Blockchain Transactions. Neither a Mainframe nor a Cloud Service are good enough. Something Entirely New is Required.

Mainframes were and still are good for business transaction scale 

Cloud services for web transaction scale (HTTP)

Blockchain is a new paradigm for transactions. 

Neither if these past two historic iterations of transaction processing systems are adequate for blockchain services.

Because they were designed to perform optimally for their own time and place using that technology.

Blockchain requires a node service to process transactions which eventually are sent to nodes (miners) to order into time stamped blocks. 

These transactions could use HTTP still. And indeed the best scaling form of this older standard so far is something called Teranode.

But is there a more optimal solution?

Of course there is. It's just not been engineered for yet.

This new transaction paradigm would be innovation by virtue of it shifting to a completely new model of processing transactions.

What will this look like? Firstly it would need to be able process transactions at a higher scale than mainframes which would not be hard. Beyond that it needs more scale than even a cloud services front end using load balancing, ingress devices etc.

The innovations would driven by the new ideas which deliver this scaling with ease. Cloud services has always been a bit of a budget up. Mashing together low quality transactions services into a scaled out solution with say kubernetes and so on. Less than ideal. Sure, it works.  But it was never really innovative because it merely stacks up low quality services at a high scale. 

What will Blockchain tech need to invent to make the new transaction scaling paradigm truly innovative? 

I'm all ears.