The Myths Of Network Utilization

The Interpath Technologies Networking Mythsinclude night-time hours? If so, your metrics are
Series (TM)already mostly useless (unless you have equal activity
This artilce is also availabe as a Podcast of "Theat night as during the day).
Sniffer Guy" though iTunes.If it is hourly--and day time only hours--it still means
Mythology is not fantasy or lies. It describes a basicthat 4 out of 24 business hour polls are 90 seconds.
truth-but as metaphor. If you understand that it isThat does NOT mean 4 times--it means 4 times
describing a fundamental reality-by telling a story-you"POLLED." What about in-between polls? It is
know to look for the reality behind the story. If youhappening many more times than the 4--that were
believe the story-you will chase your tail until youable to be seen!
decide to ignore everything. So, when I say thatIt could be happening more than enough to
automated testing and metrics like utilization aresignificantly damage production and good will.
myths, I mean that they are metaphors-stories thatIs the logic in your automated tool really taking this
can direct you to the truth. They are not the truth ininto account? If so, what does it really tell you
themselves.anyway? It can't say why--and it can't say how. It's
What is the IT person's mission? The mission is toa fire alarm that only rings if you pull it at the right
keep a set of instructions (apps) interfacing withtime and frequently enough.
humans (users), over an enormously complexThat discussion opens the door to wondering about
environment (enterprise networks and the Internetmetrics in general.
itself). Many senior and seasoned IT professionalsUTILIZATION METRICS:
have developed a sense over the years thatWhat does it mean when a tool tells you that your
something isn't quite right. There is too much "hoping"network/segment/WAN is 42% utilized? Does it
going on and not enough "knowing."mean that the wire is only used 42%? No.
An IT support person is attempting to visualize theIf you are measuring a transport medium, that
transport of discrete units of data flowing across anmedium is utilized 100%--whenever it is used. If you
"Interpath" between all the applicableput an electrical current on a wire-can you use only
components--for hundreds of applications and42% of that wire? No. It is either hot or cold. On or
thousands of users. How is that done? Black magicoff.
mostly--and a bag of black box tools.Does it mean that you only used it 42% of the time?
Often these tools are purchased by less technicalNo.
management and the support personnel receive littleIt is getting closer though. To say that it means that
training. Even if they do receive training, that trainingthe segment is 100% utilized-but only 42% of the
is focused on how to manipulate the console andtime-is ALMOST correct. Time is the key word here.
install the product. Seldom is there much rawHow is it measuring time? Is it always monitoring or is
technology exchanged. Let's face it, most of theit polling? It is polling--probably.
time we don't really know what the tool is(Sorry--polling and metrics are always hand-in-hand).
doing--only how to run it--yet we take their resultsWhat is your polling interval? Every 10 minutes-every
as law.hour-once per day? How long do you monitor
Tools poll and push and query and ask a selection ofbetween polls? Is it a "continuous" value or is it only
the components involved to report in aboutchecked on every hour? Is it an interval that the
themselves. Then, by algorithms known to only amanufacturer has never even disclosed or
few people at the vendor's office, (if they are stillacknowledged?
employed there), the data is SUMMARIZED andFor the sake of discussion--let's answer these
presented to you, the IT professional who hasquestions.
received that week of training.Imagine we have a tool that monitors a wire for 1
Once you think about the fact that the information isminute every hour. That tool reports that your
just a summarization--the problem--and"Utilization" is 42%.
mythology--begins to come clear.That means that for the one minute every
AUTOMATED TESTING:hour--that your wire is monitored--for that one
Picture a business user sitting at a PC workstation inminute ONLY-- it is 100% utilized. In other
a remote office of your company. Many times awords--42% of those polls showed 100% (rather
day--a network transaction that should take 15than 0%) utilization-for 1 minute per hour.
seconds--instead takes 90 seconds--maybe longer.Does that mean that it has 58% more room and all is
You have an automated tool that "simulates" thatwell? No.
transaction every hour. In 24 polls we see theSo, what do you know about the utilization of your
following:wire that can help you? Not as much as the
* 4 Samples show 10 seconds (much faster thanmonitoring tool's manufacturer would like you to think.
normal)Are you attempting to reconcile the results of two
* 16 Samples show 15 seconds (expected)different tools? That is extremely
* 4 Samples show 90 seconds (long enough to drivedifficult--impractical--and possibly ruinous.
a user crazy.)How much money are you spending-not only on the
* The average is 26.7 seconds.tool itself-but on decisions made as a result of that
Would an alarm sound in your monitoring system?tool's reports? However, this doesn't mean that there
With an expectation of 15 seconds--would you beis no point to those metrics. They are signpost
concerned enough to escalate this issue? Sure, it'spointing towards reality--fire alarms. Nevertheless, you
nearly double what is expected--but there are timesshould certainly not quote them when reality seems
that are quite excellent and anyway--it's still not evento imply different things. Reality is going to be more
30 seconds! Maybe it's not so critical...accurate that such metrics-every time.
It is Critical--to that user experiencing 90 secondThis is the reason for the Interpath Transactional
delays from time to time--or more frequently!Analysis approach to monitoring and reality based
What is your Polling Interval? Hourly? -- Does thattesting.