Going up – Linux for Lifts
But shouldn't it be embeds for all?
The Embedded Linux Consortium, formed recently as a vendor-neutral trade association is intended to provide a Linux-based alternative to Microsoft operating systems like Windows CE and embedded NT. Linus Torvalds has given ELC his blessing. The ELC was proposed by Rick Lehrbaum of LinuxDevices, who runs the PC/104 Consortium, and who will serve as interim chairman. More than 20 companies -- notably excluding Microsoft -- expressed their support and have pledged sufficient funds to launch ELC. Lehrbaum has a useful background paper on embedded Linux in the current issue of Dr Dobb's Journal, where he draws attention to the acceptability of soft-real time (SRT) Linux where millisecond responses are sufficient, as in many real-world situations. It would make a great deal of sense for ELC to include in its scope real-time OSes, since some key problems that confront the developers of such systems -- rapid hardware evolution, and the need for some standardisation for example -- are similar to those facing embedded-OSdevelopers. There is no universal agreement that Linux is necessarily the universal answer for RTOSes, or embedded systems come to that. And although small-footprint and "headless" operation appears to be required at present, this may not be the case in the future -- witness embedded NT. It appears ELC that their focus is really on reliable and efficient embedded systems, rather than Linux per se, and that in certain circumstances UNIX could be important. Maybe a better name for the organisation might be the Embedded and Real-Time Consortium, with a Linux Group within it. Such a solution could solve the problem of the ELC membership fee being a bit steep for small or one-person Linux outfits, by making it possible for them to have a lower-cost membership. By not insisting on Linux, ELC could encourage what Lehrbaum described as the non-Microsoft embedded OS developments to be part of ELC, rather than being a third alienated and fragmented group, especially as at the moment there are rather few embedded Linux developers. Although ELC is not intended to concerned with standards, it is desirable that reliability measures and certification be developed - and the expertise to achieve this is likely to be found largely in the ELC membership. After all, if such testing were not Linux-specific, Microsoft could be encouraged to submit its products for approval. ®
Sponsored: RAID: End of an era?