Zee Prime Capital: Further Discussion on the Specific Categories and Development Logic of Middleware

Author: Cam, Zeeprime Capital
Compiler: Biscuit, Chain Catcher

Introduction: A year ago, Zee Prime Capital co-founder Matti Gagliardi wrote an article titled "Infrastructure Lego: Middleware", which pointed out that middleware is gradually encroaching on the crypto industry from the inside out. It cited the definition of middleware by open-source software provider Red Hat: Middleware is software that provides common services and functions for applications, handling data management, application services, messaging, authentication, and API management. Middleware can help developers build applications more effectively, acting as connective tissue between applications, data, and users. At the same time, Zee Prime Capital also elaborated on the operational mechanism of middleware using Pocket Network as an example.

Recently, Zee Prime Capital has elaborated on the current categories of middleware and the investment logic and development logic of the entire middleware track, in conjunction with some middleware projects they have invested in. Below is Chain Catcher's translation of the article:

As blockchain DeFi projects continue to be developed into more advanced applications, the demand for infrastructure such as middleware is also expanding. For years, Zee Prime has been optimistic about middleware products. Our previous article "Infrastructure Lego: Middleware" emphasized the importance and economy of data relays in developing decentralized platforms.

This article will expand on the current middleware ecosystem, particularly decentralized applications. While public chains like Dfinity are key end-to-end solutions to the problem, we can only use middleware technology to bridge this gap until Dfinity achieves widespread adoption.

Since the early days of DeFi, a plethora of data networks, indexers, access control, and other middleware tools have emerged, which are key adhesives for the next generation of applications. As the complexity of DeFi applications continues to increase, we have seen an explosive growth in developers' demand for infrastructure.

Middleware is a vague concept that is difficult to describe in specific use cases. Fundamentally, middleware refers to any project that supports other higher-level applications. The last time we discussed this topic, we emphasized that middleware enables connectivity within applications, but mapping this theory of interactive networks onto a 2D plane can be quite chaotic.

Is this a map of Web 3 middleware?

In this middleware article, we will update other middleware categories that were not covered previously, while also clarifying their use cases and fundamental principles for builders. This is not an exhaustive list, and more middleware categories may be included in the future.

Storage and Data

Storage is one of the key elements in the decentralized application stack and serves as the foundation for data computation. The Cambrian explosion of Web 3 requires more storage solutions, not just simple account state recording at the base layer. Decentralized applications tend to reduce centralized points of failure or seek solutions related to Web 2 censorship.

Every application requires services provided by Web 3 middleware. However, the challenge lies in development and operations, and not every developer possesses the expertise to implement Web 3 middleware in their projects. Therefore, developers need to be able to more easily assemble these infrastructure Lego blocks into new projects.

This reflects the challenges faced by cryptocurrencies. Concepts such as wallets, mnemonic phrases, and gas are inherently unfriendly to users, and developers must handle the complexities for ordinary users.

Storage networks like Arweave and Filecoin are already operational, providing a distributed matching system for the supply and demand of storage, serving as foundational modules for any decentralized technology stack. Storage can be divided into two types of protocols: base storage layers and aggregators, which serve as scaling solutions and facilitate broader adoption.

Banyan DAO, as an important aggregation layer for storage networks, focuses on proxy storage and improving the economic incentives of existing storage protocol bridges, ensuring that applications can utilize Web 3 storage in a network-agnostic manner and track their usage sources.

Banyan also serves as a potential bridge for Web 2 applications to integrate web3 services, providing decentralized storage solutions. Currently, the implementation of these storage solutions is quite complex, and Banyan's abstraction layer and market resources enhance the accessibility of DeFi Lego blocks.

Additionally, Spheron Protocol acts as an abstraction layer for middleware solutions, aiming to be a one-stop shop for deploying and automating Web 3 projects. It features an "app store"-like interface, allowing Web 2 users to easily select decentralized infrastructure products.

Data Models and Validity

Blockchain is a state machine that continuously changes its state and generates data during the execution of computations. Over time, the number of accounts, states, and smart contracts has rapidly increased. The explosive growth of data can lead to various issues, from indexing to the synchronization and backup of initial nodes, ultimately affecting the scalability and security of the underlying state machine.

KYVE is a decentralized data lake protocol (for unstructured raw data) used for storing, validating, and retrieving data streams. KYVE utilizes the Arweave network to provide storage support for applications and protocols, offering more refined solutions for specific data validity.

Due to the long synchronization time of initial nodes, which continues to grow, if the number of validators decreases and it becomes difficult to access new nodes, the security of the entire blockchain network may be jeopardized. The market fit of KYVE's initial product is node synchronization. By providing easily retrievable and verifiable archival state data, decentralized applications can significantly reduce initial node synchronization time, allowing new validators to be added at any time while maintaining network security.

While we have introduced where and how data is stored, we also need to consider data models and parsing principles. For applications built on these state machines, the data generated during activities may require storage and computational resources beyond account balances.

Ceramic Network is a decentralized underlying data model network, and its importance to decentralized applications lies in providing solutions for L1 storage account balance states. KYVE aims to provide data validity for L1 state transitions, while Ceramic's goal is to store the state and model of application data beyond account information at the base layer. This solution allows users to create collections of IPFS data (streams), enabling static data (such as data on Filecoin or Arweave) to become dynamic data.

Furthermore, Ceramic offers composability of data models through its open-source API. For instance, Ceramic proposes a data standard that can be used across applications, bringing the same composability to data as seen in DeFi. Creating data Lego in a monetary Lego manner allows social networks like Facebook to utilize their open APIs.

The decentralized SQL database Kwil adopts a traditional SQL-compatible approach to enable web 3 data models. The biggest advantage of this model is that it makes it easier for SQL developers to get started. Kwil uses a network of nodes to maintain relational databases, which act like moats, maintained by subnetworks of nodes, and keeps nodes updated by scanning write and query events. Nodes can run a high-level request gateway to enable effective logical execution for database interactions.

Indexers

As applications and networks generate an explosion of data, decentralized applications need an interpretive layer. Similar to the early internet, users had to manually remember and maintain an IP address book, later products like DNS and search engines provided readable indexing layers.

The scale of indexing data has grown with the development of the internet, and data querying capabilities have become more user-friendly. Similarly, indexing functionality is crucial in L1 blockchains and storage networks. Due to the nature of distributed systems, data may be scattered across different locations and difficult to retrieve. The indexing layer helps accelerate the querying process and create standardized procedures.

Zee Prime portfolio company Subsquid focuses on Web3 data indexing, with the ultimate goal of becoming the next-generation API for Web 3. The project adopts a decentralized multi-layer approach to index on-chain data, supporting the Substrate and EVM ecosystems, defining the types and patterns of on-chain data, and subsequently switching the calling method for new indexed data from RPC calls to API calls, enhancing the retrievability of new indexed data.

The layering consists of two types of nodes: Squids classify data and support subsequent API queries, while Archives continuously extract raw data from the underlying state machine and save it to the database.

Similarly, SolanaFM is an indexer serving the Solana ecosystem by processing raw blockchain data into a queryable format. Similar to Graph and Subquery, these two solutions target various terminal markets.

The migration of Web 2 applications to Web 3 brings a large amount of data. While the increase in data aids the development of Web 3, it also presents a daunting task for developers and communities regarding data storage and indexing. Glitter Protocol addresses another issue: decentralized storage. Glitter provides barrier-free indexing services to exchange crowdsourced data, offering a win-win solution for developers and communities. This model has proven effective in collaboration with several social applications that store data on Filecoin.

Access Control

Access control is one of the most important yet missing infrastructures in Web 3 applications. This is an important philosophical question: which types of users can see all content on the internet? It becomes increasingly important when considering the security issues of national/company/personal sovereignty. The semantic characteristics of public blockchains/Web 3 technologies enable users to better distinguish their access content and how to access it. Although these systems have inherent openness, access control frameworks will allow for encryption/decryption based on specified supply frameworks.

