<h2>Sei network:&nbsp;Built for DeFi</h2>We often try to categorize blockchains into two categories: general purpose and app chain. Chains for general-purpose innovation and app chains for permissioned and particular use cases. However, "app chains" are not binary, but rather a spectrum that is up to the chain's choice. Sei Network is a new Cosmos chain that aims to be a "DeFi-optimized" layer 1.What does this imply? It entails making major changes (and making sacrifices) to the base layer in order for DeFi apps to thrive. A built-in order matching engine, instant finality, order parallel processing, single block order execution, and other featuresAll of these changes are made at the base layer. To be clear, Sei Network is a layer 1 optimized for DeFi, not a DEX. Not a "pure" app chain like THORChain, which concentrates entirely on cross-chain swaps, but one meant to support apps like DEXs, perps, and options.We may turn to Serum and Solana to see why you would wish to make these modifications at the base layer. Solana is a general-purpose layer 1 promoted as bringing NASDAQ on-chain, with block speeds targeting 400ms and extraordinarily high throughput. Solana's key assumption is that order book exchanges will ultimately replace AMMs, and the on-chain metrics confirm this. Serum, an order book Dapp developed on top of Solana, is the most frequently used entity, accounting for one-third of all Solana program executions. Serum is Solana's "order book layer," which projects such as Atrix, Bonfida, Jupiter, and others may use. When most people think of Solana, they picture Serum.<img src="https://cdn.builder.news/news/11ed4dcac1ad54b4a8e6023f659fee92.jpg" alt="Most used projects on Solana" title="Most used projects on Solana" width="600" height="306" class="lykafe0">Most used projects on Solana. Source: NansenThis architecture, however, has certain downsides. Because Solana is a general-purpose network, Serum (and the programs developed on top of it) is always fighting for resources with other apps. Activities unrelated to Serum, like gaming and NFT minting, can generate chain congestion and, as we've witnessed a few times, bring it to a standstill. Sei Network "eliminates the fat" by removing all non-DeFi activity from their chain.&nbsp; There is an opinion that Sei Network is Serum's own L1; making special compromises to improve the base layer for DeFi and providing these programs an "unfair edge" over non-DeFi apps.Because apps that seek to build on top must be whitelisted by governance, Sei will not be as permissionless as Solana. While part of the joy of permissionless innovation is lost, you may design a better optimal environment for your specialty. Sei's architecture includes a native order-matching engine, a pricing oracle, parallel execution, and single block order execution. Sei network is an app chain, but don't be misled; its on-chain order book enables a composable architecture, allowing CosmWasm programs on Sei to have simultaneous composability with one another and share liquidity via the native order-matching engine. Everything else is async composable in an IBC-enabled Cosmos chain. And, while we're on the subject of Solana, there's a unique Sei Solana perspective that we'll discuss at the conclusion of this article.Sei network achieves some of its improvements utilizing ABCI++, Cosmos' planned enhancement to ABCI that enables for programmability of every phase of consensus. Sei has been working with ABCI++ to enhance three things: optimistic block production, intelligent block propagation, and parallel execution.&gt;&gt;&gt; Read more:&nbsp;<a href="https://www.builder.news/crypto-insights/cosmos-vertical-scaling-solutions-KluWM" alt="Cosmos: A fresh wind with vertical scaling solutions" title="Cosmos: A fresh wind with vertical scaling solutions" target="_blank" rel="" class="">Cosmos: A fresh wind with vertical scaling solutions</a><h2>ABCI ++ Interface</h2>With an emphasis on order books, market makers prioritize block timings, finality, and latency. Because market makers must update their prices every block, shorter block timings result in smaller price updates between blocks, narrower spreads, and less risk for market makers. Anything more than a few hundred milliseconds is unworkable. Order books are suboptimal on a regular Cosmos chain because of the 6-second block times. However, the beauty of Cosmos is in its modifiability, and Sei network has been working hard on making improvements to push consensus and make it as quick as possible (goal 300-600ms). The three primary areas of concentration have been:<h3>Optimistic block production</h3>Sei Network does this by utilizing ABCI++. ABCI is the interface that exists between an application and consensus. Its primary function is to execute blocks that have been agreed upon by consensus. ABCI only engaged with consensus at decision time, giving it limited control over which transactions to select from the mempool. ABCI++ increases programmability at every phase of consensus, enabling applications to rearrange, modify, drop, postpone, or add transactions, as well as shorten block delays by providing the ability to build blocks optimistically.Following the proposed consensus phase, the application can begin optimistically executing the block in conjunction with the pre-vote and pre-commit steps. Sei Network will then begin "optimistically" implementing the state modifications to a temporary candidate state until unanimity is reached. If it is not approved (which is unusual), it is discarded. This stage involves a large amount of data processing, which can be time-consuming. By optimistically processing state changes, we can drastically shorten block timings and minimize latency (30% for 1s blocks).<img src="https://cdn.builder.news/news/11ed4dcac1ad59fda8e6023f659fee92.jpg" alt="Sei block production 1" title="Sei block production 1" width="600" height="450" class="lykafe1"><img src="https://cdn.builder.news/news/11ed4dcac1ad5d87a8e6023f659fee92.jpg" alt="Sei block production 2" title="Sei block production 2" width="640" height="671" class="lykafe2">Optimistic block production. Source: Sei Founder<h3>Intelligent block propagation</h3>In addition to producing blocks optimistically, they are also improving block propagation. In Tendermint, when a block proposer offers a block, all transaction data are included. This is a large amount of data. However, validators already have access to 99.9% of these transactions via the gossip mechanism in their local mempool and do not need to wait for them from the block proposer. Instead of transmitting complete information, the proposer will now simply submit a hash of each transaction in the block, and the validators will indeed be able to rapidly rebuild the block using the transactions within their own internal mempool. Sei Network refers to these two adjustments as "Twin-Turbo Consensus," and argues that adopting them improves performance by 83%. (optimistic block production &amp; intelligent block propagation).<h3>Parallel execution</h3>The third change to the block manufacturing process revolves around parallel execution. The sequential order procession of cosmos chains using ABCI. Orders are handled one by one in a sequential procedure, independent of the market, drastically reducing throughput and dramatically increasing delay as the demand increases. Unique markets that do not overlap can be handled concurrently using a parallel method. Instead of executing market B's initial order aftermarket A's, you can process both orders simultaneously. Orders inside a single market must still be processed sequentially in order to avoid non-determinism, which occurs when two separate validators provide different outcomes for the same state.Sei network did some stress testing around parallelization (as well as colocating validators) to evaluate what type of gains in block times, latencies, and throughput they might obtain. In general, they were able to witness 75-90% reductions in block durations vs. sequential and 40-120ms latency for parallel vs. 200-1370ms latency for sequential by parallel execution. They were able to cut block durations from 1.33s to 0.81s, latency from 371ms to 48ms, and TPS from 7,500 TPS to 12,200 TPS under the 10,000 orders/block scenario with 20 separate contracts (markets). Significant gains may be demonstrated across all load levels (orders/block), with bigger marginal gains as the load increases.Sei network has added extra features to the basic layer, in addition to the three major advancements mentioned above:<ul><li>Native Price Oracle: Oracle is integrated into the base layer; validators must agree on pricing before committing to a block. The block will not be created unless validators reach an agreement on the pricing. Allows other modules to get consistent price feeds from on-chain marketplaces.</li><li>Order Execution in a Single Block: Enables an order to be submitted and processed in a single block (in Serum it takes multiple blocks).</li><li>Order bundling: allows market makers to update prices across different marketplaces in a single transaction.</li><li>Frequent Batch Auctioning: At the completion of the block, market orders might be aggregated to clear at a single price, with the purpose of minimizing frontrunning.</li></ul>Sei Network has been evaluating smaller validator structures and increasing hardware requirements, in addition to software advancements. While there are sacrifices in decentralization, they are offset by large performance advantages, highlighting what distinguishes Cosmos: customizability.<h2>Higher hardware requirements</h2>The suggested requirements in the original edition of the Sei documentation were identical to a normal Cosmos chain. They've recently been revised to enhance those requirements even higher in certain load testing. Order books may be highly demanding on hardware, and lower-spec computers have impacted overall testnet performance. While not Solana-level standards, Sei Network has stated that they want its validators to go above and beyond what we see in standard blockchains. They are also aiming for smaller validator sets. While they have received 2,500 validator applications for testnet, they are now cycling between 250 validators at a time, with groups of 50 verifying at a time. Furthermore, they are advocating for validators to be colocated in order to minimize latency even further.<img src="https://cdn.builder.news/news/11ed4dcac1ad6523a8e6023f659fee92.jpg" alt="Sei hardware test" title="Sei hardware test" width="600" height="282" class="lykafe3">Source: Delphi digitalWhy should you co-locate? With globally scattered validators, messages take longer to travel between them, resulting in increased latency for reaching consensus and creating blocks. Latency is the name of the game in high-frequency trading, and an order book exchange must minimize it as much as feasible. Sei network has released the results of several of its colocation testing once more. What are the important takeaways?