Concrete Hydration Industrial IoT Remote Monitoring Solution - 1 year

Holt, Michigan USA

Archive Cases

“…and how are you going to connect that?”

2021 06 Concrete IoT

Introduction : The Internet of Things

After   its   inception   by   British   pioneer   Kevin   Ashton    back   in   1999,   the   term   Internet   of   Things (IoT)   has   come   a   long   way   since   becoming   more   relevant   as   the   digital   transformation   age   moves forward.   In   fact,   and   after   two   fast   decades,   its   evolution   has   seen   most   of   the   protocols   and associated   technologies   already   on   their   way   down   the   so   called   “ peak   of   inflated   expectations ”, passed   the   “ trough   of   disillusionment ”   and   well   into   the   “ slope   of   enlightenment ”,   as   shown   below per Gartner’s Hype Cycle for IoT Standards and Protocols : With   this   new   digital   pace   of   ‘plateauing’    from   places   to   people   to    ‘things ’,   classifying   and   identifying all   the   potential   verticals   and   solutions   can   be   a   daunting   task.   Yet   the   overall   current   consensus   in the   industry   has   targeted   IoT   around   two   main   applications ,   from   which   all   others   are   branched: asset   tracking    and   remote   monitoring .   And   when   it   comes   to   the   latter,   especially   on   the industrial    side   of   it,   temperature    will   almost   always   be   the   #1   attribute    on   the   list   to   measure, monitor and report. From   monitoring   precise   levels   in   the   soil   for   agriculture,   or   in   critical   freezer   containers   for   Covid- 19   vaccines,   setting   ambiance   thresholds   for   hospitality   and   buildings,   or   mechanical   variances   of industrial   equipment,   temperature    in   today’s   Fourth   Industrial   Revolution    is   no   longer   the trivial   measurable   attribute   it   once   was   since   the   first   thermometers   were   invented   back   in   the early 18th century. Such    is    the    case    of    rēd    wireless’    geotechnical,    structural    and    environmental    engineering colleagues   and   clients   at   HAE   &   Associates    located   in   Canton,   Michigan.   Since   1994,   HAE   & Assoc .   has   been   trusted   by   dozens   of   public   and   private   clients   for   their   multi-disciplined,   cost- effective,   yet   practical   solutions   in   construction,   planning,   surveying,   management   and   testing services.   Having   their   own   in-house   state-of-the-art    testing   lab   that   operates   24   hours   a   day,   7   days a   week,   HAE   &   Assoc.   is   among   a   selected   few   in   the   industry   certified   to   provide   advanced materials   evaluation   services   at   all   stages   (from   product   development   to   materials   already   in place)   and   help   diagnose   failures,   develop   methods   for   improving   existing   products   and/or   find new   uses   for   materials   for   different   applications   such   as   geotechnical   engineering,   environmental sciences and construction monitoring. One   of   their   lab   services   involves   the   placement   of   various   client’s   concrete   ‘ cylinders ’   and   other construction   materials’   ‘ core’    samples   inside   a   tank   filled   with   a   precise   pH    temperature-controlled solution   to   monitor   the   samples’   characteristics   and   performance   before   implementation   in   the real   world.   As   such,   it   is   imperative   for   HAE   &   Assoc.   professionals   to   keep   accurate,   real-time records   of   each   sample   and   monitor   the   temperature   inside   each   tank   daily ,   thus   responding quickly   to   any   changes   in   the   tanks   and   adjusting   conditions   as   needed   –   clearly   not   a   trivial   process   by   any   standard.   An   Industrial   IoT   remote   monitoring   solution   seemed   like   the   ideal candidate for this use case, so a ‘ pilot’  project was proposed by rēd wireless.

