners, Charter Ventures, Mingly Ventures, Ltd, Red Hat, Inc., Signal Lake Ventures II LLP, and Smart Technology Ventures. The current round of $20 million brings the total invested in Cradle to date to $34 million. Additionally, Cradle disclosed that it is working with Red Hat to launch an Open Source Initiative for Real Time Embedded Processing based on Cradle's “stream processing” silicon platform.

“Cradle's revolutionary processor architecture represents a breakthrough for critical hardware applications that are key to both convergence and the Internet's expansion,” commented Matthew Szulik, president and CEO of Red Hat. “System developers will be able to achieve designs for complex, high-performance real-time processing applications entirely in software, without resorting to complex and time consuming chip designs. We expect the efficiencies of this new design process to be substantially enhanced by the extension of our Red Hat development tools and the open source model to this unique, new platform.”

Stream processing, the real-time processing of many parallel streams of data, is a characteristic requirement of almost all contemporary multimedia, communications and convergence products. “Platform” is a term used to describe a microprocessor chip — such as the one inside every personal computer — which is so versatile in application that it can serve as the fundamental building block — or platform — upon which hundreds to thousands of useful software and hardware applications can be built. Platforms are supported by sophisticated software development tools — typically designed by third parties such as Red Hat — and a wealth of software applications and utilities. Cradle's architecture represents the first true stream processing platform ever to be announced.

For a new platform, this open source “effect” can be an important accelerator to widespread adoption. According to Satish Gupta, Cradle president and CEO, this was a key consideration in Cradle's alignment with Red Hat. “In the future a broad range of applications will be defined chiefly in software running on a stream processor, rather than on hardware designs such as custom chips or 'ASICs' [application-specific integrated circuits]. This will attract tens of thousands of talented software engineers to participate for the first time in the 'hottest' areas of technology. Having critical pieces of intellectual property potentially available in an open forum will dramatically accelerate this process.”

According to Gupta, such applications include digital video cameras, editing equipment and digital content severs, enterprise network routers, home gateways, broadcast and telephony equipment, HDTV receivers and advanced set top boxes, cellular phones and base stations, telephony equipment, and even products such as printers, copiers, faxes, and PDAs. The types of software he expects to be most popular in such an open forum include MPEG codecs — such as those necessary for direct broadcast satellite receivers, Open GL engines, TCP/IP and ATM software stacks — crucial to virtually every class of networking equipment, and hardware interfaces such as ADSL PCI, IEEE 1394 (Firewire), Ethernet MAC, CCIR 601/656 and Utopia.

Open Source Initiative for Real Time Processing

The Open Source Initiative based on the stream processing platform will be launched by Cradle and Red Hat later this spring. The initiative will be focused on creating a large on-line and royalty-free library of system-level functions and interfaces which will be open for the world to contribute, modify, and use. This new “library of open intellectual property” allows original equipment manufacturers (OEMs) to instantly download standard functions, experiment with product concepts, and then tailor the functions to meet their exact needs, all in a continuous and efficient fashion. As a result, OEMs realize orders of magnitude time to market gains, dramatically lower product development costs, and better quality products.

Traditionally, OEMs develop these functions and interfaces by designing custom or semi-custom chips, purchasing off the shelf chips and using “glue logic” to fit them to their specific needs, or by licensing code and paying royalties to third parties to incorporate the functions.

According to Red Hat's Michael Tiemann, CTO, “The creation of the library, enabled by Cradle's platform, obsoletes inefficient and costly hardware based approaches. Moreover, shifting the functions from a proprietary to an open environment further reduces any unnecessary cost burdens placed on bill-of-materials as a result of royalty charges and also eliminates dependencies on third parties to maintain key pieces of the product design. For the first time, the OEM is in total control.”

Cradle and the Universal Microsystem Platform

The Universal Microsystem, or UMS, invented by Cradle, was designed to join the classic microprocessor, microcontroller, and digital signal processor chip families as the fourth — and newest — single-chip computing platform upon which a significantly wide range of future electronic devices will be built.

The architecture of the UMS differs from its predecessors — chips like the familiar Pentium or PowerPC — in that it contains dozens of small, fast processors that all work in parallel. This strategy is ideal for the internal computing required by the coming generations of imaging, multimedia, digital video, communications, navigation, and high-speed networking products. A particular feature of the UMS is that it allows highly complicated applications to be created predominantly in software — rather than with custom chips or expensive combinations of conventional processors. The development of complex applications in software rather than hardware yields tremendous economies in product design. Additionally, the UMS offers nearly an order-of-magnitude performance improvement in terms of computing speed per dollar over previous platforms. This step-function improvement in performance suddenly makes fanciful products such as a $200 handheld DVD player attainable.

 
This article was originally published on LinuxDevices.com and has been donated to the open source community by QuinStreet Inc. Please visit LinuxToday.com for up-to-date news and articles about Linux and open source.



Comments are closed.