Aside from the exponential increase of data traffic the devices would need an similar increase in processing power. A central hub in a location would help. But the you're partially defeat the idea of a universal block chain.
Not on blockchain. Blockchain technology is just not efficient enough (scaling, throughput, fees, processing power) to support IoT use cases. It uses a different approach based on Direct Acyclic Graphs (DAG) that they decided to rename "Tangle" (a pretty poor marketing decision imho)
I did not read the article but the fact it does not even mention iota (even to say it is not based on blockchain but on DAG technology) indicates it is probably very poorly researched and/or very biased
Iota isn't actually a crypto currency - there's a server run by the iota team that "coordinates" all transactions. See previous discussion here: https://news.ycombinator.com/item?id=16457120
The plan is to shut down the coordinator when network will be big enough. Even with the coordinator running, to say that Iota is not a crypto currency is quite wrong (also it is much more than a currency). In any case, to not mention Iota in an article about "IoT and blockchains" is just plain dumb considering this is the certainly the major project in this field (even if it only to mention that Iota is not running a blockchain but a DAG, and that it is not currently fully decentralized)
What happens to those 10 year batterylife IOT devices when they start doing blockchaining and maybe even needing to get info from server to just send the damn temperature data. Also these kind of devices might use wireless networks like LoRa with very limited bandwidths[1] and usually extremely low data caps(50 bytes per minute anyone?)
IMHO biggest problem with blockchain hypers is the denial of current status of the blockchain technology in regards of the number of transactions. For example bitcoin get around 2000 transactions per second and median confirmation time sitting in somewhere 8-9 minutes[2]. Would you like to wait 8 minutes at counter for them to confirm your payment? Likewise in IOT applications for critical temperature monitoring, would you like to know only after some time find out that your subject(person/machine) has overheaten?
Granted not all blockchain tech is equal in performance and compromises can be achieved to avoid those performance pitfalls. But at what point it would just be better to use basic signing and leave it at that(crypto engines are more or less standard features in microcontrollers).
You do not get the same guarantee with a credit card that you do with an irrevocable block chain transaction. Checking the source accounts balance (which is public with BTC) you could quickly verify funds and have high certainty the transaction will proceed. This is in effect how credit card transactions work under the hood as well. Finance still heavily relies on night time batch processing.
While I'm not sure what IOT blockchain is supposed to do - I guess it's a bit hypothetical at the moment - it might make sense for payments. For example your smart light socket could send some currency to order a new bulb when it goes. Or your router could charge some passer by a few cents to supply data service. While bitcoin is slow other stuff like Stellar take approx 5 secs and cost like 0.001 cents per transaction.
As an aside I wonder if your ccvt's AI system detected a party was happening and ordered some recreational drugs using some currency it came with when bought, who would be guilty.
10 comments
[ 2.8 ms ] story [ 32.1 ms ] threadwat
IMHO biggest problem with blockchain hypers is the denial of current status of the blockchain technology in regards of the number of transactions. For example bitcoin get around 2000 transactions per second and median confirmation time sitting in somewhere 8-9 minutes[2]. Would you like to wait 8 minutes at counter for them to confirm your payment? Likewise in IOT applications for critical temperature monitoring, would you like to know only after some time find out that your subject(person/machine) has overheaten?
Granted not all blockchain tech is equal in performance and compromises can be achieved to avoid those performance pitfalls. But at what point it would just be better to use basic signing and leave it at that(crypto engines are more or less standard features in microcontrollers).
[1] https://www.lora-alliance.org/technology [2] https://blockchain.info/charts
As an aside I wonder if your ccvt's AI system detected a party was happening and ordered some recreational drugs using some currency it came with when bought, who would be guilty.