In the article
The Cognitive Net Is Coming published in the recent IEEE Spectrum, the author (Antonio Liotta from the Eindhoven University of Technology) states:
Perhaps as early as the end of this decade, our
refrigerators will e-mail us grocery lists. Our doctors will update our
prescriptions using data beamed from tiny monitors attached to our
bodies. And our alarm clocks will tell our curtains when to open and our
coffeemakers when to start the morning brew. (
I would say, coffeemaker and curtains should be smart enough - if they are not yet - to know when to start brewing or when to open. Rather than be hard programmed by time, they should sense when these functions are needed).
By 2020, according to forecasts from Cisco Systems, the global Internet
will consist of 50 billion connected tags, televisions, cars, kitchen
appliances, surveillance cameras, smartphones, utility meters, and
what not. This is the
Internet of Things, and what an idyllic concept it is.
But here’s the harsh reality he says: Without a radical overhaul to its
underpinnings, such a massive, variable network will likely create more
problems than it proposes to solve. The reason? Today’s Internet just
isn’t equipped to manage the kind of traffic that billions more nodes
and diverse applications will surely bring.
Then the author proceeds to devise a more intelligent (cognitive) Internet protocol which would presumably solve the problems of today's global networks. Not going into the details of the
cognitive protocol we may find some useful (even though not necessarily totally new) idea to
endow
every connected computer/processor equipped device with the ability to route data. Given the
computational capabilities of today’s consumer devices, there’s no
reason for neighboring smart gadgets to communicate over the core
network. They could instead use any available wireless technology, such
as Wi-Fi or Bluetooth, to spontaneously form “
mesh networks.”
This would make it possible for any terminal that taps into the access
network—tablet, television, thermostat, tractor, toaster, toothbrush,
you name it—to relay data packets on behalf of any other terminal.
By off-loading local traffic from the Internet, mesh networks would free
up bandwidth for long-distance services, such as IPTV, that would
otherwise require costly infrastructure upgrades. These networks would
also add routing pathways that bypass bottlenecks.
To handle data and terminals of many different kinds, it is suggested that the routers
(including the terminals themselves) use methods for building
and selecting data pathways borrowed from a complex network that already exists in nature: the
human
autonomic nervous system.
Comparing a complex system to a human body is a popular metaphor. The human body system controls breathing, digestion, blood circulation, body heat,
the killing of pathogens, and many other bodily functions. It does all
of this, as the name suggests, autonomously—without our direction or
even our awareness. Most crucially, the autonomic nervous system can
detect disturbances and make adjustments before these disruptions turn
into life-threatening problems.
In fact, the parts of the brain that control this process rely on a multitude of inputs from many subsystems, including taste, smell,
memory, blood flow, hormone levels, muscle activity, and immune
responses. Does the food contain harmful bacteria that must be killed
or purged? Does the body need to conserve blood and fuel for more
important tasks, such as running from an enemy? By coordinating many
different organs and functions at once, the autonomic system keeps the
body running smoothly.
By contrast, many current systems (the Internet included) address a disturbance, such as a spike in
traffic or a failed node, only after it starts causing trouble. Routers,
servers, and computer terminals all try to fix the problem separately,
rather than work together. This often just makes the problem worse rather than to fix it.
One idea, proposed by IBM, is the Monitor-Analyze-Plan-Execute (
MAPE)
loop, or more simply, the
knowledge cycle. Algorithms that follow this
architecture must perform four main tasks:
First, they
monitor a router’s environment, such as its
battery level, its memory capacity, the type of traffic it’s seeing, the
number of nodes it’s connected to, and the bandwidth of those
connections.
Then the knowledge algorithms
analyze all that data. They use
statistical techniques to determine whether the inputs are typical and,
if they aren’t, whether the router can handle them.
Next, they
plan a response to any potential problem, such as
an incoming video stream that’s too large. For instance, they may figure
the best plan is to ask the video server to lower the stream’s bit
rate. Or they may find it’s better to break up the stream and work with
other nodes to spread the data over many different pathways.
Lastly, they
execute the plan. The execution commands may
modify the routing tables, tweak the queuing methods, reduce
transmission power, or select a different transmission channel, among
many possible actions.
An integrated building energy technology advanced system architecture (
ASPA) can utilize some of these principles above. For example, above mentioned curtains can react on the level of light and open when needed, but also close when the temperature in the room is raising due to excessive sunlight approaching the limit when air-conditioning needs to be turned on. In the hydronic systems, why do we need to maintain the maximum level of the temperature in the storage tank during the period of low or no demand for hot water - only to waste energy to the heat loss? Instead, the system can monitor the pattern of daily use and predict the demand, reducing the energy use to the minimum. Outside temperature follows the pattern but can be of course a subject of significant variations. A combination of knowledge-based algorithm with a feedback loop including adaptive capability would control the air circulation. An irradiation sensor can tell the solar thermal collector if it needs to adjust the flow rate in the system in order to either increase efficiency or prevent overheating.
An evolution of this system architecture, Advanced Sustainable Control Energy Network Technology, will be key to keeping the system in check. Not only will it help prevent individual components from
failing, but by monitoring data from neighboring nodes and relaying
commands, it will also create feedback loops within the local network.
In turn, these local loops swap information with other local networks,
thereby propagating useful information across the Sustainable Network.
