Fireside with Linera: A new multi-chain layer-1 built for scalability
"Scalability inherently requires exploiting parallelism" - Joseph Bonneau, Research Partner at A16z Crypto
Iterative Ventures is a web 3 investment fund & community with the aim of supporting digital asset adoption. We have a focus on infrastructure and technology owing to our technical background.
Please reach out (email@example.com) if you would like to learn more or collaborate, including being featured in our community events/ media outlet.
Thanks for reading Iterative Ventures! Subscribe for free to receive new posts and support my work.
Follow us on Twitter and join our Telegram community to get the latest update!
Big thanks and shoutout to SignalFire and Josh for hosting us at their San Francisco office for the fireside chat.
This week’s article will be a summary of the fireside conversation as well as diving into the problem that Linera seeks to address at a very high level without going into step-by-step execution mechanism.
Here is the recording from Linera’s fireside chat. Thanks again to Mathieu for accepting our invitation to speak and his feedback on this article.
In general, blockchain architectures are divided into execution and consensus layers. However, thanks to advances in Byzantine Fault Tolerant (BFT) consensus protocols (See Bullshark consensus mechanism for example), the new bottleneck for transaction rate appears to be the sequential execution of transactions rather than consensus ordering. This lies at the heart of Linera’s focus and design.
At the height of web3 and crypto bull run from 2020 until early 2022, many have witnessed the limitations of existing blockchain infrastructures. Some of the common issues surfaced within existing heavyweights (such as Ethereum) include:
High gas fees: Many have experienced or have heard of high gas fees. In a nutshell, high gas fees are the result of network congestion where each transaction is competing to be included in the next block (thus the blockspace scarcity problem).
As a user, what this means is that during high usage periods, it can get quite expensive to make a transaction.
For projects, this can be discouraging as users will have less incentive to make transactions and trade thus decreasing the overall profitability of a protocol or project.
For context, users spent $9.9 billion in transaction fees in 2021. Source
Unpredictable response time: Furthermore and stemming from the same root cause, as transactions are competing for block space, users may have to wait for longer than usual in order to execute the transaction.
Worse yet, there may be a chance that the transaction may not be included at all.
Furthermore and with many existing blockchain architectures designed with mempools (where transaction data are stored before confirmation in a block), there presents the opportunity for validators (miners if on a proof-of-work chain) to extract value by arbitrarily reordering, including, or excluding transactions within a block at the detriment of users.
In the aforementioned Paradigm Research’s article, it is mentioned that “[a]ll else equal, MEV causes gas prices to be higher than they would be if MEV did not exist. This implies that miners are already indirectly extracting a fraction of the total MEV that gets extracted by traders, estimated at ~12% today.”
*This article goes in-depth describing MEV.
Due to the aforementioned issues (especially regarding network congestion), newer generations of blockchains all have various angles to address scalability issues through parallelization. (Other solutions including transaction batching and compression such as roll-ups, collectively called layer 2, are also popular solutions as of late.)
Continuing on the previous quote by Joseph Bonneau, scalability means the ability of a system to improve performance by adding resources.
It is different from the common measure of transactions per second (TPS) as defined and published by many other blockchains.
On the flip side, for a system to scale through adding resources, the state feeding into the system to be processed also need to be designed in such a way to allow independent data structures (such as sharding in other blockchains) - splitting state into pieces so that different process units (validators for blockchains) can process independently.
Ideally, if we add one more processing unit, the throughput will also increase proportionally. However, the relationship is not linear as validators will also need to communicate to reach quorum as well as to update states across other validators to the same state.
These were the considerations taken in designing Linera.
Linera’s design and consideration
For blockchain systems, the state of an application is replicated across several partially trusted nodes called validators. A transaction is submitted to the validators when a state change is required (eg. an update to a user’s wallet). Typically, the proposal and validation of a block is done by validator nodes.
Over the last few years however, there is a gradual shift from operating a single chain in a blockchain system to a multi-chain world for better scalability and better security (as bridges are notorious for getting hacked. Read more here).
Multi-chain, however, can mean many things such as roll-ups on the Ethereum blockchain. (Read more here.)
At Linera, the state of Linera applications are organized across many parallel chains of blocks, called microchains. Furthermore, the responsibility of extending a chain with new blocks is assumed by the owners of the chain as opposed to validators and is separate from block validation.
There are various types of microchains that Linera supports:
Single-owner chains: Only one user (Likely to be the chain owner as identified by its public key) is authorized to propose blocks
Permissioned chains: Where only a well-defined set of cooperating users are authorized to propose blocks.
