Polygon: Web3's MongoDB and More

I first heard of Polygon in mid-2021. Friends of mine who were web3 entrepreneurs and NFT artists mentioned it as a promising project, providing a solid scaling solution for Ethereum. After some initial research, I learned that the founding team members were mostly from India, which immediately piqued my interest. I’ve been watching and researching Polygon ever since.

Last week, Stripe made a big announcement by adding crypto support in USDC (a stablecoin pegged to USD) for its global payout product, partnering initially with Twitter and deploying on, you guessed it, the Polygon network! This is not only a major validation for polygon, but also a testament to the potential and progress of the web3 ecosystem in India. As I expounded in my first post of 2022, in an age of technology abundance enabled by open source and web3 innovation, developer-turn-entrepreneurs can (and do) come from anywhere.

There is no shortage of content on the Internet about Polygon’s token price, fundraising, transaction volume, and traction in the developer community (more than 10,000 decentralized apps are being built on Polygon to date). However, the project warrants a comprehensive profile of its origin story, core technology, and potential as an investment opportunity.

That's what this post will attempt to do.

Origin Story

The founding story of Polygon began in a co-working space in Mumbai, India. It was pure serendipity and “love at first sight”.

Jaynti Kanani, a scrappy developer and early Bitcoin adopter, was hacking on an Ethereum scaling solution in 2017. Anurag Arjun, also a Bitcoin enthusiast, happened to be sitting right across from him at the same co-working space. Jaynti started talking to Anurag about what he was building (called Matic at the time) and asked Anurag to join him as a co-founder. Anurag said yes.

Soon after that, Jaynti met Sandeep Nailwal via a local crypto trading community. (Back then, most of the crypto communities in India were all about trading and making money, not building technologies or products.) Jaynti and Sandeep hit it off, and a 3rd co-founder of Polygon was “born”. Jaynti fondly recounted Polygon’s serendipitous founding story in this interview:

Mihailo Bjelic, the 4th co-founder and a Serbian entrepreneur, did not join Polygon until a couple of years later. Mihailo was working on a similar Ethereum scaling solution. Instead of continuing solo, he joined forces with the trio from India. Jaynti, Sandeep, and Mihailo all do a good amount of interviews and public presentations. Anurag is the most private one of all the co-founders. Sandeep and Mihailo now often appear together to talk about the latest news from Polygon or the ever-changing web3 space on crypto podcasts like Bankless. They are also the two co-founders most quoted in the media.

The co-founding team used to have C-suite titles like a traditional company: Jaynti was the CEO, Sandeep was the COO, etc. In late 2021, the team decided to do away with all these official titles, because the project is fully decentralized and having such C-suite titles would be anathema to that decentralized ethos. In other words, the co-founders want to “walk the walk’ as much as possible when it comes to decentralization.

Polygon counts famed investor Mark Cuban as an early backer. In February 2022, it announced a massive $450 million fundraising round, done via a private token sale. The round was led by Sequoia Capital’s India arm, with participation from the who’s who of crypto venture capital – Softbank, Tiger, Accel, Union Square Ventures, Digital Currency Group, Alameda Research, Dragonfly, etc. According to Sandeep, this fundraise should give Polygon another “three to five years of runway”.

Technology (By Analogy)

I like to understand technology by analogy, so I’ll apply that lens to explain Polygon without web3 jargon (e.g. L1, L2, etc.). There is a great analogous technology that is older, better understood, and applicable to almost every blockchain technology: databases.

At its core, a public blockchain is a distributed database that is not controlled by any single entity, thus decentralized. (A private blockchain, of course, is controlled by a private entity, and thus not decentralized.)

Ethereum is a decentralized, distributed database. Polygon, as an Ethereum scaling solution, is also a decentralized, distributed database. To achieve more scale with lower cost on top of Ethereum, Polygon makes certain technical tradeoffs. What kind of tradeoffs? A classic distributed database tradeoff, which Jaynti revealed in this interview.

In the interview, Jaynti was asked about what tradeoffs Polygon made between Consistency and Availability, according to the CAP Theorem. His answer was:

“Polygon chooses availability and gets eventual consistency in the current PoS network same as Ethereum. We made a design choice to replicate the same behavior as Ethereum consensus and reduce our engineering efforts at the beginning. And of course, it’s easy to implement :)” [emphasis mine]

You don’t have to know what the CAP Theorem is (an often-debated theoretical framework in database design) to understand the tradeoffs Polygon made. By prioritizing “availability” and settling for “eventual” consistency, Polygon is optimized to be fast and always on, always accessible, and never down. (To power use cases like payment, gaming, and decentralized finance, downtime is a big no-no.)

