On Nov 2, 2008, at 7:49 PM, Xiaofan Chen wrote:

> So it seems to me that determinism is measured with the
> jitter (variation of the latency). Is this correct?

I don't think so.  It's like a "hard real time" operating system.   
"Determinism" means that you can guarantee packet delivery within a  
particular maximum time; latency and jitter UP TO that time are free  
to vary all over the place.

> "CAN has no particular advantage in latency --- it has one in  
> *determinism*, i.e. its latencies won't vary as wildly as those of  
> an Ethernet at the same throughput load relative to their respective  
> capacities."

Badly worded, I think.  Variation of latency is "jitter";  but it is  
the inability to guarantee a certain MAXIMUM latency, to put an upper  
bound on latency, that makes the network non-deterministic...

I remember being singly unimpressed back in the late 1980s, when BBN  
announced advances in their butterfly network router that guaranteed  
packet forwarding in less than one millisecond; at the time, cisco had  
begun shipping their first HW-assisted router, which would route on  
the order of 10000 pps.  More than ten times faster.  It wasn't till a  
couple years later that I understood how non-deterministic the cisco  
was, and how significant an accomplishment that 1ms guarantee was...

BillW

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist