IoTChat

Edge computing - @floyer addressed this recently here: http://wikibon.com/t...

On the industrial side, we see a continuation of the migration from closed systems to classic more open IT systems.

"Noise". The scale of noise is going to push innovations in big data and related intelligent and learning systems.

The acquisition of Kepware by PTC fits that.

Kepware is sort of a babble translator for all of the closed systems, enabling classic IT communicate with lower level sensors/components.

Followup Q: There will be winners and losers in how #iot or #data is architected. Any thoughts on what the winners are doing right now?

The starting point is low cost sensors - MEMS & nanotechnologies have brought Dow the sensor costs to a few dollars. The biggest single cost is physical placement, connection & calibration.

It requires architectures that do some processing at the edge and only send the "gold" back

David: cost of connecting should be low so if there is issues with network and power doesn't it make it costly? remember the Truck roll issues in broadband? Thoughts

@jtmcarthur56 the technical intelligence is one piece to noise reduction and we are going places, the other side is the cognitive psychology & designing for THAT. That is a diff story.

@dfloyer don't sensors just get embedded at the point of manufacture going forward?

the biggest change up in the cloud is the need for big data and fast data

IoT is a big change in how apps/infra are built because it puts hardware design/engineering at forefront. Think of Steve Jobs/Jonny Ives and the Apple dev culture, and swirl machine learning into it in a seamlessly intelligent design

lost that: key is for #bigdata and #fastdata #analytics to be less of a trade-off

There is a lot of talk of the technology so far in this conversation, but little about how this is going to improve people's lives by adding value to it. Solve for that first, and the tech will come.

@jameskobielus that's interesting thinking - "data elegance"

@jameskobielus The assumption is that the data is there in Apple's case they were vertically integrated and the data was available. This supports the vertical integration for apps with data. Do you agree?

yup - the last mile, or in the case of IoT the first few hundred feet from sensor to edge computer. Challenge is to remove labor from that step.

@dvellante It requires architectures that do the vast majority of high-powered cognitive computing locally in semi-autonomous fashion. Think connected cars.

@dfloyer Bingo that is what I was thinking. That is one of the "holy grails" for #IoT

@daveixd "Internet of Dave" - what can instrumenting us do for our lives - good bad ugly???

@daveixd You're right. I see disaster avoidance technology, such as SafeKick on oil rigs.

@daveixd take a hospital patient as an example: machine measuring vital signs has several streaming outputs: 1 to nurses/drs, one to service company about maintenance, and one to device manufacturer about usage scenarios

@dvellante Think cognitive elegance: your smart devices gracefully complete your thoughts, remind you of what you may have forgotten, proactively guide you per your unstated intentions.

@ggilbert41 And hopefully stop waking up patients every few hours to check their vitals.

The cost of connecting/engagement with IoT needs to be asymptotically zero: free/reliable power at edge nodes (e.g., solar, motion), ubiquitous free wireless connections, free open algorithms/data.

Staying on Q2 for a second, I think a lot of data gets processed locally, filtered/analyzed/curated data gets sent to the data center for further analysis to inform distributed monitoring systems.

@jtmcarthur56 yes no question that's the prevailing model...so it comes down to what assets to instrument (everything?) - the prerequisite is connectivity (to avoid a "truck roll")

@jtmcarthur56 Sensors will not just be "embedded" at point of manufacture: every surface and substance within smart materials needs to be a seamless sensor in its own right.

Alan Patrick ‏@freecloud said on Twitter "Various, but one is end devices are very dumb/thin datasteam, can do very little, diff to modern ICT archiectures"

I do not think IOT platforms are any different vs other big data analytics; There will be always some differences for each market but lot of commonality exists for IOT with other markets like Cyber security, ADtech analytics, finServ analytics

Future of connected devices is not just the device but the "swarm".

Swarming or the need for swarms will force collaboration between competitors to enable establishment of swarming protocols and inter-connectivity.

good qeusiton. identifying key use cases for IOT is important for IOT market success

Companies will need to fathom approaches where they compete at device level but collaborate at the swarm level.