Lit Protocol aims to address this issue by using threshold cryptography, allowing users to provide some public credentials (such as NFTs in their wallets) for access to network resources and content. The network of nodes running the protocol is used to verify proofs and approve interactions, validating both the provided proof and whether it meets previously set access control conditions. Once verified, users can access the requested content. To some extent, the Lit protocol is seen as a reading solution for Ceramic.

Guild.xyz is also attempting to solve the access control issue from a different angle. Initially, Guild focused on creating Discord servers for decentralized projects, but it has now expanded into a multi-chain access control gateway.

Integration Platforms

To further achieve blockchain integration in the 3D bridging world, the Polywrap development platform (formerly known as Web3api) makes the integration of Web 3 protocols more efficient. While Web 3 protocols have openness and composability, achieving this composability is much more challenging than in Web 2 applications. This is because each protocol runs specific business logic and is often packaged into SDKs in specific languages.

Due to the lack of standardization, integrating different SDKs is highly inefficient. Additionally, SDKs in specific languages mean that protocol developers often publish redundant SDKs in multiple programming languages, making maintenance difficult.

Polywrap's solution is to leverage standardized patterns and WASM to alleviate the burden on applications. Developers do not need to preload various protocol SDKs into their applications; instead, Polywrap integration provides developers with easy-to-read patterns for calling applications (similar to REST APIs). This allows applications to download SDKs only and execute the module package when needed. This means that applications equipped with Polywrap can access Web 3 protocols seamlessly.

The user experience of Web 3 applications is still not refined enough. As emphasized above, inputting gas fees can lead to a poor user experience. By integrating the multi-chain protocol Biconomy's API, applications can enhance this user experience. The Biconomy platform provides a range of tools to enable gasless transactions, fast transactions, and instant cross-chain transactions.

Biconomy achieves gasless transactions through the use of the ERC2771 protocol and clever forwarding design. Cross-chain functionality is supported by on-chain liquidity pools, using off-chain servers (executor nodes) to monitor cross-chain transactions in the liquidity pool, releasing assets on the other end after verification.

To provide the next billion crypto users with a smooth user experience, these types of tools are crucial. Our goal should be to continuously strive for a more seamless interaction process between Web 3 systems.

Although not explicitly categorized, Sepana is building a search engine for Web 3. Whether it's DeFi, SocialFi, DAO, or NFT, Sepana's solution is to provide a search engine for Web 3 applications and data, allowing users to browse the entire Web 3 data, with the protocol serving as a gateway to various ecosystems.

Additionally, Sepana's open-source algorithms can be used to enhance other applications, such as social media feeds in database solutions like Ceramic or Kwil, allowing users to adjust the display and acceptance of content on social media based on specific sentiments using open-source algorithms.

How Does Everything Fit Together?

The business models of most modern tech companies and applications can be summarized as data production/extraction, data/model control/distribution. The smooth user experience of modern web applications is based on these fundamental processes.

For the workflow of data processing, we hope middleware solutions can develop in demand and support various decentralized products in the Web 3 environment. Through the projects/middleware types listed above, we can clearly see their position in data processing.

In reality, many categories span multiple ranges, and due to these overlapping attributes, it is currently difficult to accurately define middleware categories.

A common example is social media networks, and this model can be extended to a broader Web 3 middleware stack.

We assume a social media network will be named twatter, and its product consists of the middleware stack components shown in the above diagram. Note that we do not consider social Web 3 to be "Twitter-but-decentralized." We believe Social Web 3 is more like an emerging form of media, and it may even reference Web 2 applications for verification in decentralized privacy login systems like Sismo (if Web 2 companies are willing to open their APIs).

In its most primitive form, all data on the platform (usernames, profile pictures, historical activities, social graphs, etc.) can be stored and indexed in IPFS format on storage networks, with the data model stored in Ceramic or Kwil. Accounts based on database solutions for Twatter will have all the previously mentioned data models.

For example, if the platform requires users to use a free-minted NFT to access the platform to reduce spam, then users need to connect their wallets to the platform first, and the access control protocol will verify before displaying. Twatter can also integrate platforms to natively enable other web 3 services or utilize Sepana's algorithms to design social graphs.

