Thursday, February 23, 2017

The complicated nature of confusing 5G

http://www.androidpolice.com/2017/02/10/what-is-5g-the-laymans-guide/

And Verizon is doing a 5G test pilot in these cities. 
Ann Arbor, Atlanta, Bernardsville, Brockton, Dallas, Denver, Houston, Miami, Sacramento, Seattle, and Washington, DC.

What that actually means and how Verizon defines it are unknown.  But 5G is interesting and challenging and exciting.  And coming. 

Monday, February 20, 2017

What are Smart Cities and how are they actually built?



  This youtube webinar goes over what is and what isn't connected in Cambridge, so it's a good primer to what might be the future of smart cities.  The bigger question becomes how do SMBs connect to this ever expanding smart world?


Friday, February 3, 2017

Dreaming of a connected Things kind of world with Analytics of Things on top of in


    Analytics encompasses a few categories. At least it's been that way for quite some time now. We've all read about analytics based upon Descriptive, Diagnostic, Predictive and Prescriptive. But how does it relate to Analytics of things and the Internet of things?

     IoT & AoT is a new old world that's changing everything.  It's also aiming for a 5G and connected everything world.  An always on and connected smart world.  But is it really smart? And how is machine learning and analytics going to help improve our lives?  We don't really need more technology to automate our lives and steal our jobs and money.  That's not a connected world. That's a poverty stricken dystopian society.

   Descriptive analytics has been around for ages, even before machines and computers took over.  What we call Descriptive now was just Business Intelligence and Data Warehousing operational reporting a few years ago.   And before that it was really just standard reporting.  
 
      Diagnostic is really a step up from Descriptive, but with more emphasis on analyzing some data to figure out  if you're losing money or the energy bill went up in the winter and summer months.   Many times it's based on some sort of regression model if it's complex or simple arithmetic aggregation and summation formulas if no form of data science is involved.
   
     Predictive is as it kind of sounds, it tries to predict the future based on historical data and trends. It's not always an exact science and a lot of it depends on the data collected and analyzed.  Before "big data' and NoSQL and Hadoop it was really based on a lot of bad sample data that didn't always create a clear picture of your past or future.
 
       Prescriptive is relatively new as it tries to not only to predict the future, but be far more proactive and make actual recommendations based on historical data, trends, and learning patterns.   Amazon shopping cart or Netflix is simple examples of recommendation engines, but that's just scratching the surface.

     Going forward machine learning and deeper learning wants to not only recommend, but basically diagnose, recommend, predict, and choose what you can and should do in a variety of domains and circumstances.   Hollywood has created worlds and movies like this.  Sadly though, most of these movies were dystopian horror shows and not a utopian peaceful and happy world.

     This leads me into Internet of Things (IoT) and Analytics of Things (AoT) and how the smart connected world is kind of like the movies, except wouldn't we rather have a utopian society vs a dystopian one?

     The Internet of Things is far more complex than people realize. If you listen to some consulting firms or companies pushing their own hardware sensors or software they will just sell you on the theory of Dump all the data into AWS or Azure or collect the data from sensors and put it into hadoop or spark.  And then run some diagnostics and predictions.   And the magic of IoT and AoT comes into view.  Except it's far more complicated than that.  Far far more complicated and people seem to be ignoring the reality and truth to these theories.

     IoT means connected everything.  Some manufacturers are building out sensors on their devices for everything. Whether it's a John Deere or Ford or GE doesn't really matter as they want to sell you on buying their devices to build out this IoT connected world.  But what does that actually mean?

     You can buy a few cheap sensors yourself, a rasberry Pi, download some open source software and create your own 'Smart home" and driveway where you get some text message or email that turns on your video camera system if somebody steps in front of your home or drives up your driveway.

     Is this IoT?  It kind of is, but all those devices need to be connected via blue tooth and wifi or hardwired.   And your phone must use some Telco service to get access to that text or email.   Somebody had to create an app or site so you can view that camera and maybe even speak to that stranger from afar.  Think of it as somebody knocking at your door while you're at work, but you can talk to them like you're home.  

     What kind of analytics does a normal person really need for a smart home?  Well homes and buildings are rarely like phones. And why would we want them to be?   It's really not a good idea to be an Apple home where every single thing in your house is bought from and subscribed to Apple or Google(well Alphabet) or Samsung or Tesla.   So you have a Samsung TV, a GE fridge, an Apple TV, an android Phone,  Whirlpool washers and dryers, a Ring doorbell system, a VW car,  some other Garage door opener,  and so on and so forth   All these devices should work together in theory. But we all know the enterprise space is filled with Microsoft products that don't exactly integrate or work with other Microsoft products.

     Now imagine adding Hadoop, Teradata, Oracle, HP, SalesForce, Vmware,  Cisco,  some local T1 and Internet provider,  some local security force, some company that maintains the elevators,  some 3rd party contractors, various lighting systems and air conditioners and heaters.   A smart building and a smart connected world is really as simple and as complex as it sounds.

     Even if we don't connect everything, connecting many things will require integration of hardware, software, sensors, and systems that have never been integrated before and many of these companies don't play nice with each other anyway.    And then building the analytics on top of all those different things.   There might be 20 different dashboards and KPIs that only matter to a few companies or individuals.  But in a IoT/AoT world, everything connected matters.  Except you just pay attention to what matters to you and finger point when the hardware or software isn't working.  That won't work though. People don't want to hear some rep blame somebody else.

    And lets not even get into how does the sensors or hardware hold up in Minnesota or Norway?   A mobile developer or web developer shouldn't care whether or not a sensor breaks down in freezing conditions.  Except a connected world means your mobile app is useless if the sensors can't survive under 32 degrees. Or above 100.  Or 10,000 feet in the sky.  Or over 55 MPH.  Or in an ocean. Or can't handle 1000 people a day pushing on actual elevator buttons. Or kids jumping up and down on a bed a thousand times a day.

      Or how do you pipe billions and billions of sensor data through the airwaves all over the world?  We have issues at times with electricity now. And that's in places like the US.  There are many countries where running water and electricity aren't exactly guaranteed every day.   So if they can't guarantee electricity, how are they going to guarantee Data and connectivity?   Smart connected devices and IoT isn't just buy some sensors, connect them, configure your software and do some analytics.  

     The Internet of Things and Analytics of Things will change the world.  It'll also be a complicated and connected system that isn't easy to implement or maintain.  Glossing over these realities helps nobody. It just creates a fake connected IoT/AoT world.  And that helps nobody but a few people and their bank accounts.



.

Fortifying the Cyber Frontier: Safeguarding LLMs, GenAI, and Beyond

In the ever-evolving world of cybersecurity and infosec, the convergence of cutting-edge emerging technologies like Large Language Models (L...