Great article, Johnny, and I love the word schmäh 😁
Totally agree with your overall point. That said: my first data job in 2009 (migrating hospital data between very different OLTP databases, rather than anything analytic) was already using a "multi-hop data management and processing pattern", on SQL Server 2005. And that company had been doing that for some years before I joined. There was nothing lakey or open about it, which is key, but I guess it was basically an ELT medallionesque approach.
Yeah, I feel like I've managed to live under a rock... In my 20-ish year career I'd never experienced the multi-hop approach until the advent of Lakehouses, but having exchanged dialogue with a few folks, perhaps it was more common than I realised. So having gotten all preachy about it, I feel a little tale between my legs now. That being said, I still maintain that it's widespread adoption is still a relatively new thing
For sure, I think the move to standardising on reasonably open/portable data formats, and the whole storage/compute separation, is also a meaningful factor. The combination of what you do, how you do it, and what you do it WITH, all have to come together.
Great article, Johnny, and I love the word schmäh 😁
Totally agree with your overall point. That said: my first data job in 2009 (migrating hospital data between very different OLTP databases, rather than anything analytic) was already using a "multi-hop data management and processing pattern", on SQL Server 2005. And that company had been doing that for some years before I joined. There was nothing lakey or open about it, which is key, but I guess it was basically an ELT medallionesque approach.
Yeah, I feel like I've managed to live under a rock... In my 20-ish year career I'd never experienced the multi-hop approach until the advent of Lakehouses, but having exchanged dialogue with a few folks, perhaps it was more common than I realised. So having gotten all preachy about it, I feel a little tale between my legs now. That being said, I still maintain that it's widespread adoption is still a relatively new thing
And schmäh is my word of the week. I've had need for it in a few different contexts so far (not all particularly positive)
For sure, I think the move to standardising on reasonably open/portable data formats, and the whole storage/compute separation, is also a meaningful factor. The combination of what you do, how you do it, and what you do it WITH, all have to come together.
Interesting.