WASM is all the rage these days, with new companies sprouting up to take advantage of the power of modern personal computers, the ubiquity of web browsers, open source, and much more. Today while recording the It Depends podcast with Sanjeev Mohan and Matt Housley on our 2023 predictions, I blurted that WASM might be key to enabling Data Mesh. Let me explain.
When I speak with Zhamak about her vision for Data Mesh, it comes from her dystopian fear that big companies and governments control access to data (I agree with her). In the end, data needs to be free. Data Mesh is about the decentralized sharing of data and giving people power and control of their data.
When I think back to my youth on the Internet (the 1990s and early 2000s), peer-to-peer (p2p) was how we’d share a lot of information. Unlike today, companies were still trying to figure out the Internet back then. The web browser was new (I actually beta-tested Mosaic as a teenager), and there were a few walled gardens like Compuserve, AOL, and others. What I loved about the Internet then was how truly free it was. In those days, anything was possible, and information flowed freely.
Fast forward to today, and the Internet is mostly a cesspool controlled by large monopolies (I’ve thought about leaving, but that’s nearly impossible to do). The big thing that’s changed is the web browser is the primary interface for almost everything you do on your computer (your mobile device is similar). Computers are insanely powerful these days. My Mac M1 has more power than I’ll need for most pedestrian things I do on the browser. Most people have similar disparities in how they use their powerful machines and how much latent power is sitting idle.
The developments around WASM and databases are interesting. Earlier this year, the folks at Singlestore gave an excellent demo of using WASM and Singlestore at my meetup. A few months later, Jordan Tigani and crew arrived on the scene with Motherduck. You can catch Jordan’s public unveiling of Motherduck during our interview here. There are several other projects where Postgres + WASM, SQLite + WASM, and more.
What does all of this mean? It means you can run a database inside your browser. This opens a whole new world of data applications, such as interactive analytical apps, in-browser OLAP queries (no need to traverse the network anymore), and much more. It also means the local creation and sharing of data products, free from the centralization of the big clouds or other intermediaries. The database is just the start. You can see where this goes. The equivalent of Napster is now possible for data products.
Of course, a lot needs to happen for WASM to enable Data Mesh. Data Mesh itself needs a lot of things to be built, such as interoperability and sharing standards. But Data Mesh is also about avoiding any single monolithic product or vendor, as this defeats the purpose of decentralization. WASM allows a way to sidestep centralization while offering an actual p2p experience for the decentralized sharing of data products.
Why WASM might be key to Data Mesh
I think an 'Intro to Web Assembly (WASM)' is probably needed for a lot of data engineers, DS, analytics engineers and analysts, etc.
(I agree with most of your points, btw, but you know this already.)
I am also personally pretty staunchly of the opinion that data products is the single most obvious use with meaningful adoption for Web Assembly as a commercially viable paradigm. There's a lot of hype and hoopla around this for crypto trading, lots of Rust vs. Zip silliness that may not actually be important.
IMO, this is the single best opportunity for data engineers and data-focused software engineers to "lead from the front" that I've seen in the past few years, and if it's executed well it will help get around the classic "data teams are secondary to software teams" narrative that is too common.
Never once in your article did you expand on WASM.
You assume everyone is intimately familiar with your jargon :)