On Friday, November 9, Parity Ethereum announced the alpha release of its light client-based wallet, Parity Fether. For the moment, the Fether is only on testnets and the Parity team does not recommend it for mainnet use, but the release marks a significant advancement in UX for the decentralized access to Ethereum.
To understand the significance of this, you have to know what a light client is and why they matter.
A client (also referred to as a node) is software that connects a website or app's front-end user interface with the back-end Ethereum smart contracts and ledger.
There are a few client-types that serve different functions. Some nodes fully download the history of the Ethereum blockchain – including every single transaction since the genesis block – and record and verify all new transactions. Some nodes are also miners. These node types are essential to the security and decentralization of the network, but to run them requires a significant level of technical know-how, time, and resources.
This presents a problem, because if you want to connect to the blockchain, you have to use a client. To circumvent this barrier to entry, many apps and wallets running on the Ethereum blockchain use Infura, a third-party node cluster that connects the front-end user interface (UI) of Dapps and wallets to the back-end smart contracts on the Ethereum blockchain. Essentially, wallets and dapps using Infura's infrastructure relay transactions to Infura nodes, which then write them to Ethereum. There are a few problems with this, though, because the Infura-running apps simply trust the Infura nodes, creating a single point of failure and somewhat defeating the point of a decentralized, trustless, secure ledger.
Many apps and wallets running Infura also give the user the choice to connect to Ethereum using their own client (if they run one), but again, running a node isn't particularly user-friendly. This is where light clients come in.
Light clients are the more user-friendly alternative to full nodes. Light clients don't download the whole history of Ethereum and they don't verify every new transaction that comes in. Instead, they rely on full nodes to send and receive transactions. It's kind of like how Infura-enabled dapps rely on Infura nodes to communicate with the blockchain, except where Infura-based dapps and wallets put trust in Infura nodes, light clients rely on full nodes to do the same – they just do so in a cryptographically-secured, trust-minimized manner. (For more on how they do that, check out Parity's explainer on light clients from back in July.)
Light clients are often touted as the solution to the ecosystem's reliance on Infura, but light clients aren't without their problems. For one, there aren't a ton of them that are as light as they need to be for wide adoption; for example, many cannot run on your cell phone (though Parity's can). Additionally, light clients aren't widely implemented in existing Dapps and wallets, creating undue friction. Without default integration into Dapps and wallets, users must download and run a light client separately and then integrate them.
Parity's Fether, though not fully ready for mainnet release, promises to be a more user-friendly, decentralized, trust-minimized Ethereum wallet. And this is big, because if Parity can address any major bugs and offer a relatively reliable and user-friendly light-client integrated wallet, more Dapps and wallets might do the same, minimizing the reliance on Infura and spelling a big win for network security.