• Add a Comment
  • Edit
  • More Actions v
  • Quarantine this Entry

Comments (6)

1 localhost commented Permalink

Hi Dave,I too have worked in my back yard this past weekend. What a sore back! Weak spots in the infrastructure i.e. bad circuits, faulty equipment and other layer 1 problems need to be identified. This usually requires labor to be sent to the site. So, you are correct, self-awareness is important. As we all know once a system is aware of a problem it must be fixed. Self-healing is the goal. Maybe some day AI bio-software can help me re-seed my lawn. Now that would be self-healing!Bernie CiongoliTech Labshttp://www.techlabsinc.com/dynatrax.html

2 localhost commented Permalink

Hi Bernie, Thanks for the response and sorry about your back:(I agree, a system needs to be aware of layer 1 problems in order to manage in an automated fashion. What are the basic ingredients to be successful?(along with the great products from Tech Labs of course) The autonomic MAPE loop, along with the autonomic standards and integrating technology, provides the necessary guidance for creating self-managed systems - Monitor, Analyze, Plan, Execute. You need all in the right balance to be successful. - Just like lawns - you need the right balance of moisture, sunlight, and nutrients. Now, what would it take for an IT infrastructure to be aware of itself being aware of layer 1 problems? would this constitute self-awareness?

3 localhost commented Permalink

Hi Dave, your mention of policies is intriguing. I have always been somewhat sceptical of the neurological metaphor for Autonomic Computing. But policies seems to be to be a really interesting way to view computing systems. I.e. a computing system can be controlled by a policy language that is much more English-like than the lower-level programming languages. That could be very powerful when applied to autonomic computing systems.Posssibly policies could even be self-generated in the future. It seems to me that Policy-based autonomic systems coupled to symptom databases that stores the most common system failure symptoms and known remedies that can be applied autonomically could revolutionize business computing. What are your thoughts?

4 localhost commented Permalink

Hi Dave, Lennart,That’s a tough one. But you got me thinking.“SELF-AWARENESS” is defined as the capacity to become the object of one’s attention. The individual whose attention is self-directed is said to be in a state of self-awareness, that is in a state in which he or she can observe his or her own characteristics and behaviors.Now with that said, self-awareness must be a closed loop with artificial intelligence.When the infrastructure's NMS has feedback of a problem, and when the problem is sensed, correction is made. This is only a response to pre-identified policies not self-awareness. The system must have some sort of AI.Can a system learn to observe its own behavior? Maybe an algorithm, which can determine the best solution path, can be embedded into the NMS?What do others think?Bernard CiongoliTech Labs/DynaTraX

5 localhost commented Permalink

Maybe it's just my daily experience with a toddler, but I think that the best approach to self-awareness is to learn the rules based on a combination of instruction and experience. To stay with the horticultural theme, this weekend I tried to teach my daughter to pick roses without touching the bush. She learned how to pick the flower by watching me (instruction), but she learned why not to touch the bush by getting stuck (experience). I think the first step is to create systems that can learn the rules, and act on them. Lennart, the notion of a policy engine tied to a symptoms db is very intriguing to Singlestep, and definitely fits in with the instruction/experience model. But even this concept needs to be primed by human experience and market need. Where do the policies and rule sets exist today? For most companies, they're in the heads of their IT Staffers, or an IT outsourcing partner. So first, we've got to find better ways to capture existing knowledge as policy.Getting a product in the hands of IT service providers that helps them to document policy/response, and then automatically trigger reaction...that may be the next step towards revolution.Chris Noblehttp://www.singlestep.com

6 localhost commented Permalink

yes, kids are really amazing to observe... We really should model more from children than advanced intellect as we build more self-awareness into technology. Afterall, wouldn't you agree self-aware IT infrastructures are really in their infancy? When we first moved overseas my kids were quite young and it was immediately obvious(and humbling) that their ability or capacity to rapidly learn new languages was far superior to us older folk. Some studies suggest the brain is more adept at learning language at the early stages of our life(I found that study somewhat comforting) but I also observed an incredible tolerance or capacity for trial and error that seems to steadily decrease with age. It would seem therefore that we need to build such tolerance into autonomic self-managing technologies. Mechanisms that can tolerate numerous trial and error responses in nanoseconds could quickly build into intelligent self-managing responses...That said, I can't wait to put my earplugs in, hop on my John Deere and turn my well fed and watered, out of control lawn, back into a well manicured fairway tonight...

Add a Comment Add a Comment