The Challenges

Not   only   this   industrial   temperature   case   was   not   such   a   trivial   ‘ attribute ’   to   collect,   but   also   the specific   pH -solution   filled   tanks   added   an   additional   level   of   complexity:   the   sensor’s   temperature measuring   device,   or   ‘ probe ’,   needed   to   be   able   to   chemically   withstand   the   various   alkaline/acid levels   present   inside   the   tanks’   solution   without   altering   the   measurements.   Moreover,   typical wireless   IoT   and   industrial   monitoring   sensors   only   alert   when   a   certain   threshold   is   ‘ triggered’ , limiting   the   number   of   messages   sent,   thus,   conserving   battery   life   in   the   process.   In   this   case   we are   requested   to   monitor   every   single   hour   of   every   single   day   of   every   single   month   –   for   12 months.   In   simple   math,   24   temperature   readings   a   day   would   be   measured,   collected,   and   sent by    the    IoT    temperature    sensor,    each    day    for    365    days    which    amounts    to,    at    least,    8,760 temperature    data   points   in   a   year    (read   ‘ at   least’    since   we   are   not   including   periodic ,   supervisory,   and   other   non-temperature   related   messages).      And   since   this   industrial   lab   environment   is   home to   not   just   the   materials   evaluation   tanks   but   also   many   other   chemical,   environmental   and geotechnical     experiments     and     applications,     the     radio     frequency     (RF)     ‘ ambient     noise ’     and interference   levels,   hardware   placement   and   antenna   specifications   all   had   to   be   considered when    selecting    the    appropriate    wireless    connectivity    technology    for    sensor    communication quality.   Finally,   and   probably   most   important   of   all,   the   pilot   would   be   deployed   right   at   the   start of the Covid-19 pandemic and under the subsequent state  and national lock-down guidelines.

The Technology

Given   all   the   requirements   for   this   project,   we   needed   a   robust   wireless   technology   that   could withstand   the   harshest   of   RF   environments,   with   minimal   or   no   human   intervention   after   setup, having    the    best    energy    efficiency    given    the    demanding    hourly/daily    polling    of    temperature readings.   Since   each   of   the   readings   involves   only   a   small   number   of   packets   of    asynchronous    data   (i.e.,    temperature    readings    at    least    once     an    hour,    regardless    of    when     within    that    hour), technologies    such    as    Wi-Fi     or    Bluetooth     were    not    considered    as    these    have    a    lot    more    data capacity   (or   bandwidth )   than   was   needed,   which   in   turn   add   range,   battery,   and   interference limitations   due   to   their   bigger   data   rates   and/or   higher   frequency   RF   bands .   Other   technologies considered    that    do     have    lower    data    and    power    specifications    with    better    range    and    lower frequency   RF   bands   were   either   proprietary   or   not   well   supported   –   thus   having   the   risk   of   typical and non-ideal ‘vendor-locked-in’  scenarios with no future-proof  guarantees of any kind. After   careful   evaluation   and   considering   rēd   wireless   have   already   tested   and   evaluated   long range    and    low    power    wireless    technologies    in    the    past,    it    was    determined    that    the    wireless communication    best    suited    for    this    project    was    the    California-based    Semtech    Corporation ’s patented    yet     royalty-free    LoRa ®    PHY     (physical)    connectivity    protocol,    in    conjunction    with    the open-sourced    LoRaWAN ®    MAC     (Medium    Access    Control)    network    protocol,    managed    and standardized   globally   by   the   LoRa   Alliance ®.   Not   only   does   the   LoRa®   protocol   offers   incredible robustness    against    ‘ noisy ’    RF    environments,    with    long    range    features    capable    of    combating interference   and   ‘jamming’    scenarios   which   dynamically   adjust   so   it   can   still   recover   the   small messages,   but   also   employs   a   secured   and   encrypted    sensor-to-network-and-application    end-to- end   AES-128  scheme that comes standard , and not just as a later security ‘ add-on ’:

The Sensor

After   selection   of   the   LoRa®   and   LoRaWAN®   technologies,   the   next   step   was   to   choose   the sensor   that   would   collect   the   ambient   temperature   inside   the   tanks   via   an   external   probe   capable of    withstanding    the    various    alkaline    or    acid    liquid    levels    present,    transform    this    information accurately   into   an   electronic   signal   that   could,   then,   be   transmitted   wirelessly    to   a   local   ‘gateway’   (or    ‘concentrator’     in    LoRa®    ‘ lingo ’),    which    in    turn    then    sends    this    data    to    the    network    and application   platforms,   secured   &   encrypted,   for   analysis,   reporting,   alerting   and   visualizing   (see diagram   above).   And   because   this   is   an   industrial   commercial   setting,   the   device   in   question would   need   to   conform   not   only   to   the   challenges   already   described   above   but   also   to   industry standards such as Ingress Protection  (IP) against dust  and water  as well as being FCC certified . Given   all   these   ‘checklist’    items,   and   after   careful   evaluation   of   many   ‘off-the-shelf’   options   available given   the   broad   adoption   of   LoRa®   with   many   device   manufacturers,   we   decided   on   Minnesota- based   Radio   Bridge ’s   industrial   RBS306-TEMP-TC-US    LoRaWAN®   Wireless   Thermocouple    Armored Sensor ™.    The    RBS306-TEMP-TC-US    can    be    deployed    ‘ standalone ’,    as    in    this    case,    or    via    the company’s   sensor-to-cloud    solution.   Each   one   comes   with   an   external   K-type    thermocouple   probe, which   we   did   not   use   since   our   project   required   one   suited   for   specific   pH    conditions.   Besides   K - type,   the   sensor   can   be   configured   for   other   popular   industrial   probe   types ,   such   as   B ,   E ,   J ,   N ,   R ,   S   and   T .   With   a   NEMA   enclosure,   IP-67   rating,   operating   ambient   temperature   range   between   -40°C and   +70°C,   16   bits   precision,   .06°C   of   accuracy,   and   lithium   batteries   capable   of   200k+   messages, this   Armored   Sensor ™   was   the   ideal   choice   for   our   project.   Combined   with   Radio   Bridge’s   unique messaging   approach,   where   event   payload    data   messages   such   as   temperature   and   humidity   are separated   from   periodic   ones   such   as   battery   status   which   can   be   scheduled   independently   by   the user,   full   decoding   documentation   available   and   unmatched   customer   support,   makes   this   sensor extremely easy and flexible to configure for the coders  and non-coders  alike. Finally,    the    correct    pH    specialized    quick    disconnect    thermocouple     ‘stinger    probe’     was    sourced locally   thanks   to   the   collaboration   of   our   geotechnical   engineering   colleagues   and   partners   at Livonia,     Michigan-based     Rhino     Wireless ,     which     was     connected     without     any     additional configuration or calibration needed.

The Platform

Once    the    hardware    challenges    and    network    needs    were    all    addressed,    rēd    wireless    then proceeded   to   focus   its   attention   on   an   IoT   platform    –   which   is   the   one   ‘ piece   of   the   puzzle’    that customers   will   interface   the   most.   As   such,   we   needed   a   highly   flexible   IoT   user   experience    (UX) environment   with   the   ability   to   create   a   uniquely   branded   application   for   clients,   easily   accessible without    the    need    of    special    ‘apps’ ,    secured,    accurate    and    reliable.    Moreover,    this    project application    required    real-time    temperature    monitoring,    with    on-demand,    weekly    scheduled reports,    threshold    programmable    alarms    and    alerts    via    e-mail    and    SMS .    After    evaluating    6   different   types,   from   ‘boxed’    and   ‘ closed-coded ’   applications,   to   do-it-yourself    ones,   we   selected   a     cloud-based Application Enablement Platform  (AEP). As   a   result,   and   with   minimal   coding,   rēd   wireless   was   able   to   provide   this   industrial   client   with   a fully   automated ,   real-time   dashboard    view   of   their   pH -solution   filled   tanks’   exact   temperature levels,    thresholds,    alarms,    alerts,    events,    and    other    relevant    data    on-demand ,    anytime     and accessible   from   anywhere   via   secure   HTTPS .   The   solution   also   allows   for   full   historical   and   current records   access,   as   well   as   two   separate   weekly   reports   that   await   the   client’s   e-mail   inbox   each Monday   morning   –   an   overall   weekly   report    and   a   detailed   hourly   report .   They   also   receive   SMS    texts alerts if the temperature levels exceed their own programmable above/below thresholds.