Public chains: any user can propose operations to be included in the next block by validators.
(See here for comparison across public, private, and permissioned blockchains)
The microchain design allows for Linera validators to be efficiently sharded across multiple workers. This is because transactions for one user in a chain don’t necessarily need to be ordered in consideration of other transactions. Inside each Linera validator, the workers responsible for particular subsets of microchains can therefore execute user transactions in parallel.
As each validator is free to add internal workers when needed, this allows Linera to scale horizontally in response to higher than normal network traffic.
Elimination of mempool
With the aforementioned issue surrounding mempool (MEV, network congestion etc), by allowing users to manage their own chains, this design allows Linera to avoid having a mempool in most use cases, as users submit their transactions directly to the validator and fully control the processing time.
Because of this design, the only external parameter affecting responsiveness (the time between when a transaction is submitted to when a transaction is executed) is therefore the network round-trip-time (or colloquially - network latency) thus achieving better control over system response time.
Running all the chains in the same set of validators also helps with cross-chain communication (see here why this is important. eg. atomic swaps) as messages are exchanged via point-to-point communication across co-located workers that trust each other and require much fewer sources. In comparison, in traditional sharded blockchain systems, cross-chain communication is done across distrusting nodes.
There has been a lot of discussion of what it means to be web3 as many in today’s crypto world are deemed as web 2.5 - that they are partially on-chain.
So what does it mean to be fully on-chain and web3 native?
The litmus test that I have been hearing is that when the developers for a project disappear, the project will still function as usual.
This means that state changes should all be registered on-chain thus to achieve decentralization.
As with the limitations, including network congestion, cross-chain security and efficiency in communication, and limitation of block size (not mentioned in this article, but you can read more about Ethereum’s block size limit here), there are still a lot of technical challenges to be resolved.
Linera is a next-generation blockchain infrastructure that seeks to address some of these issues.
Here is Linera’s blog, whitepaper, and Twitter for further learning and updates.
Shoutout to NEAR SF and OnePiece Labs as our channel partners for spreading the word. We look forward to seeing more of you at our next event! As always, keep iterating 💪
2/7 - Metaverse and Gaming Conversations and Demos - RSVP
[Partner event] Networking, Drinks, and Metaverse and Gaming conversations by Open Future Forum
Meet founders in the Metaverse and Gaming spaces and share what's new.
If you have made something cool to share in the Metaverse and Gaming space OR something made by someone else please, come and share / demo. It is in a bar, but please do feel free to bring your laptop to share.
2/10 - Intro to deal structuring with Andrew and Adi, Goodwin - RSVP
One of the pain points a first-time founder has to deal with is to understand the intricacies of a deal. What do you do, and how do you understand when you receive a term sheet from an investor? And/or how do you negotiate?
A16z wrote about this article a few months ago on the various deals and structures that can be offered.
We will have a talk about this with Andrew and Adi from Goodwin.
Time/ Date: Fri, Feb 10, 2023, 11 am PT
Virtual Event: a link will be available once you sign-up here
2/21 - ChatGPT and AI + ML Revolution - RSVP
[Partner event] Meet people in ChatGPT and AI + ML Revolution by Open Future Forum
If you have made something cool to share in the ChatGPT and AI + ML Revolution space OR something made by someone else please, come and share / demo.
2/27 - Fireside chat with Evan from Mysten Labs / Sui - RSVP
This is a fireside chat with Evan at Mysten Labs to learn more about the Sui ecosystem. Sui is one of the premier next-generation parallel processing blockchains backed by A16z (read more here)
Time/ Date: Mon, Feb 27, 2023, 7 pm PT
Location: 44 E 3rd Avenue, San Mateo, CA (Draper University)
March will enable withdrawals of 17.2M ETH staked or deposited in the Beacon Chain since December 2020. Only 43,200 ETH can be unstaked per day, however, the total staking reward of the past two years, ~ 1M ETH (1% of free-floating ETH), can be withdrawn instantly. (Source)
The California DMV is teaming up with Tezos blockchain for digitizing car titles and transfers. In partnership with blockchain software firm, Oxhead Alpha, the DMV aims to modernize its record-keeping and reduce the potential for fraudulent activities from car sellers. The agency hopes to have the system up and running within the next three months. The next step will be the introduction of digital wallets for non-fungible token car titles, with the DMV overseeing the transactions. (Source)
Thanks for reading Iterative Ventures! Subscribe for free to receive new posts and support my work.