The process might look like this

At the time of writing this article, we discovered an application called Orbis Social, which has built a social network using the aforementioned stack technology. Next-generation applications are under development, and we hope to see more unique use cases in the coming months.

A key point in the above diagram is that the further right the process goes, the more it becomes chain-agnostic. This structure allows applications to exchange data with competing products, in contrast to the Web 2 monopolies, which ultimately redistribute this pseudo-standardized composite value back to platform users.

Next-Generation App Builders and Web 3 Middleware

As Web 3 tools continue to emerge, we need to constantly ask whether these tools truly facilitate Web 2 personnel.

Web 3 middleware should adhere to the same fundamental principles as early crypto pioneers. Teams should choose Web 3 middleware based on its merits, whether in terms of security, durability, or censorship resistance, the advantages of web3 middleware are very prominent. There are even features of Web 3 middleware that we cannot imagine, which may unlock more new functionalities for applications.

These infrastructure Legos can enable deeper network integration. Tim Berners-Lee believes that an open and composable internet can provide cheaper management and computing solutions than Web 2 counterparts. As Dennis Nazarov pointed out, a complex computing system can be built through modular and specialized infrastructure. In the worlds of Web 1 and Web 2, users sacrificed the ability to manage state information to connect to the web, while Web 2 giants privatized state information to capture more value.

Public state machines will disrupt this model by maintaining state information in an open manner and introducing token economic models to enhance consistency on both sides. This is the natural law of resisting capital plunder.

Zee Prime's Perspective

In many ways, middleware belongs to the B2B segment of the crypto industry. Therefore, middleware solutions are often highly technical and non-intuitive for the typical end user (the user is not the direct target audience). Instead of focusing on new DeFi protocols, NFT projects, or GameFi studios, we believe that focusing on infrastructure is crucial for developing new applications.

In summary, these infrastructure Legos (and future Legos) will perform the following operations:

• Increase censorship resistance
• Facilitate positive-sum economic games
• Improve efficiency
• Innovate business models

Another potential impact of these interchangeable infrastructure modules and abstraction layers on applications is:

1. Moving further away from the underlying layer;
2. Becoming more chain-agnostic.

This is not a rebuttal to fat protocols, but rather focuses on the impact of the continuous advancement of middleware technology. In principle, this can be seen as reducing switching costs. On-chain applications (primarily DeFi) will have the same nature as base chains (i.e., products built on financial accounts).

More complex non-financial applications will have unbounded connections with such blockchains, thereby reducing switching costs. For instance, the functionality of free-minted NFT access control is very easy to port to new blockchains and wallets, and some applications are already using this feature to attract users across chains.

We firmly believe that increasing value transfer in information transmission is very meaningful, but to achieve this process and improve the variety of applications and user experience, a large number of infrastructure Lego blocks are needed.

When discussing middleware investment, value capture is one of the most controversial topics. In a sense, a truly critical middleware looks like public infrastructure, although some might argue that this also applies to certain successful applications (like Twitter wanting to be public infrastructure).

Therefore, one might expect the profit margins, usage fees, and revenues of middleware to be minimized, but a more reasonable approach is to investigate the range of fees acceptable to the public.

Although middleware may seem unattractive, in the context of a global technological revolution, it is easy to reach the scale of billions of dollars in Lego blocks. Since middleware performs specific functions for applications, their TAM is not related to the underlying public chains or upper-layer applications at any given time.

While both middleware and DeFi are influenced by feedback mechanisms of token economic models, they differ in their ability to capture value. Middleware projects typically benefit from the supply and demand relationship of their tokens (e.g., network nodes) to deliver the services they provide. In contrast, the demand relationship for tokens in most DeFi projects is less clear, and regulatory concerns about cash flow projects further complicate the situation.

For these reasons, we continue to seek new middleware solutions that will enable the next generation of applications to adopt crypto technology on a large scale. We believe that the next generation of applications will split financial and online business activities. The a16z way of saying it is: we want native applications, not reified applications.