The Deployment

Once   all   the   hardware   and   software   requirements   were   tended   and   the   project   scope,   planning and   goals   had   been   discussed   and   approved,   rēd   wireless   finally   deployed   the   pilot   system   at   HAE &   Assoc.   lab   on   March   16th,   2020   –   exactly   one   week   before   Michigan’s   Governor   1st   statewide stay-at-home   order    for   all   non-essential   workers   due   to   the   start   of   the   Covid-19   pandemic. Following    early    mask    mandates    and    social    distancing,    we    validated    that    the    location    and placement   of   the   hardware   was   secured,   including   the   gateway   and   its   connection   via   physical ethernet   medium   ( not    Wi-Fi)   and   to   the   LoRaWAN®   network,   as   well   as   the   Armored   Sensor ®   and probe   which   were   mounted   on   a   temporary,   custom   made   wood   frame   next   to   the   pH -solution filled tanks at the lab. rēd   wireless   then   proceeded   to   conduct   an   RF   survey   study    of   the   building   utilizing   a   handheld spectrum   analyzer    to   measure   and   identify,   among   other   key   performance   indicators    (KPI),   ambient noise   levels   and   potential   sources   of   interference   to   the   LoRa®   physical   channels.   Once   the survey   was   completed   and   the   results   were   analyzed   to   be   within   acceptable   thresholds,   we conducted   an   additional   end-to-end    network   bidirectional   ‘ live’    test   using   a   LoRa   Alliance®   certified Field    Test    Device     (FTD)    with    confirmation    of    messages    to    validate    the    performance    of    both   communication   directions:   the   ‘uplink’    (UL,   or   sensor-to-gateway-to-network ),   and   the   ‘downlink’    (DL, or   network-to-gateway-to-sensor ).   Although   many   LoRaWAN®   live   applications   will   use   unconfirmed   messages   (meaning   that,   for   example,   the   sensors   will   typically   send   periodic   messages   to   the network   without   expecting   a   confirmation   that   the   messages   were   received   successfully)   due   to the   high   confidence   in   the   LoRa®   protocol   and   its   error   correction   features   that   maximize   battery and   capacity   efficiency,   when   testing   commercial   solutions   on-premises,    one   must   validate   that   the system   is    working   as   expected   by   measuring   metrics   such   as   Packet   Error   Rate    (PER)   among   other KPIs   –   a   common   misconception   of   some   in   the   industry   deemed   as   ‘ worthless ’   or   ‘ time   consuming ’. This   type   of   network   testing   is   key   to   have   as   a   baseline    for   any   future   problems   and   since   no wireless   network   will   ever   be   100%   error   free   all   the   time.   Both,   the   RF   survey ,   and   network validation   testing ,   come   standard    with   all   rēd   wireless   commercial   projects.   rēd   wireless   also provides   RF   Indoor/Outdoor   Prediction   Design   Studies    for   bigger   and   more   complex   commercial projects,    featuring    1,    2    or    more    gateways    with    multiple    sensors.    This    deployment    took approximately 1-2 hours to complete.

