Network World's Network/Systems Management Newsletter, 02/05/07
Management software or Magic 8-Ball?
By Denise Dubie
A couple weeks back on an especially stressful Sunday in January, I picked up my Magic-8 Ball post half-time of one NFL Playoff game and desperately asked, "Will the Patriots make it to the Super Bowl?" It's cryptic answer: "Outlook not so good."
Being from New England, I didnít bother to ask my handheld psychic about the Bears and the Colts match (yes I am that kind of football fan), but I also knew it couldn't tell me Super Bowl results based only on the fact that the Patriots lost. That experience coupled with a few recent talks with management vendors reminded me of my desire to gain an edge on future events with some predictive analysis. I'm sure it depends on the network manager, but I'd bet getting an advance take on how their networks will handle certain peak loads, deliver services to more end users or respond to new applications is just as comforting as my picking up the Magic 8-Ball or dialing Miss Cleo's hotline.
Last week, two vendors in particular prompted me to think more on predictive analysis. To start, Integrien updated its Alive software with capabilities to speed problem diagnosis and offer known fixes to problems. Alive 5.5, introduced at last week's Demo '07 conference, works to learn the path an application takes across the infrastructure and determine a baseline of normal behavior to alert network managers when someone is amiss.
Separately, eG Innovations in February will release an updated version of its flagship service management software. eG Enterprise Suite 3.3 operates either with agents or without and installs on the server to be managed. It can also learn from constant monitoring what normal behavior is and then alert IT managers to anomalies.
The value of this type of feature -- also available in products from Netuitive among others -- is that network managers get a heads up to at least be on the look out for problems. From what I understand, predictive analysis tools can tell you that a CPU is maxing out its utilization, that a server is lagging in terms of response time or that an application is chatting back and forth needlessly with one server -- all valuable information to an IT team working to keep systems optimized and services flowing. What the technology can't say definitively is what exact service outage or performance degradation will happen because of such conditions.
Just like my Magic 8-Ball query didn't tell me who would win the Super Bowl
if the Patriots lost the playoff game, predictive analysis can't see the exact
events yet to be played out. Instead it can communicate likely outcomes as a
result of certain events. Most important, the technology can warn network
managers to make concessions for failing elements in their IT infrastructure
before they have to learn what might have been. And if vendors or IT managers
integrate tools such as service modeling and topology mapping, management
software could be just what network managers need to connect the dots and
determine that if Server A fails, potentially Application X won't be able to
perform as expected for End-user Group 3.
Questions or problems regarding this web site should be directed to firstname.lastname@example.org.
Copyright © 2008 Art Beckman. All rights reserved.
Last Modified: March 9, 2008