I've done an extensive review of LinuxBIOS recently, for potential 
use in IBM products.  I find there are several distinct advantages for 
clusters of the future in LinuxBIOS. Namely, LB provides: 
 
1) ultra fast boot 
2) terminal server support 
3) remote administration (boot device manipulation) 
4) device support in BIOS, particularly high speed interconnect 
 
These are many of the features that I proposed in the ClusterBIOS 
suggestion at the last SCIDAC meeting. The ClusterBIOS suggestion went a 
little further -- to provide an API for all these functions.  The advantage 
of the API is that cluster compliant nodes that meet the API could 
implement the API any way they want -- LinuxBIOS might be just a single 
instantiation of the API. 
 
 
So, Narayan and Al, from a planning perspective, we can either say that 
we're just going to use LinuxBIOS, or that we're going to push for an API, 
of which LinuxBIOS is one implementation.   In 3-5 years, an API might be 
possible, but it is certainly attractive to just work directly with the 
LinuxBIOS guys, I'm sure they'd be interested in making LinuxBIOS the 
standard for large clusters of the future... comments? 
 
thanks, Rich 
 
Richard Ferri 
Senior Programmer, IBM Linux Technology Center 
Poughkeepsie, NY 
rcferri@us.ibm.com 
845.433.7920