The Results

The   final   dataset   consists   of   8,760   data   points    (hourly   temperature   readings,   every   hour,   every day,    for    1    year)    from    the    Radio    Bridge’s    industrial    RBS306-TEMP-TC-US    LoRaWAN®    Wireless Thermocouple     Armored    Sensor ™.    The    errors,    or    ‘ lost ’    messages,    were    categorized    as    1.    non- connectivity    (non- wireless    and/or   network ),   2.   wireless    (RF)   and   3.   network .   A   mandatory   network Application   Programming   Interface    ( API )   configuration   change   on   the   week   of   October   26th,   2020, tallied   93   missing   messages   across   4   days    which   were   excluded   from   the   final   count   since   these were   non-connectivity    related   ( API   config ).   For   the   connectivity   related   errors,   there   were   a   total of   24   network   connectivity   interruptions    spread   across   4   different   weeks   ( network    errors)   and only 7 wireless related missing messages  spread across 5 different weeks ( RF errors): Our   original   goal    target   for   wireless    ( RF )   connectivity   related   errors   was   2%   or   less ,   taking   ‘ cue’   from    our    3G/4G    cellular    performance    ‘ old    testing ’    days.    To    say    that    our    expectations    were exceeded    is    an    understatement,    having    only    a    .35%     message    ‘ network / RF ’    final    error    rate. Moreover,    this    pilot    project’s    test    LoRaWAN®    Network    Server    was    not     SLA ’ d     (Service    Level Agreement)   or   QoS ’ d    (Quality   of   Service),   still   we   were   most   happy   with   these   results   of   just   a .27%    network   ‘ downtime ’   /   99.73%    network   ‘ uptime ’   (not   counting   API   config )   -   just   shy   of   the typical industrial and commercial ‘ three nines ’ ( 99.9% ) SLA standard: The   Radio   Bridge’s   industrial   RBS306-TEMP-TC-US   LoRaWAN®   Wireless   Thermocouple    Armored Sensor ™   performed   brilliantly,   including   in   the   ‘ energy   department ’,   maintaining   the   same   initial 3.1    volts     averaged*    all    throughout    the    year    without    having    to    replace    the    original    batteries (* values   fluctuated   between   3v   and   3.1v ).   Once   again,   these   exceeded   our   expectations   considering the   large   amount   of   temperature   data   points   collected,   in   addition   to   all   other   periodic   and configuration    messages,    making    the    business    case    for    this    device    on    stricter    scheduling requirements (less than 1hr reports) as well as its sustainability  and battery waste reduction . Finally,    and    without    even    realizing    it,    we    were    able    to    validate    some    of    the    well-known connectivity    ‘corner   cases’    currently   being   addressed   by   many   LoRa   Alliance®   members,   including questions    such   as   “ what   happens   when   a   device   cannot   connect   for   over   3+   days ”   and   “ how   well   does a   device   join   back   the   network   after   a   3+   days   interruption ”   –   which   following   our   mandatory   ‘ API configuration ’   4   day   ‘outage’    in   October   2020,   the   answer   to   both   questions   is   :   “ this   device   re-joins the network flawlessly ”. These   results   speak   to,   not   only   the   resiliency   of   the   LoRaWAN®   technology,   the   quality   of   the chosen   hardware   components,   and   the   overall   reliability   of   the   networks   and   platforms,   but   also of   the   critical   importance   system   integrators    (or   any   other   IoT   implementer   for   that   matter)   and their   wireless   connectivity   skills   and   experience   play   in   the   success   (or   failure)   of   any   IoT   project   – big   or   small.   From   Beecham   Research’s   own   web   report,   whyiotprojectsfail.com ,   to   the   LoRa Alliance®’s   Certification   Test   Tool   (LCTT)   and   RF   Performance   Evaluation   Procedure ,   it   has   finally become   noticeable   that   the   wireless   #connectivity    side   remains,   as   industry   publication   IoT -NOW   puts   it   in   their   2021   Q2   edition ,   “ a   complicated   jungle   of   telecoms   industry   technology ”,   something even   ourselves   have   witnessed   since   rēd   wireless’   inception   back   in   2016,   seeing   (and   advising) many in the IoT industry that have chosen to overlook this critical piece of the overall IoT puzzle  for the    past    7    years.        With    over    25+    years    of    cellular    and    non-cellular    wireless    experience, connectivity    is   at   the   very   core   of   all   rēd   wireless   operations   and   forms   the   foundation   for   all   our services and solutions – and this project’s success serves as testament of that.

