Will Ethereum emerge as the de facto next-gen transaction processing platform?

*Originally published by our expert, Prabhu Eshwarla, on LinkedIn

Ethereum is a transaction processing infrastructure based on Digital Ledger Technology(DLT) aiming to power the next-generation of our digital lives, economy, and socio-political governance.

Ethereum was announced in early 2014 by Vitalik Buterin, and has been rapidly rising in popularity ever since. Let’s look at a few numbers. It processes over 630,000 transactions per day on the blockchain (with peak being 1.3 million on a single day).

Ether, the native cryptocurrency of Ethereum, is the second most valuable next only to Bitcoin. 35 million unique addresses exist on Ethereum blockchain. The most widely used token standard in the crypto-currency industry is ERC-20, a standard developed by Ethereum. 94% of the top 100 tokens by market cap, and 87% of the top 800 tokens are built on Ethereum. ERC-20 tokens account for about $5.6 billion in funds raised in 2017 through ICOs. In the first 3 months of 2018 already the corresponding value is $6.3 billion.

The Ethereum developer community is estimated to be the largest in the world at 250000 , about 30 times more than the next blockchain community, and it has attracted many of the best technical minds in the industry to work on its open source stack. Over 1700 decentralised apps (dApps) have been launched to date. In short, the Ethereum blockchain ecosystem is diverse, strong, filled with some of the smartest minds in tech and is flooded with innovation and activity at breakneck speed.

But all this has not been sufficient for Ethereum to become the de facto blockchain platform of the world and declare victory. There are two primary reasons for this:

1. Technical limitations: Technically, there are limitations in the Ethereum platform. According to Vitalik Buterin , the crucial problems that they are working on to push Ethereum to the next level are: privacy, consensus safety, smart contract safety and scalability. Of these four, the real showstopper is scalability. Ethereum can deliver about 15 transactions/second (tps) throughput compared to a payment network like Visa which handles about 2000 tps (with peak capacity of ~ 45000 tps). Stock exchanges can handle several tens of thousands of tps , with High Frequency Trading(HFT) handling message throughput upwards of 150,000 tps. If Ethereum has ambitions to power the global digital economy, clearly it needs to scale in order to compete. Linear, incremental scaling will not cut it. What is needed is massive, non-linear scalability.
2. Enterprise needs: Ethereum was conceived as a public, permissionless tech emphasising transparency (meaning anyone can view all transactional data) and decentralisation (meaning no one controls it), but enterprises are very concerned about the privacy of their data and protection of intellectual property. They do not want their transactions to be publicly visible. Ethereum’s original design goals are at odds with these enterprise requirements.

What is the Ethereum team doing about technical limitations?

Let’s first understand the root of the scalability problem of Ethereum. The ‘Blockchain Trilemma’ (a trilemma is a difficult choice from among three options) is a term coined by Vitalik Buterin, which states that blockchains can achieve only 2 out of the three desirable properties at one time: Decentralisation, Scalability and Security.

If Ethereum had picked scalability and security, then it would have become centralised and this would have allowed it to have a higher throughput. But, Ethereum as a design choice chose to prioritize decentralisation and security over scalability. So, how do we include scalability in the current Ethereum architecture ?

Figure: Ethereum scalability solutions

Ethereum scalability solutions:

The core Ethereum team has embarked on a very ambitious multi-year roadmap to improve scalability. The solutions can broadly be classified as follows:

1. Layer 1 aka on-chain solutions: These are improvements implemented at the base level protocol of Ethereum. The primary scalability initiative here is Sharding.
2. Layer 2 aka off-chain solutions: These are also known as Layer 2 solutions because they are built “on top of” the Ethereum main-chain (Layer 1). Two of the key scalability initiatives are Raiden and Plasma.

Let’s take a closer look at Sharding, Raiden and Plasma.

Sharding: In Ethereum’s current state, every node processes every transaction in the network. One of the special mining nodes selected by the consensus algorithm processes these transactions into blocks, which are again validated by each node . This level of validation give the blockchain a high level of security, but limits throughput and scaling. Blockchain sharding is a technique where the entire state of the network is split into multiple partitions called shards that each have their own independent state and transaction history.

Each node processes only certain shards (i.e., only verifies a subset of the transactions). If a node needs to know about transactions or blocks that it doesn’t store in its shard, then it finds another node with the information it needs. One of the goals of sharding is to move away from requiring ‘full’ nodes — those which store the full state of the network and every transaction that occurs. All these allow the overall throughput of the main blockchain to be much higher by distributing the processing, than having a single shard do all the work as it is now. Sharding has its origins in a traditional database scaling technique called ‘database sharding’, which splits a database into parts and puts each part on a different server, all of which is coordinated by a distributed database management system. If you are interested to understand the technical details of Ethereum sharding, I recommend this article by Vitalik and Ethereum Wiki’s Sharding FAQ.

