Crypto space cannot see widespread adoption with smart contracts and dApps.
Decentralized web must support real world data-rich and compute-heavy apps sooner or later. Web3 cannot be the foundation for the decentralized web because of astronomical developer cost. STORE introduces CPT — Cost Per One Thousand Transactions — a measure that compares developer/user cost to send transactions across multiple decentralized networks.
1/ Here's an interesting discussion on the current state of affairs with crypto today -- https://t.co/MHfuoyjwky
2/ It almost looks like crypto is looking for *problems* to fit its *solutions*. Granted, smart contracts and dApps have their place, but pretending that they will be the foundation for the next Internet (web3) is unrealistic.
4/ Web became what it is today because of the *freedom* developers enjoyed in building their apps and services. This resulted in a rich set of tools and frameworks to improve their productivity and the quality of the software they produced.
5/ With little plumbing, different services could interoperate and in the process, it resulted in cheaper deployment options for developers and lowered the cost of service for enterprises and end users.
6/ And, this progress happened in an *evolutionary* manner with unlimited choices for developers on how they could deploy their apps. They could use *any* language, *any* design pattern, and *any* deployment architecture (on-prem, public cloud, hybrid cloud, serverless, etc.)
7/ dApp platforms changed that in a big way. App developers must now rethink their business logic in the form of smart contracts and are constrained by platform-specific languages to develop their apps.
8/ And, they lose all the freedom to optimize how their apps are run and how data is stored. This is a big price to pay for "decentralization premium". In addition, developers or users of their apps must buy or reserve *network resources* to be able to run these dApp instances.
9/ The cost of these network resources is astronomical. We @storecoin set out to compute what these costs are going to be on various platforms. We wanted to compare these costs against similar apps developed on a centralized cloud platform like AWS.
11/ Economic feasibility is critical for developer revenue, which in turn influences their adoption. Our work led to comparing the developer cost using a metric called, "Cost per Thousand" transactions or CPT.
13/ It is impossible to make an apple-to-apple comparison because of the nuances and differences between various projects, so we must focus on the "trend" this analysis shows rather than specific CPT values.
14/ We can observe that dApp platforms are cost prohibitive to deploy even simple apps requiring a storage of ~512 bytes per app instance and taking ~20ms to execute.
15/ So, these platforms cannot practically support data-rich and compute-heavy apps. Such apps are already the norm in the AI/Machine Learning era we live now.
16/ So, any optimization to run such apps must come from within these platforms, which is why the comparison to AOL makes sense -- https://t.co/uJtlFBhDqh.
17/ In other words, developers are at platforms' mercy, just like they were at AOL's mercy in the early days of the web. So, for crypto to garner widespread adoption, we must think differently on how apps will be run on decentralized web.
18/ Because, sooner or later, we will have real world apps running on these platforms and the existing mindset on how decentralization should work will not be able to meet that requirement.