What gets traded off is the guarantee of data consistency, thus “eventual consistency”. (Think of “consistency” as accuracy at any given moment.) This is a similar and common design choice made by a generation of popular NoSQL distributed databases, like MongoDB, Cassandra, AWS’s DynamoDB, and Azure’s CosmosDB. (For more on MongoDB, I wrote a deeper profile on the company in December 2020.)

The “easy to implement” (with a smiley face) part of Jaynti’s answer is also worth unpacking. The surest way to keep a distributed database “available” and always-on is to add more replicas (or copies) of the same data. If the machine that stores one data replica goes down, just route the application’s traffic to another replica, and another replica, so the entire system is always “available” for that application. Adding new replicas (or “producing more blocks'' in blockchain parlance) is relatively easy (and fast to do), compared to keeping all the replicas consistent, which is quite hard and gets harder the more replicas you add. Polygon can live with these tradeoffs, because it sits on top of a slower but more consistent and more secure distributed database, called Ethereum.

Polygon's architecture

With this technical understanding of Polygon’s characteristics in mind, Stripe’s choice of Polygon makes total sense.

Stripe has a long history and lots of experience using this type of distributed databases, specifically MongoDB. According to a case study published back in 2012, one of the reasons why Stripe engineers liked MongoDB was because getting new replicas set up and running was “incredibly easy”. Even today, Stripe’s job posting explicitly looks for engineers with experience in MongoDB, because its real-time database infrastructure is based on MongoDB – a massive distributed database system that has “strict requirements for security, durability, availability, latency, and scalability.” Sounds familiar?

Polygon is web3’s MongoDB, and more.

Web3’s MongoDB (and More)

Most investors don’t have a good framework to compare and analogize crypto projects to the more traditional investment universe. Bitcoin to gold is the closest thing we have. But if you grasp the technical tradeoffs made by different crypto projects, it is not hard to find analogous products in the more mature (and still growing) database industry.

If Polygon is web3’s MongoDB at the minimum, we now have an adequate comparison to assess Polygon’s future potential as an investment.

At MongoDB’s all-time high, achieved in November 2021, it reached a market cap of $42 billion. At Polygon’s all-time high, achieved on Christmas Day 2021, it reached a market cap of $20 billion. At the moment of this writing (April 23, 2022, Saturday afternoon Pacific Time), MongoDB has a market cap of $27 billion, while Polygon is trading somewhere between $10.5 to $11 billion (crypto never sleeps). To put this comparison in perspective, MongoDB was founded in 2007; Polygon was founded 10 years later in 2017.

Conservatively, Polygon’s market cap can grow 2x to 3x, just to match MongoDB’s market cap.

However, Polygon’s ambition appears much larger than MongoDB’s. It is betting aggressively on zero-knowledge (zk) proofs to increase its network’s scalability by committing $1 billion worth of Matic tokens to partnering or acquiring solutions in the zk space. It has already spent $250 million to buy Hermez and $400 million to acquire Mir. By comparison MongoDB’s three acquisitions to date, totaling a little over $100 million, look like pittance.

Polygon might just be the faster horse, and MongoDB is already pretty fast.

Just like how Polygon used to have a different name, Matic, MongoDB also used to have a different name, 10gen. When 10gen pivoted to focus on its distributed database product, the team chose the word “mongo”, which is a slang that means “huge” or “extremely” – perhaps to describe the amount of data such a database can scale to, perhaps to project the ambition of the company, or both.

If web3 is as big and as early as many developers and entrepreneurs believe, Polygon could be bigger than “huge”, bigger than MongoDB.

Disclosure: I hold some $Matic tokens.

我第一次听说Polygon是在2021年中,当时一些朋友是web3创业者和NFT艺术家,他们提到Polygon是一个很有前途的项目,为以太坊提供一个扎实的扩展解决方案。经过一些初步调研,我了解到Polygon创始团队的成员大多来自印度,这立即引起了我的兴趣。从那时起,我就一直在关注和研究Polygon。

上周,Stripe发布了一条重磅消息:它决定为其全球支付产品增加对USDC(一种与美元挂钩的稳定币)的加密货币支持,初步与Twitter合作,并部署在......您猜对了,Polygon网络上!这不仅是对Polygon的重大认可,还证明了印度web3生态系统的潜力和进步。正如我在2022年第一篇文章中所阐述的那样,在一个由开源和web3创新促成的科技产品丰富的时代,从开发者转成创业者的过程可以(且确实)来自世界的各个角落。

互联网上不乏关于Polygon的代币价格、筹资、交易量和在开发者群体里的使用程度(迄今为止有超过1万个去中心化应用程序正在Polygon上被构建着)的内容。然而,Polygon值得一篇更全面的分析,对其起源故事、核心技术和投资潜力进行全面介绍。

这就是本文打算做的。

起源故事

Polygon的创始故事始于印度孟买的一个共享办公空间。几位创始人的相遇纯属偶然,犹如“一见钟情”。

Jaynti Kanani是一位勤奋的开发者,也是比特币的早期采用者。2017年的时候,他正在做一个以太坊的扩展解决方案。Anurag Arjun,也是个比特币爱好者,碰巧在同一个共享办公空间工作,而且就坐在Jaynti的对面。Jaynti开始和Anurag聊起他正在做的东西(当时还叫Matic),并邀请Anurag作为联合创始人加入他。Anurag答应了。

此后不久,Jaynti通过当地的加密货币交易社群认识了Sandeep Nailwal(当时,印度大多数加密货币社群都是关于交易和赚钱的,没有开发技术或产品的)。Jaynti和Sandeep一拍即合,Polygon的第三个联合创始人就这样“诞生”了。Jaynti在这个采访中深情地讲述了Polygon充满偶遇的创业故事:

第四位联合创始人Mihailo Bjelic是一位塞尔维亚创业者,一两年后才加入Polygon。Mihailo当时正在捣鼓一个类似的以太坊扩展解决方案,但他没有继续单干,而是选择加入来自印度的Matic三人组。Jaynti、Sandeep和Mihailo都做很多采访和演讲,Anurag是所有联合创始人中最神秘的一位。Sandeep和Mihailo现在经常一起出现在像Bankless这样的加密币播客上,谈论Polygon的最新消息或调侃不断变化的web3空间。他们也是最常被媒体引用的两位联合创始人。

Polygon的创始团队曾经像传统公司一样,拥有“首席”的头衔:Jaynti是首席执行官(CEO),Sandeep是首席运营官(COO),等等。但2021年底,他们决定取消这些官方头衔,因为Polygon的项目是完全去中心化的,拥有“首席”这样的名号无疑与项目的去中心化精神形成了矛盾。换句话说,在去中心化的问题上,Polygon的联合创始人们希望尽可能地“身体力行”。

Polygon有像 Mark Cuban这种明星投资人做为其早期支持者。2022年2月,Polygon宣布了一轮4.5亿美元的大规模融资,通过私下代币销售完成。该轮融资由红杉资本的印度分部领头,集齐了加密货币风险投资领域的名人堂——软银、Tiger Global、Accel、合广投资、数字货币集团、Alameda Research、蜻蜓资本等。据Sandeep说,这轮融资应该能给Polygon带来 “三到五年的runway”。

核心技术(由类比来理解)

我喜欢通过类比这个方式来深度看懂一门技术,所以将用这个方式来解释Polygon,而不使用web3的术语(如L1、L2等)。我们其实有门很适合类比的技术,它存在的时间更长,更容易理解,而且适用于几乎所有的区块链技术项目:数据库。

首先,公共区块链在本质上就是一个分布式数据库,不受任何单一实体控制,因此是去中心化的(当然,私人区块链由私人实体控制,也因此不是去中心化的)。

以太坊是一个去中心化的分布式数据库。而Polygon作为以太坊的扩展解决方案,也是一个去中心化的分布式数据库。为了在以太坊之上以更低的成本实现更大的规模,Polygon做了一些技术上的权衡。什么样的权衡?一个典型的分布式数据库的权衡,Jaynti在这个采访中透露了这点。

在采访中,Jaynti被问到,根据CAP定理,Polygon在可用性和一致性之间做了哪些权衡?他的回答如下:

"Polygon选择了可用性,并在目前的PoS网络中获得与以太坊相同的最终一致性。我们在设计上选择了复制与以太坊一样的共识(consensus)模式,这样减少了一开始团队在工程上需要的投入。当然,这也比较容易实现 :)" 【加粗是我加的】

您不必知道CAP定理是什么(数据库设计中一个富有争议的理论框架),就可以理解Polygon所做的权衡。因为优先考虑“可用性”,从而妥协于“最终”一致性,Polygon像优化的是速度、永远在线、永远可访问、永远不停机的高可用性。(停机,对于支付、游戏和去中心化金融这样的用例而言,是无法接受的。)