However, before sharding can be implemented, Ethereum protocol must be upgraded to Casper (proof-of-stake consensus mechanism) , which is currently still in development. Currently, both Bitcoin and Ethereum use proof-of-work consensus algorithm. Moving to proof-of-stake (POS) from proof-of-work (POW) helps achieve better latency with less computation and storage. There are many different types of POS but the ones that use Byzantine Fault Tolerance based on Consortium consensus have better scaling properties, wherein validators for each block are randomly chosen by the network, who then agree if the block should become part of the chain.

Figure: Raiden payment channels

Raiden (Layer 2): The Lightning Network was originally created as a way to scale Bitcoin. The central idea of the Lightning Network is to rely on the main blockchain only when needed. Certain types of transactions (smaller value, recurring, or less significant ones) are executed ‘off chain’ (using payment channels), and are validated independently before being sent back to the main Blockchain, where they can be processed as one single transaction. The Ethereum Raiden Network is a scaling solution very similar to the Bitcoin Lightning Network.

The main difference is that Bitcoin Lightning Network is used only for payment (bitcoin) transfers whereas Raiden can be used for tokens and smart contracts. One can transfer ERC-20 tokens and also Ether (ETH). Decentralised apps can also run on Raiden.

An example of Raiden network in action would be where all your daily purchases at the local grocer are maintained in ‘offline’ records, and finally, at the end of the month, you make a single consolidated payment to the grocer. How it works is that you and your grocer would open a bi-directional payment channel (for you to pay for goods or for the grocer to give you refunds) and both deposit some money committing to the trade, in the form of an initial transfer of crypto currency or tokens from the main chain to payment channel.

After this, there can be any number of transactions between you and the grocer on the payment channel. At the end of the month, the final net settlement transaction (depending on who owes whom) to transfer crypto back from the payment channel happens on the main blockchain. This has the ultimate result of reducing multiple transactions between parties to just 2 on the main Blockchain, and the rest of the transactions occur off-chain. Off-chain transactions are faster and cheaper than on-chain transactions because there is no need to wait for block confirmations and pay a miner fee to include such transactions in the next block.

A side benefit of Raiden (or payment channel technology) is that it will make micro-transactions much more feasible for uses such as IoT. Micro-transactions may not always be viable on the main chain because miner fee has to be paid for every transaction. For example, if one has to pay $10 processing fee for a transaction worth $1, it is not acceptable.

Source: Plasma: Scalable Autonomous Smart Contracts whitepaper

Plasma (Layer 2): A Sidechain is a separate blockchain that is attached to its parent blockchain using a two-way peg. A 2­way peg (2WP) protocol is a protocol that allows transfers of a cryptocurrency from a main blockchain to a secondary blockchain and vice­ versa at a predetermined rate. The original blockchain is usually referred to as the ‘main chain’ and all additional blockchains are referred to as ‘sidechains’.

Plasma is a hierarchical tree of chains (sidechains) in a parent-child relationship. Child chains are created on the Ethereum Blockchain , with their own validators. Child-chains can spawn their own child-chains, who in turn can spawn their own child-chains, and so on. Since there’s no limit to how many blockchains you can put in the tree, this allows for near infinite scaling. Once the plasma tree of blockchains is ‘accepted’, the hash of all of those is stored on the main chain.

This way you would know if someone tried to tamper with the sidechain data, as the hash of the sidechain would not match what is stored in the main chain. With Plasma, we could perform many complex operations at the child-chain level, running entire applications with many thousands of users, while only interacting with the Ethereum main-chain occasionally. The end result would be an increase in transactions per second and lower transaction fees. The drawback in this approach is that you may have a lower level of decentralisation and security in the sidechains, compared to the respective guarantees you get from the main Ethereum chain.

Plasma and Raiden are both scaling techniques for blockchains. But there are several differences. Raiden uses state channels (which are not separate blockchains) but Plasma uses sidechains (which are separate blockchains). Here is an excellent article explaining the differences, but to wrap your head around this, you can view state channels as a technique used for low-fee, scalable and highly secure payment transfers , while sidechains have broader capabilities.

In summary, scalability improvements from layer 1 and layer 2 initiatives are expected to have, not additive, but multiplicative effects. So, if 100x improvement is obtained from Sharding and 100x from Plasma, this would increase the scalability of Ethereum by 10,000x, which should be powerful enough to handle most current-day scenarios . Beyond that, for reaching one million tps or more, Ethereum team is pursuing more advanced techniques like quadratic sharding.

To read the complete piece, please click here.

And don’t forget to subscribe to our updates!