Alime.230
net.lan
utzoo!decvax!harpo!floyd!vax135!lime!martin
Fri Apr 23 09:23:53 1982
IBM local net

From: lime!vax135!houxi!ihnss!ihuxs!steffen

Reliability questions on the token-passing ring network

	1) What happens when the master machine goes down?
	
	2) What happens when two machines in the ring go down?
	
	3) What verifies that each machine's sanity timer works?
	
	4) How do you add a machine to the ring without breaking the ring
	and becoming vulnerable to other machines' failures?
	
You can solve most of these problems with duplex, reliable machines like
telephony processors; but IBM has simplex, unreliable processors.  A final
note: I'll trade reliability for efficiency any day.

				Joe Steffen
				Bell Labs, Indian Hill
				(312) 979-5381
				UUCPnet: ucbvax!ihnss!ihuxs!steffen