It was good to read this press release from ARM today. They are planning to pull resources together in order to build an open source Linux platform that others can use to build ‘Connected Mobile Computing’ devices. This is a very similar approach to that which Intel is using with MIDs and MobLin and it can only be good for consumers. The full press release is here
CAMBRIDGE, UK, – Oct. 3, 2007 – To address the rise in consumer demand for access to the Internet and advanced applications on larger display mobile devices, ARM [(LSE: ARM); (Nasdaq: ARMHY)] and six companies today announced a collaboration that will result in the development of a Linux-based open source platform for next-generation mobile applications. The collaboration, announced at the fourth annual ARM® Developers’ Conference being held this week in Santa Clara, California, builds on the ARM architecture and its rich ecosystem of Partners to deliver a standards-base platform based on Linux. This group of companies are all working to accelerate the enablement of truly always on, connected mobile computing (CMC) devices.
Its not the first time someone has put Firefox on an Linux/ARM platform though. Pepper did it with their original Pepper Pad and decided to move to an x86 architecture for the following product. Either the software development costs were too high or they needed to get better performance. I suspect it was a combination of the two factors. This new development by ARM will help solve the first problem. A combined effort will benefit everyone but it won’t do anything for the current performance problem. As i said before, the mobile Internet experience needs to be compatible and quick and if either one of those elements fails, then the customers Internet experience fails.
There are better performing ARM cores now, its true, and even today ARM has announced Cortex A9 which they say will provide 4 times the performance of ARM11 within the same, 250mw power envelope. (I’m not quite sure what ‘envelope’ means in this context.) 4 times the power will certainly help and could actually be enough to provide the perfect handheld Internet experience but its going to take take time. Two years in fact until devices appear on the market and in that time, x86 cores will be moving into the sub 1-watt range. If you add the costs of software development and component costs (the most efficient x86 cores are going to costs way way more than ARM cores) its going to be an extremely tight race.