The Customer Impact

After   only   the   first   3   months   of   operation,   rēd   wireless   asked   Mr.   Gus   Haengel ,   President   of   HAE   & Associates ,    3    insightful    key    questions    concerning    the    ongoing    project    results    –    below    are    his answers: 1) How satisfied are you with the project so far? “ I am extremely satisfied with the project.  Excellent data, fast and easy to make decisions .” 2) Any questions or highlights so far? “ No question[s].  The data information is clear and easy to read .” 3) Does this solution bring value to your organization? “ Yes,   as   we   engineers   do   not   have   extra   time   in   our   days.      This   type   of   data   is   very   valuable   and   time saving for our precious time.  And is excellent to make fast and clear decision on the systems. ” Indeed.   As   rēd   wireless   became   more   familiarized   with   the   many   aspects   of   the   geotechnical industry   through   this   remote   monitoring   pilot   project,   we   quickly   realized   this   was   no   typical, trivial   ‘ temperature   logging ’   exercise.   Due   to   the   tanks’   specific   heater   element   characteristics,   lab ever-changing    ‘dynamics’    and   strict   required   pH    levels,   once   new   core   samples   have   entered   the tank    (as    seen    in    the    graph    below    at    the    beginning    of    the    week),    if    the    temperature    goes above/below   the   specified   thresholds   it   could   take   up   to   16   hours    of   manual   daily    ‘ adjustments ’   to bring it back to ‘ normal ’ and in compliance: To   keep   the   integrity   of   these   core   samples,   which   is   critical   to   the   success   of   the   lab,   careful inspections   and   methodical   measurements   must   be   accurately   logged   hourly ,   daily ,   and   weekly . The   consequences   for   any   errors   or   omissions   can   be   costly,   both,   in   time   and   money   and   cannot simply   be   ‘ pencil-whipped ’:   each   core   sample   (which   can   be   as   many   as   20   or   more   present   in   the tanks   at   any   given   time)   must   be   meticulously   extracted   from   the   field   by   a   specialized   crew,   using sophisticated   and   expensive   ‘coring ’   equipment,   and   shipped   to   the   lab   –   to   the   tune   of   about $1,000   to   $3,000   per   sample .   This   price   can   easily   double   in   the   wintertime,   as   the   extraction   of the samples become more challenging. With   rēd’s   remote   monitoring   solution,   all   hourly,   daily,   and   weekly   trends    can   now   be   monitored , visualized ,   alerted ,   reported ,   and   reacted    in   real   time   before   ‘ hitting’    the   specified   thresholds, thus   cutting   the   critical   time   once   passed   the   thresholds   to   4   hours    or   less,   greatly   minimizing risk  while increasing productivity  by about fourfold. Mr.   Haengel   has   also   expressed   how   invaluable   it   has   been   for   him   to   receive   via   e-mail   every Monday   morning   the   two   automated   reports:   the   one-page   summary   weekly   overview ,   and   the multiple-pages   detailed   log    for   each   individual   hourly   data   point.   It   allows   him   a   quick   and   effective analysis   and   sharing   of   the   data   with   his   team   on   what   has   transpired   the   week   prior   and   what’s trending in the week ahead.

Artificial Intelligence (AI)

In   collaboration   with   our   friends   and   colleagues   at   Elipsa.ai ,   we   decided   to   go   even   further   with this   project.   Providing   Elipsa   with   the   last   6   months’   worth   of   data   for   them   to   process   through their   ‘ AI   outliers ’   algorithm,   we   wanted   to   see   if   any   additional   trends   or   other   relevant   information would   come   up.   And   although   we   were   collecting   just   hourly   temperature   readings   and   the   alerts when   temperature   levels   were   above/below    a   certain   specified   threshold   from   just   1   device,   an interesting   ‘ pattern’    did   come   up   as   a   result:   the   AI   model,   based   on   roughly   4,380   temperature   data    points    collected    and    their    associated    above/below     thresholds,    suggested     changing     the present   above/below    thresholds   to   more   precise   values    for   better   performance   and   less   ‘ reaction’   time,   lining   up   almost   perfectly   with   our   own   findings,   results   and   recommendations   and   serving as   proof   of   the   benefits   and   efficacy   of   AI   services   such   as   Elipsa’s,   something   we’re   keen   in offering as well as part of our solutions in the future.

What Is Next?

