Click here to sponsor Fudge Sunday and post jobs!
Music: Gary Wright - Dream Weaver (1975)
This week we take a look at Google Service Weaver and the next generation of distributed applications.
Last week we looked at buf. Now, let’s talk a bit about the potential implications of a blog post announcing Google Service Weaver from Srdjan Petrovic, Garv Sawhney, Urs Hölzle, Parveen Patel, Mark Lohmeyer, Amin Vahdat, Michael Whittaker, Robert Grandl et al.
There’s a website: https://serviceweaver.dev
There’s a /docs.html (well done): https://serviceweaver.dev/docs.html
There’s a hello world blog too: https://serviceweaver.dev/blog/quick_intro.html
This will probably catch your eye…
We found out that Service Weaver improves application latency by 15x and reduces VM costs by 9x when compared to a typical microservices solution built on top of gRPC and protocol buffers. 1
So, I don’t have any inside knowledge — but it is safe to say that Service Weaver didn’t just come into existence a few weeks ago. This problem space is very likely an old itch that got scratched within Google and now the “new” backscratcher is being shared with all.
For a great roundup post, check out what Matt Campbell put together over at InfoQ.
Remember that Shakespearian thing about life being a stage and we are but actors? Well, let’s go back to 2009 and learn about “Actors in Erlang”2 from Alex Miller.
And speaking of weavers and dreams… A quarter century has passed since Macromedia Dreamweaver launched3 to capture the hearts, minds, and checkbooks of budding web developers in the late 1990s.
A few years after capturing everything Microsoft had not already captured in the web authoring tools market, Macromedia became part of Adobe in 2005.
What can we draw from the comparison of a “new” framework meant for building and deploying distributed applications to a 90s era web focused integrated development environment (IDE)? Well, I don’t know, yet — not for sure by any means.
The Service Weaver potential for “a better way” will be interesting to watch over the next three (3) years or so. In just a few months, Service Weaver has gathered impressive GitHub vanity metrics (as of 12-Mar-2023):
For comparison, buf already has impressive GitHub vanity metrics (as of 12-Mar-2023):
Graphs like these can be easily conflated with patterns for success in tooling. In reality, these graphs are looking backwards at what might be a trailing indicator of success in tooling.
Ultimately, as Lee Ditiangkin might argue, growth in adoption comes from making a strong business case that happens to include tooling which improves key metrics used for measuring developer experience.
So, what will be the next big thing in distributed applications?
Until then… Place your bets!
P.S. Special thanks to Neeloy Bhattacharyya for the feedback last week as inspiration to create this issue of Fudge Sunday!
As a reminder, after a +25 year walkabout, I’m an IBMer (again). For 2023, in “Work Plug”, I’ll share a new link each week that is educational, accessible, and relevant to platform engineering from fellow IBMers4 in the wider IBM Community.
More Links of the Week:
I am linking to my disclosure.