Mailing List Archive

Mailing List: techdiver

Message Display

In all the talks about failures, I think people are talking about two
different things...

Rich uses examples of, say, putting a Y-valve on a tank and attaching two
complete regulator systems to the valves.  The complexity has doubled (two
regulators) but he asserts that the chances of a life threatening emergency
has decreased.

There are several problems with this that makes the statement both true and
false...

With the regulator example, if one regulator fails it may or may not affect
the other regulator.  Thus, you do NOT have two separate systems.  For
example, assume a blown hose in one regulator.  The other regulator, in a
Y-Valve system, is most assuredly affected.

In addition, with two regulators, the chances of _a_ regulator failure has
doubled.

Lastly, there is still a single point of failure - the Y-Valve.  Should the
Y-Valve fail, all the redundancy in the world won't help you.

Now, that's only part of the story.  A true analysis of whether the
redundant system has to take into account the sum of the probabilities of
_all_ failure modes of interest in both the redundant and non-redundant
system.  Only after summing the probabilities can one state that System A
is more or less likely to fail than System B.

The short of it, then, is that there is no intuitively obvious answer to
the question, "is Redundant weebles more or less likely to wobble than
non-redundant weebles?"

      -- Kevin --

"I'll get around to it today --
          For sufficiently large values of 'today'"

Navigate by Author: [Previous] [Next] [Author Search Index]
Navigate by Subject: [Previous] [Next] [Subject Search Index]

[Send Reply] [Send Message with New Topic]

[Search Selection] [Mailing List Home] [Home]