Data Mesh: The Next Stage in the Evolution from Time Share to Data Share

From that pivotal moment more than 50 years ago when Bell Labs first shared a copy of UNIX with UC Berkeley, UNIX became the most influential time-sharing operating system, continuing as the foundation for the modern compute environment.

But times change. Allocating formerly scarce resources like CPU and memory has given way to data becoming the critical resource to share. However, sharing data remains cumbersome, expensive, and merely portable for ML and analytics. There has to be a better way…

Enter Data Mesh. Zhamak first introduced the Data Mesh concept at a 2019 O'Reilly conference as an approach to radically democratizing data access. Both criticized and praised by analysts, Data Mesh now defines data sharing for many organizations.

However, the lack of a universally accessible reference architecture and an open technology built natively for Data Mesh, combined with overly ambitious marketing and messaging, has led to widespread confusion within the industry.

Zhamak aims to demystify Data Mesh from a technical implementation [standpoint | perspective]. Shedding light on the current flavors of Data Mesh found in the wild, exploring their trade-offs and peculiarities and their influence on her work on NextData's "nxd os" Data Mesh operating system.