A new kind of player, a swarm manager/enabler, will emerge which will focus on enablement of the swarm and monetizing it. Google could be one such player (Think where Nest may be headed...and the self-driving cars). Social Media companies like Fac

Social Media companies like Facebook etc. may also move up the value chain from just being people connectors to swarm enablers.

Smartphones, smart cars, smart homes, smart appliances, smart wearables, smart health monitoring/optimization devices, industrial M2M, etc.

Are there any hot verticals more then others? Oil & Gas, Healthcare, ??? thoughts

http://deepaksethspe...

Can you define #IoT #hyperconvergence ?

#IoT will span across traditional verticals

Fleet management use cases of all sorts. Check out my post today on the swarming nano-agents of algorithmic cognition in a world of drones for military, precision agriculture, etc.: https://www.linkedin...

Retail.

office elevators talking to cellphones

Check out @SparkCognition for real-time analytics for security/fraud prevention in retail.

@jtmcarthur56 I'll let @ggilbert41 define #IoT #Hyperconvergence - Go George

Consumer IOT is different from Industrial IOT

I believe healthcare, utility/infrastructure, logistics, manufacturing are the core verticals that will be strongest successes. As GE puts it, the Industrial Internet.

@daveixd Agreed, but I wouldn't leave out retail.

use cases around Consumer IOT will not offer compelling value props yet; these are NICE to have

@smuddu refrigerators vs turbines?

We have some serious #bigdata players on here and I know that @smuddu has a security angle - #security ?

@jtmcarthur56 to me, #IoT #hyperconvergence means that the trade-offs between limited analytics on #fastdata and the richer #analytics on #bigdata start to dissolve

@jtmcarthur56 consumer facing retail, or logistics? When I was at Moto we were toying with RFID in logistics. The margins are so low that a) investment w/o clear wins are hard to do; but b) it is so easy to win big. ;-)

@smuddu ROI on Industrial IoT is much clearer than consumer IoT

Early adopters will be in industrial IOT; and need to pick verticals like OIL/gas, Manufacturing etc

Seeing a great deal of action in healthcare IoT these recent weeks - expect it to be a hot vertical for 2016

Here's another recent post on deploying analytic microservices into the IoT: https://www.linkedin...

Then use cases need to look at Asset management, service/maintenance management etc

@dvellante turbines already doing it. With a turbine, you want good, real-time monitoring and failure analysis. Unscheduled maintenance is expensive on a turbine. And so is a turbine spinning out of control.

@jtmcarthur56 the problem with #analytics is that #bigdata needs to get stored first, then analyzed. #fastdata gets analyzed and then, maybe, stored - these need to merge!

@ggilbert41 that is hyperconvergence

@jameskobielus Smart cities are too vague use cases and not going to happen in next 2 years; may be after 10 years :-)

@jtmcarthur56 agree - this is a real use case for IOT

That latter post isn't about any particular vertical app, but it points to the fact that microservices is SOA for this new era of nano-endpoints that do more of the analytic heavy lifting in fogs.

Security and Health Care - 50% of the data will come from surveillance video

@daveixd logistics, inventory control, real-time marketing apps, real-time fraud prevention.

Healthcare is a key focus of IoT app dev. As can be seen from IBM's partnerships and announcements at CES, eg https://www-03.ibm.c...

Consumer appliances is another big IoT app dev focus. Once again, IBM's partnerships per CES are indicative of this trend: https://www-03.ibm.c...

@daveixd One of the companies incubating at @TechSandBox1 is working on technology to help marketeers and shelf managers monitor the impact of marketing and inventory in real time. Stores don't tell the producers about out of stock situations in real time.

Energy & utilities is a perennial hot IoT market: smart grid.

@jameskobielus this seems to be virtual reality; it will not be practical deployment in next 2 years

@jameskobielus agree; healthcare and medical devices have practical IOT use cases

GE predix is still not a working product; they are not building it again for 3rd time; when will we see this delivered to market?