I am seeing three types of blockchain:
- Blockchain that uses Bitcoin to deliver a consumer proposition. For example, Lighthouse (covered last week).
- Blockchain that uses some other “interim token of value” to deliver a consumer proposition.The best example is Ethereum with Ether. These “interim token of value” are sometimes called Altcoins, but that is confusing. You won’t buy pizza with them. They are simply used within the Blockchain system to pay for system resources. For example, Augur.
- Blockchain that is used as an IT tool for B2B processes such as settlement & clearing.
Most of the debate centers around 1 and 2 – whether to keep the block size constant and use Sidechains or whether to increase the block size or whether to use a new Blockchain such as Ethereum.
In this post, I focus on the 3rd use case. When I first encountered this, it was explained to me like a database that was totally decentralized. My first thought was “those overloaded folks in IT hardly need another database paradigm to suss out”. Just when they are getting to grips with when to use SQL and when to use NonSQL, you expect them to look seriously at yet another totally different paradigm?
One company at this layer is Eris. They forked Ethereum (which anybody can do and which IBM did with Adept). Their proposition appears to be Ethereum without Ether. In other words, Eris is an internal IT tool.
“Who will be the Red Hat of the Blockchain era?”
I am on record as an Ethereum fan. It now looks to me as significant as Linus Torvalds releasing Linux in 1991. So, it now looks like the question is:
“Who will be the Red Hat of Ethereum?”
That could be Eris.
Building an Internet scale decentralized P2P system is technically really, really hard. This is as hard as building an operating system, so the analogy with Linux fits.
Ask the guys who built Skype how hard it was to build an Internet scale decentralized P2P system. Building a value transfer system is far harder than a VOIP system because the risk of loss is so much higher. Some noise on the line that forces you to ask your buddy to repeat something is OK and a small price to pay for getting something free; losing some money through a technical glitch is not OK.
It is a hard technical problem because you have to deploy to millions of machines of varying power and type that are only intermittently connected to the Internet and deliver a service that is as fast and reliable as the centralized server based competition; it also must be as easy to build and deploy to a decentralized P2P network as it is to build and deploy to AWS or the machine in your closet.
The question is, why would this be relevant to Internal IT projects?
Databases work very well. They are scalable, reliable and cheap (most are open source). Making the data totally decentralized and immutable seems like a solution looking for a problem. That would true for the historic use of IT, which was internal facing.
I cannot imagine any conceivable reason why a decentralized and immutable blockchain based database would be useful for an internal ERP back office system.
Those systems fall into the “if it ain’t broke don’t fix them, just keep patching until they eventually fall apart under the weight of dependency conflicts”.
These are intra-company systems. Everything happens within a single enterprise. Orders might come in from outside and be sent outside, but those are not peer relationships.
The uses cases for a decentralized and immutable blockchain based database relate to inter-company transactional systems between peers. These are the kind of systems that are very common in the world of institutional trading systems for settlement and clearing processes. The parties to a transaction are peers – for example two investment banks. Today those inter-company transactions go through trusted neutral intermediaries such as DTCC or Stock Exchanges.
It is unclear what is the value of replacing those trusted intermediaries. It is unlikely to be price, because these are hyper efficient markets where the customers are big and have negotiating clout. Maybe it is security. More likely it is time.
Reducing the time taken for settlement and clearing processes creates competitive advantage.
The key feature is the immutability of blockchain databases. The parties to a transaction have to be 100% confident that nobody can change the transaction.
Update: if you think this is all theoretical, read this story in Bloomberg about the problems with the Dell buyout.