被权衡掉的是对数据一致性的保证,因此叫“最终一致性”(“一致性”可以被看作是在任何特定时刻数据的准确性)。这与一代非常流行的NoSQL分布式数据库常做的设计选择很类似,如MongoDB、Cassandra、AWS的DynamoDB和Azure的CosmosDB(关于MongoDB的更多信息,请读我在2020年12月写的一篇更深入的介绍)。

Jaynti的答案里,“容易实现”(带笑脸)这个部分也值得解读。保持分布式数据库“可用”和始终在线的最可靠方法,是增加更多相同数据的副本(或拷贝)。如果存储一个数据副本的机器发生了故障,只需将应用程序的流量路由到另一个副本,再另一个副本。这样一来,整个系统对该应用程序来说就总是“可用”的了。增加新的副本(或者用区块链的说法,“生产更多区块”)是相对容易的(而且可以很快实现)。相比之下,保持所有副本的一致性相当困难,而且副本越多,难度越大。Polygon可以作出这些权衡,因为它位于以太坊之上。以太坊虽然速度较慢,但却是个能保证一致性程度更高,且更安全的分布式数据库。

考虑到这些Polygon特性的技术本质,Stripe选择Polygon就非常合理了。

Stripe使用这种类型的分布式数据库,特别是MongoDB,已经有了很长的历史,累计了很多的经验。根据2012年发表的一份案例,Stripe工程师喜欢MongoDB的原因之一是,附加和运行新的副本的体验是 “令人难以置信的简单”。直到今天,Stripe的招聘启事还在明确寻找有MongoDB经验的工程师,因为Stripe的实时数据库基础设施是基于MongoDB的,而MongoDB是个大规模的分布式数据库系统,对“安全性、耐久性、可用性、延迟和可扩展性都有着严格的要求"。听起来是不是很耳熟?

Polygon是web3的MongoDB,还有余。

Web3的MongoDB(且有余)

大多数投资人没有一个好的框架来比较加密货币项目与更传统的投资领域,比特币与黄金是目前最接近的类比了。但是,如果您掌握了不同加密货币项目所做的技术权衡,就不难从更成熟(而且仍在增长)的数据库行业找到类似的产品。

如果Polygon起码是web3的MongoDB,那么我们现在就有一个有说服力的比较对象,来评估Polygon作为投资的未来潜力。

在MongoDB的历史最高点,即2021年11月,它的市值达到了420亿美元;在Polygon的历史最高点,即2021年的圣诞节,它的市值达到了200亿美元。在这篇文章出版的时候(2022年4月23日,太平洋时间周六下午),MongoDB的市值为270亿美元,而Polygon的交易额在105亿到110亿美元之间(加密货币永不眠)。了解下两家公司的年限,您可能会对上述比较更有概念:MongoDB成立于2007年,Polygon成立于10年后的2017年。

保守地说,仅仅只对标上MongoDB的市值,Polygon就有2至3倍的增长空间。

然而,Polygon的野心似乎比MongoDB大得多。它正积极押注于零知识(即zero-knowledge或zk)证明,以提高其网络的可扩展性,并承诺将价值10亿美元的Matic代币用于建立合作关系或收购zk领域的解决方案。Polygon已经花了2.5亿美元收购Hermez,4亿美元收购Mir。相比之下,MongoDB到目前为止的三次收购总额仅略高于1亿美元,显得有些微不足道了。

MongoDB已经是匹相当快的马了,但Polygon可能比MongoDB还快。

就像Polygon曾经有个不同的名字,叫Matic,MongoDB也曾经有个不同的名字,叫10gen。当10gen转向专注于其分布式数据库产品时,团队选择了 "mongo "这个词——是个俚语,意思是“巨大的”或“极度的”——或许是为了描述这样一个数据库所能扩展融入的数据量,或许是为了投射公司的雄心壮志,又或许两者皆有。

如果web3的未来潜力真的像许多开发者和创业者认为的那么大,发展周期还那么早的话,Polygon可能会比“巨大”还要大,妥妥的是web3的MongoDB还有余了。

备注:我持有一些$Matic代币。

如果您喜欢所读的内容,请用email订阅加入“互联”。要想读以前的文章,请查阅《互联档案》。每周一篇新文章送达您的邮箱。请在TwitterLinkedIn、Clubhouse(@kevinsxu)上给个follow,和我交流互动!