<ul><li>Colocation decreases latency by 46% when compared to international dispersion.</li><li>The number of validators that may be added without increasing latency is limited to 50.</li></ul>Although there are clear drawbacks to having all validators in the same geographic location, the performance improvements are difficult to reject. When Sei network releases mainnet, they will most likely move in this colocated, smaller validator set path. Although testing for both the software and hardware designs is still continuing, interested teams are considering building on Sei network. When the mainnet goes live, they will have a slew of protocols ready to go.<h2>Sei Network Ecosystem</h2>Sei network&nbsp;has revealed 50+ projects that are interested in developing on Sei. While just a handful have been disclosed, significant ones include Vortex, UXD, and Axelar. The fact that these three initiatives emerge from diverse ecosystems is fascinating.<ul><li>Vortex: Vortex is the result of Retrograde's fusion of Cosmos and Terra-based projects. Retrograde's goal of being the "Convex of Terra" has shifted to becoming a top perpetual exchange on Sei. Again, Sei is not a DEX in and of itself, and projects like Vortex are required to build on top of and leverage Sei's foundation layer infrastructure. They are the major perp protocol announced and are now live on testnet.</li><li>UXD: Solana's UXD is a fully collateralized stablecoin. Users generate UXD by depositing an asset (ETH, BTC, etc.) and then shorting a perp on the opposite side with UXD. UXD will work with Vortex to handle the short side of the transaction (similar to how they use Mango on Solana). Launching on Sei network offers UXD access to IBC and the larger Cosmos ecosystem, and stablecoins are a major missing piece in Cosmos DeFi. Kujira's freshly released stablecoin USK and Agoric's IST are both on the way, however, both are CDP-style, which can be difficult to expand owing to capital inefficiency. There is a significant market opportunity for UXD to extend outside of Solana</li><li>Axelar: Axelar is a Cosmos cross-chain communications protocol whose primary application is to bring EVM assets to IBC. Osmosis employs Axelar as their fundamental bridge/asset provider for BTC, ETH, and USDC, and Sei's integration of Axelar's wrapped assets contributes to Sei's supremacy in the Cosmos ecosystem.</li></ul>Other protocols, such as data solutions and synthetics, are mentioned below, with more to be released in the following weeks. Nitro, a Sealevel (Solana's VM) execution environment that has settled on Sei as a Sei L2, is a recently announced project that is highly fascinating.Nitro has caused some misunderstanding since it has been referred to as a "Solana L2," although the most correct definition is a "Sei L2 running the Solana VM." This is an optimistic rollup that runs the Solana VM Sealevel (the same as on Solana) but as a Sei rollup. Sei's objective is to facilitate the transfer of Solana developers into the Cosmos ecosystem, with future iterations potentially releasing as general-purpose or app-specialized rollups on Solana or Ethereum. While Nitro v1 doesn't really settle on Solana and so does not collect SOL gas costs, it is a positive step in terms of the SVM gaining more mindshare from the EVM.Today, the EVM is by far the most used execution environment. If you feel that the EVM is Ethereum's moat (as many do), then the SVM being standardized for rollups is a good thing for Solana. Solana's greatest innovation, one may say, is in its execution engine. Nitro gives weight to that (and Eclipse, informed shortly after on Celestia).Because Nitro is a Sei L2, it must utilize Sei's token for data availability and settlement fees. In terms of a timescale, we've seen how long progress has taken with the Ethereum rollups, and an extra twist here is that the fraud process for a Sealevel rollup will look a little different, given there is no global state source to commit to.<h2>Sei Network token use case</h2>As an L1, the Sei token will be utilized for staking and gas fees in the same way that any other PoS blockchain would. However, in order to keep fees low, value accrual is kept to a minimum (as we see on Solana). MEV redistribution is predicted to be the largest component of value accrual. Sei network proposes to use a private off-chain flashbots-style auction in which bots may bid on common MEV transactions like liquidations and arbitrages.What about airdrops? Since July, Sei network has been conducting an incentive testnet for two months and has now extended it for another two months, with the goal of distributing 1% of the supply to users. Sei has indicated possible airdrops for the Cosmos, Terra, and Solana communities in addition to the testnet. More announcements are planned in the future.<div class="pr-block-special">Disclaimer: This article is intended for and only to be used for reference purposes only. No such information provided constitutes advice or a recommendation that any investment or trading strategy is suitable for any specific name.</div>

Sei Network aims to be a "DeFi-optimized" layer 1. It entails making major changes to the base layer in order for DeFi apps to thrive. 

Sei Network: A Layer 1 Optimized For DeFi