The last time we looked at Blockchain economics was in late 2016 in one of our meetups when Dinis Guarda presented a paper on its likely evolution in the IoT space (see summary over here). Since then we were asked to look at how blockchain may be used in specific industries, in this case professional services.
I won’t go into the specifics of how blockchain works (it’s in the article about Dinis I linked to above) but it is worth summarising the plusses and minuses of blockchain here:
Good News:
- Highly Distributed – uses a highly meshed network which increases system resilience
- High degree of inbuilt security – chained blocks of transactions and the logic of nodes agreeing transactions before being confirmed can survive many hacks to a system
- The very high level of cryptography used (the “Hashing” calculations) puts a high load on anyone wishing to hack the system by brute force. You can click here to get various security solutions for protecting your confidential data.
- This Distributed architecture means you don’t need a centralised Hub to run it, or to have a trusted central party.
Bad News
However, as with all engineering, everything is a trade-off – cost vs. capability, security vs. speed, network resilience vs. network operation load, etc. And blockchain cannot avoid this, its benefits bring problems:
- Highly Distributed hits system scalability – a fully meshed system increases the amount of network communication required by the square of the number of nodes.
- Inbuilt security and cryptography takes a very long time to process, so the system is not capable of agreeing transactions speedily – it can only process a handful of transactions a second, and can take hours if not days to confirm a Bitcoin transaction for example
- In addition this adds costs – the hashing calculations are extremely energy intensive, calculations at the moment are that it costs in the order of $1000 to process a Bitcoin, and there are 12.5 per block, so it’s c $12,500 per block. So a “block” of about 2,500 transactions (that’s the Bitcoin size) means about $0.50 (50 cents) per transaction. Compare that to fractions of a cent for a typical financial transaction today.
- Bitcoin transactions are very simple and have a low data requirement. However, the more data required per transaction (say a smart contract) the fewer transactions possible per block (a Bitcoin block for example has a 1 Mb limit) and up goes the price per transaction. In addition if data has to changed up goes the processing load.
- It’s still not fully secure – Cryptocurrency Blockchains can be (and have been) hacked by exploitable flaws in the data stored in transactions, and by rogue actors on the inside.
In summary
So in essence blockchain technology (as designed today) is great if you want:
- highly secure,
- highly resilient technology
- no trusted 3rd party.
But, for blockchain to work it also needs transactions to be:
- relatively low volume (c 6 a second)
- not particularly time sensitive (hours, or even days to complete)
- relatively high transaction value to mitigate the blockchain operating costs.
In addition you have to value the distributed/low trust required feature of the system, as if you don’t need it then it’s a very high overhead, with transaction costs several order of magnitude higher vs. other existing approaches.
Worryingly, we see a lot of mooted applications in the press where it is clear there is not a hope in hell that Blockchains (in their current form), will be fast enough or cheap enough to work (see Dinis’s talk on IoT – linked to above – for a typical example). Somewhere between the hype, hope and heuristics is a major disconnect. In our view a lot of applications currently being suggested for blockhain at the moment won’t be able to cope with its operational and economic overheads.
Now as it happens (and fortunately for our client), some key Professional Services applications do fit the requirements of this profile, so it is worth them looking at blockchain technologies for applications in their businesses. So good news then.
But, even so, one should always compare the blockchain options to the existing, cheaper “Good Enoughs” around today. Also, keep in mind Dinis’s thesis for the evolution of IoT blockchains, which may also occur in Professional Services, i.e. there will also be a huge temptation to employ more scalable/lower cost/faster performance blockchain system designs with:
- Less distributed architecture to scale it for speed and cost
- Less complex security in the blockchain to reduce the processing load time and cost
But this, in Dinis’ view, will probably be partly done by implementing them as “private and less distributed” systems behind IT datacentre security walls.
Note that in this case the owner of such a “Walled Garden” blockchain will very probably put themselves in the position of the “Trusted 3rd Party” supplier and, as Dinis noted for IoT blockchains:
…there will be a “race” in each industry segment (and maybe across them) as major players vie to be the Industry champion in their segment, and let network effects drive them to dominance. This will of course give these players major control, information asymmetry, and the ability to price as they see fit.
This is not unlike most network businesses’ evolution – telecoms, electronic payments processing, internet, social networks etc. where initially players try to build “walled garden” services to maximise their revenues, and only after a long time does it commoditise, eventually, into a regulated utility.