Now that most of us read that the #Bitcoin network is consuming more power than Ireland, namely 30 TWh per year (3.4 GW) for only 4 TPS (transactions per second), time has come to reflect on the desirability and economic viability of #blockchain technology.  Each bitcoin transaction consumes 235 KWh : driving from Brussels to Barcelona in a Tesla Model S...

By comparison a Nokia Cloud Mobile Gateway (ePDG) which Mobile Network Operators use to admit Wi-Fi callers on their core networks runs on a single server (two Intel 12-core CPUs) and consumes around 1.75 MWh per year (200 W) to deliver 70K TPS (IKE, GTP or Diameter) : 25 Wh per transaction or driving that Tesla for... 150m. OK we need 8 transactions to establish a connection between your smartphone and the network : 1.2 km. Even that is 1000 times more efficient than spending a Bitcoin.

A blockchain addresses the problem of lack of a central trusted party safeguarding access to a central transaction ledger - a bank, insurance company, social security agency, utility or telecom operator. A blockchain removes the need for any central organization to verify your personal ID card (US : driving license) prior to allocating a bank account number, insurance contract number, social security account, electric connection number or phone number in their database (respectively).

To join a blockchain anyone can pick a random keypair as his/her private and public identity, then declare creating assets (16 Bitcoins in this case), but the terrible price to pay for that privilege is to solve a very difficult computational problem, namely to find a random key for which the SHA-256 signature of the transaction block has at least 40 leading zeroes. A problem requiring so many unsuccessful attempts that the reward for solving it is constantly increasing (currently a quarter million €). A riddle so difficult that 3.4 GW worth of FPGA/ASIC compute power only yields 16 Bitcoins every 10 minutes.

Now imagine a bank or telco making the following proposal : pick any 64-digit (256-bit) "bank account" or "phone number" and if the SHA-256 signature has 10 leading digits "0" we'll credit your account with 256K€... People would start mining for such 64-digit number, but would you call this a "game", "contest", "marketing campaign", "sweepstake" or "illegal lottery"? Perhaps the latter, on the basis that bank accounts and phone numbers do not consist of 64 digits?

By common legal standards Satoshi Nakamoto is the criminal organizer of an illegal lottery disguised as a distributed database - an extremely inefficient one. So is Vitalik Buterin of #Ethereum, Charles Lee of #Litecoin and Chris Larsen of #Ripple : they could stand trial one day. Also of course for organizing a Ponzi scheme, for counterfeiting € and $.

Yes, blockchains have interesting properties such as immutability, privacy and erasing old transactions. But a well-administered Oracle database holding account or phone numbers is also immutable for me, as long as only my bank or telco has write permissions into it. My bank or telco are legally bound to respect my privacy, not to reveal my home address for example. And I'm sure their DB admins must have written a script to erase my old transactions and calls I did years ago.

Mobile Network Operators are in high need of a better identification and authentication scheme than the SIM, characterized by its 15- or 16-digit E.212 International Mobile Subscriber Identity (IMSI) and secret private key Ki. NB-IoT is e.g. hampered by that mandatory SIM. We thought that #3GPP would remove the burden and open up for #SIMless authentication in #5G, as 3GPP SA1 had specified, but in the end the SIM card lobby successfully kept it out of Release 15, only permitting EAP-AKA' and EAP-5G for the #SUPI (Subscriber Public Identity). Probably claiming that "SoftSIMs" would be there soon, as they have done for the past 15 years.

There's thus an enormous opportunity for #IoT manufacturers and users to conceive their own object identities as fulfilling a difficulty condition, but whether 6000 blockchain nodes would need to check that condition... hum... I don't think we can afford that. As soon as a more efficient, truly forward-looking Standards Development Organization or Object Manufacturer will fix the difficulty condition, it will be sufficient for any IPSec Security Gateway, ePDG or TLS server in the world to verify it before admitting the Connected Object to the core network. Forget 3GPP, the GSM Association or the NGMN Alliance... IPSec and TLS are over-the-top and access-agnostic.

Declaration : I do not own #BTC, #ETH, #LTC, #XRP nor other cryptocurrencies (but I have an 11-digit mobile phone number, 15-digit IMSI and some 16-character bank accounts :-)