I also don't mean to put down the creators of this product, it probably has its use cases. But every time I see something like "SQL with the scalability of KV" (which happens a lot these days), it's like someone promising a perpetual energy source. In other words, there is usually some kind of catch involved.
In this case it's that JOINs are still expensive. Well, this is actually a pretty big deal, because JOINs are typically the reason I settle for either a NoSQL or a SQL store.
Again, I'm sure this has other merits that may distinguish itself from the rest of everyone and their grandma building a database these days.
Only JOINs across actors are sacrificed. Most problem sets fit quite well into the actor model and you can keep JOINs where it matters most (inside an actor).
In this case it's that JOINs are still expensive. Well, this is actually a pretty big deal, because JOINs are typically the reason I settle for either a NoSQL or a SQL store.
Again, I'm sure this has other merits that may distinguish itself from the rest of everyone and their grandma building a database these days.