It    should    be    no    surprise    to    anyone    reading    at    this    point,    and    rēd    wireless    is    also    happy    to announce,   that   we   have   officially   begun   the   transition   of   this   project’s   status   from   ‘pilot’    to,   now, ‘ commercial’    implementation.   In   collaboration   with   our   geotechnical   engineering   colleagues   and partners   at   Livonia,   Michigan-based   Rhino   Wireless,   and   after   a   100%   satisfaction   rating    from HAE   &   Associates,   this   project   becomes   our   2nd   successful   industrial   IoT   implementation    (a score   of   2   out   of   2 )   in   less   than   3   years   –   a   1st    under   the   rēd   wireless   name;   the   second   (and   prior one)   in   2018   seeing   our   involvement   only   as   professional   contractors   for   one   of   the   largest   global automotive   companies   in   Detroit,   Michigan,   with   equally   successful   completion   of   our   wireless connectivity ‘ piece of the puzzle ’ in just under 6 months. Below   are   some   of   the   more   relevant   ‘ next   steps ’   in   the   ongoing   development   of   this   now   proven, successful industrial IoT implementation: 1) Add   new   sensors   with   specialized   probes   to   monitor   the   water   levels   of   the   specific   pH- solution   filled   tanks    –   As   the   temperature   rises   –   or   changes   –   within   each   tank,   the   water   of the   specific   pH -solution   slowly   evaporates,   something   which   also   needs   remote   monitoring   in conjunction with the temperature levels. 2) Lab   environment   also   needs   temperature   and   humidity   remote   monitoring    –   we   are   also deploying   temperature   and   humidity   ambient   sensors,   not   only   on   the   lab,   but   all   throughout the building as well - which may include other attributes such as CO2 . 3) Develop   a   new   Elipsa.ai   routine   to   identify   an   upward/downward   temperature   trend    –   As   the temperature   inside   each   tank   trends    upward   (or   downward)   for   a   specific   amount   of   time   and before   a   specific   threshold   value,   identify   and   create   a   sort   of   ‘ pre-alarm ’   condition   just   before hitting the actual threshold limit. 4) Switch     to     an     SLA’d/QoS’d,     commercial     and     fully-technical-support-backed     LoRaWAN® Network   Server   (LNS)    –   Last   but   not   least,   although   the   results   for   this   project   exceeded   all   our expectations,   the   current   LNS   used   in   this   pilot   was   NOT   backed   by   any   typical   Carrier   Grade   commercial   Service   Level   Agreement    (SLA)   and/or   Quality   of   Service    (QoS)   requirements, which    is    perfectly    normal    for    ‘testing’    purposes    and    expected    of    this    type    of    so    called ‘ community ’ LNS servers like the one chosen. But    the    bigger    issue    is    the    fact    that    with    these    community    LNS    services    there    is    simply    no guarantee   of   any   kind   of   network   uptime ,   with   minimal   or   no   technical   support   available   other   than   public   ‘ forum ’   online   spaces   -   regardless   of   problem   origin   or   source.   As   such,   we   were ‘ left   in   the   dark’    for   4   days   when   our   non-connectivity   related   API   mandatory   config    in   the   week   of October   26th,   2020,   left   us   ‘ AWOL ’   -   resulting   in   93   missing   messages   in   one   week .   rēd   wireless was   forced   to   deal   with   the   situation   in   the   most   creative   of   ways,   even   though   it   was   found   to   be an   unforeseen   internal   LNS   problem   (read   ‘ not   rēd ’)   –   something   totally   unacceptable   for   any/all commercial solutions (even for home/office cases as well, at least in our professional view). For   this   and   many   other   reasons,   and   after   evaluating   6   other    different   commercial   LoRaWAN® Network   Server   offerings   from   around   the   globe   in   the   last   3   years,   all    rēd   wireless   commercial remote   monitoring   solutions   moving   forward   will   now   be   hosted   and   supported   in   the   US   by global    IoT    company    and    Swiss-based    LORIOT     –    including    this    HAE    &    Associates    industrial commercial solution. Among   the   many   features   and   advantages   of   selecting   LORIOT   as   our   commercial   LNS   of   choice and,   subsequently,   of   rēd   wireless’   offerings   and   solutions   are   :   Carrier   Grade,   99.9 %   ( three   nines ) network   SLA ,   end-to-end   encrypted    bidirectional   data   and   device   protection,   LoRaWAN®   AES- 128     encryption    combined    with    high-grade    TLS    v1.2     communication,    ‘battle-tested’     security, scalability   with   built-in   redundancy,   high-availability,   minimal   maintenance,   MQTTS    integration (rēd   wireless’   protocol    of   choice   for   communication   between   LNS   and   AEP),   multi-tenancy,   gateway logs, alerts and alarms and many others.