So far, however, all these variations have only really been about refining cryptocurrencies or extending them to support another type of transaction. This brings us to the second generation of blockchains.
This demonstrates one of the big limitations of current blockchains: speed and capacity. (Speed is often measured in transactions per second, or TPS.) Several approaches have been suggested to solve this, from sharding to sidechains and so-called "second-layer" solutions. The need for more innovation here is strong.
With the words "smart contract" in the air and a proved—if still slow—technology to run them, another idea came to fruition: permissioned blockchains. So far, all the blockchain networks we've described have had two unsaid characteristics: They are public (anyone can see them function), and they are without permission (anyone can join them). These two aspects are both desirable and necessary to run a distributed, non-third-party-based currency.
As blockchains were being considered more and more separately from cryptocurrencies, it started to make sense to consider them in some private, permissioned settings. A consortium-type group of actors that have business relationships but don't necessarily trust each other fully can benefit from these types of blockchains—for example, actors along a logistics chain, financial or insurance institutions that regularly do bilateral settlements or use a clearinghouse, idem for healthcare institutions.
Once you change the setting from "anyone can join" to "invitation-only," further changes and tweaks to the blockchain building blocks become possible, yielding interesting results for some.
For a start, proof of work, designed to protect the network from malicious and spammy actors, can be replaced by something simpler and less resource-hungry, such as a Raft-based consensus protocol. A tradeoff appears between a high level of security or faster speed, embodied by the option of simpler consensus algorithms. This is highly desirable to many groups, as they can trade some cryptography-based assurance for assurance based on other means—legal relationships, for instance—and avoid the energy-hungry arms race that proof of work often leads to. This is another area where innovation is ongoing, with Proof of Stake a notable contender for the public network consensus mechanism of choice. It would likely also find its way to permissioned networks too.
Permissioned blockchains can avoid certain complexities that public, non-permissioned ones can't, but they still have their own set of issues. Proper management of participants is one: Who can join? How do they identify? How can they be removed from the network? Does one entity on the network manage a central public key infrastructure (PKI)?