SEARCH searchBY TOPIC
right_division Green SCM Distribution
Bookmark us
sitemap
SCDigest Logo
 
 
distribution

Focus: RFID and Automated Identification and Data Collection (AIDC)

Feature Article from Our RFID and AIDC Subject Area - See All

 

From SCDigest's OnTarget e-Magazine

- Oct. 20, 2015 -

 

RFID, IoT and AIDC News: The Internet of Things - It Won't be Easy

 

Five Key Challenges IoT System Designers will have to Navigate

 

SCDigest Editorial Staff

We are certainly in the hype mode of the Internet of Things (IoT), with the emphasis mostly on actual and potential successes, such as Caterpillar's connectivity with millions of its machines across the globe, or the Amazon Dash button for automated home order entry.

But the reality is that there will be many technical, legal, regulatory and process challenges to have tens of billions or perhaps even trillion of things connected to the Internet.

SCDigest Says:

start

Decisions about what data needs to be stored locally or stored in the Cloud, and with what frequency, will be a key aspect of IoT system architecture.

 

close
What Do You Say?

 

Click Here to Send Us Your Comments
feedback
Click Here to See Reader Feedback

In a recent witty column on EngineringLive web site, Tony Paine, CEO of Kepware Technologies, notes that the many promised benefits of IoT are "contingent upon the resolution of key challenges; most notably the interchange of data between the mechanical, digital and human components of an industrial automation process."

Paine identifies as what he sees as five key challenges to mass IoT deployment:

Challenge 1: Identifying Things within the Internet

In order for things to be able to communicate with each other, they need to be uniquely identifiable within the Internet, Paine says. Until now, this has been accomplished through a unique Internet Protocol (IP) address. So the first question becomes, do we have enough IP addresses to go around?

Paine says there is a strong move to adopt something called the Ipv6 standard, which defines a 128-bit address capable of uniquely identifying 340 undecillionth (we're not sure, but it sounds like a very big number) of addressable items (compared with only 4 billion addressable items using today's Ipv4 standard).

While that will address the pure number of things to connect, Paine says there will be process problems too.

"It will be difficult, if not impossible, to manage this [IP assignment] effectively on a global Internet scale," Paine writes. "Typically, such tasks have historically been achieved by Naming and Number Authorities, with the aid of Network Administrators, but this will likely be an impediment, as things become added at an unprecedented rate.

Challenge 2: Discovering Things and the Data they Possess


Once a thing can be identified, Paine says the next challenge is how other interested parties will discover that it exists and what data it possesses. Of course, a thing should be able to restrict discovery of some or all of its data, based on security requirements. Just consider: what will be the process for adding a new "thing" on a factory floor or consumer's home?

"Balancing ease of discovery with the rigid constraints of security will be fundamental to the success of IoT and must be achievable without a PhD in cyber security," Paine adds.

Challenge 3: Managing massive amounts of data

These trillions of things we eventually connect will produce something much larger than trillions of data points, all of which will need to be collected, analyzed and possibly archived.

"Moving this amount of data over the Internet will consume new levels of bandwidth, which could result in the degradation of service as well as higher costs for internet carriers, service providers, and ultimately end users," Paine notes.

In addition, archiving this data for future analysis will require massive amounts of data storage and a new generation of scalable analytical applications capable of honing in on individual points of interest in a timely manner.



(RFID and AIDC Story Continued Below)

 

 
CATEGORY SPONSOR: SOFTEON

 
 


Indeed, decisions about what data needs to be stored locally or stored in the Cloud, and with what frequency, will be a key aspect of IoT system architecture.


Challenge 4: Navigating connectivity outages

Don't assume the things that make up IoT, as well as the communication systems that connect them, will be available 100% of the time. Instead, expect intermittent or even long term outages., Paine notes.

"This could severely impact systems where data loss is unacceptable or where variances in the data needs to be known in real-time," he writes.

Challenge 5: Integrating existing infrastructure into new IoT strategies

For decades, industrial things have made data accessible over private, internal networks. Paine says that in the past, security has largely been ignored in these networks, in order to achieve optimization and integration between dissimilar industrial systems. It has been believed that there was safety when using a private network - though of course that seems rather deluded given today's major cyber risks.

"Given that the typical lifecycle for industrial things exceeds 20 years, there will be an expectation to integrate the existing into new IoT strategies. Opening these private networks and the data they contain to the Internet will require detailed security assessments to minimize risk of exploitation," Paine adds.

SCDigest would say Paine doesn't make that point strongly enough - the risks for hacking and data breeches would seem huge.

What are your thoughts on these five IoT challenges? Let us know your thoughts at the Feedback section below.

 

Recent Feedback

A point of view on addressing the challenges listed:

 

1. Identifying Things within the Internet: Agree that Ipv6 is the right decision. Nice to have but not necessary today. Not every 'thing' needs to connect to the internet. Good design principles will save the world a lot of grief for no value added.
2. Discovering Things and the Data: Given the concerns of security, variations of 'Publish & Subscribe' models with authentication makes sense.
3. Managing massive amounts of data: Good design principles will again minimize the need for everything to connect to the internet and therefore reduce the need to move big volumes of data. Different Cloud, Fog/Edge, smart device level computing are ways to address this. There is a big difference between a social media company’s data collection for potential monetization vs an enterprise running a business and supporting operational decisions. Bad enterprise data problems have not gone away.
4. Navigating connectivity outages: Some of the breakthrough innovations using IoT don’t require internet connectivity. This gets into the realm of M2M vs IoT.
5. Integrating existing infrastructure into new IoT strategies: This is most probably the biggest challenge for existing enterprises as they figure out their ‘transformation’. A much quoted reference is what UBER did to transportation. A very compelling proposition exists for not integrating to legacy infrastructure.

 

In conclusion some design principles to consider:

 

1.Fix problems at the source;
2.Fix what you control first; and
3.Don't create non value added work.
 


Subhash Chowdary
CEO
Aankhen Inc.
Oct, 21 2015
 
 
   
.