As a new player in operational intelligence, Machine Data takes Big Data into really small places, doing timely things we've been trying to do for a long time. Designers can use a pattern of its influence, to help imagine new solutions.
Anomaly detection and data imputation within time series
The Mind of a New Machine
1. The Mind of a
New Machine
Intelligent Operations
2. Solutions for Operational Intelligence automatically enable and enhance progress.
The goal of operational intelligence is to
make progress the normal outcome of
circumstances where information having
real-time relevance can drive the
awareness, initiation and method of a
successful action.
LEARNING
value
DOING
value
States
Boundaries
Procedures
Expectations
Options
Models
The information part is critical; and so is the technology. Machines can work quickly and
productively on understanding volumes of data and levels of detail that we otherwise
might not seriously consider. However, by telling them what we want, and having them
tell us what works and what doesn’t, we learn how to teach them to do more for us.
3. Having action automated is of course the most fundamental ambition of I.T., and we
have several generations of advancement in that regard.
Still, today, there is the equally important and often vexing question of how we know
what action to take is the “best” action; and the ambition to achieve that certainty is
the driving force of current advancements, aimed at practical real-time intelligence for
operations of any scale.
As part of that ambition, “computing” has always had the attraction of being about
“machines that know more than we do”… We rely on the fact that machines can and
will pay attention to things when we ourselves can’t or won’t. But then we want that
attention to guide practical action. There are at least four general kinds of results that
we’re happy to have provided for us by automation.
ControlOptimizationConnection RemediationAction Issues
Operational Intelligence: proactive supply and interpretation of key facts
4. More than ever before, machines can do the intelligence-based work. With the help of
contributions like “machine data”, we’re more likely to discover that:
• separate items and entities can logically interact in newly valuable ways;
• the elegance and performance of the interactions can both be maximized;
• the availability and behavior of requested efforts can be produced on-demand
within desired ranges of effect;
• and the complexity or complication created by unintended after-effects or risky
breaches can be confidently navigated and resolved back to desired conditions.
Part of our ambition is to have all four of those discoveries working together, with
automation continuously “realizing” (determining and executing) the where, when and
how of using what we know – or what the machines know.
ControlOptimizationConnection RemediationAction Issues
Operational Intelligence: having the right information at the right time
5. Intent Enabled: Where When How
Connection With other functionalities Just in time Communication
Optimization In context Before engagement Economical
Control Within circumstances Full-time Efficient
Remediation Underlying and on surface At whatever priority Permanent
As automation, a “capability” combines know-how with execution. Intelligence supports
the know-how; newer I.T. often concentrates on expanding or enhancing the sources of
the intelligence. More know-how expands the range of things that are “do-able”.
Operational intelligence can offer breakthroughs or unusual value in occasions that are
identifiable generically, as shown below. These value “propositions” make the impacts and
relevance of new I.T. capabilities measurably significant and noticeable.
ControlOptimization AbilitiesConnection Remediation
6. Challenges: Where When How
Detection Mobile, stationary, micro, macro Before, during, after Sensing
Description Localized, portable, generic, specific,
private, public
Implicit, explicit Recognizing
Validation Self, external Prescribed, ad hoc Testing
Instruction Local, remote Just-in-time Programming
Action On-demand, scheduled Solo, co-operative Performing
More specific value is also circumstantial. It may be notable initially to people whose roles
routinely bring certain situations to their attention. The contrast of manual versus
automatic methods of making circumstantial progress sets the stage for evaluating new
technology. Things are deemed advancements when automation solves “action” issues, to
allow the acting party’s intentions to be met at new scale, scope or types of events.
ControlOptimization Action IssuesConnection Remediation