Oh thank you for sharing! I had tried describing this earlier in my career and had call it "The complexity shell game" as in attempts to hide complexity only serve to move it around and to trick yourself it's gone. I would much rather use an existing standardized term, so I'm switching to Waterbed Theory, thank you!
I like minimal irreducible inherent complexity of a problem or a system (waterbed or zero-point would work I guess) and then you can derive some form of a law of conservation of complexity… if you can measure it. You can definitely feel it, though.
edit: see also ‘No Silver Bullet’ and its essential vs accidental complexity; but don’t disregard Kolmogorov, either, even if it isn’t strictly engineering.
You're getting downvoted but I had the same thought. This sounds like engineering. I do believe there are some cases where pushing the problem to a different place is the right thing to do, sometimes that different place is contextually better suited to that task, so you do actually gain performance or capability from moving the problem around.
When I saw the title the first thing I thought of was schema on read versus schema on right when in data platforms.
You can make writes faster and part of that is by not dealing with schema resolution but you do push the work somewhere else there too.
I guess the same principles apply on many different levels, from when you write to the file system up to how you deal with conflicted data types during data ingestion.
2 angles I think DB designers don’t often think about:
1. Durability extends to the client. Replicated db might ack a write to client, but what if that ack gets lost on the way back over network? If client talks to the DB over simple HTTP, the write might first look like a failure. Can the client retry?
2. Human perception times are biological and don’t change much. But everything in the tech stack has gotten so so much faster since the 80s. Throughput matters, sure, but latency (relatively speaking), is much less of a constraint now than it was.
2) depends, i have a database that farks up a pretty complex distributed system when clients write from another az, latency really can be an issue for some workloads
Another fun failure mode in the same vein: if you are emulating an NVMe device and declare that a write operation is successful once bytes are in memory, you will quickly find yourself buffering arbitrarily large amounts of data as "persist to disk" (SSD speed) falls behind "acknowledge writes" (RAM speed). If you do not add backpressure to your system intentionally, it will be added for you – and you may not like where it's placed!
(This also happens at the SSD level: burst writes can be very fast as data is buffered in the SSD's own RAM, then performance steady-states at the true write speed once that's saturated)
Today I learned there’s a name for the general version of this idea.
https://en.wikipedia.org/wiki/Waterbed_theory
At a certain point in a solution everything you do to optimize (“push”) in one area will cause a negative effect in a different area (“bulge”).
But this is a nice concrete example.
Oh thank you for sharing! I had tried describing this earlier in my career and had call it "The complexity shell game" as in attempts to hide complexity only serve to move it around and to trick yourself it's gone. I would much rather use an existing standardized term, so I'm switching to Waterbed Theory, thank you!
Its amazing how humans come up with the same things independently.
I call it the conservation of effort. If you are able to save mental energy in one place, it necessarily has been spent in another.
My friend calls it the lump under the carpet. You can move it around, but you can never get rid of it.
I like minimal irreducible inherent complexity of a problem or a system (waterbed or zero-point would work I guess) and then you can derive some form of a law of conservation of complexity… if you can measure it. You can definitely feel it, though.
edit: see also ‘No Silver Bullet’ and its essential vs accidental complexity; but don’t disregard Kolmogorov, either, even if it isn’t strictly engineering.
I had heard of it in reference to auto sales being a balloon. If they show you a squeezed price, fees bulge on the other end of the deal.
It’s called constraints and tradeoffs yeah
Sure, but I like how this feels more tactile and how things may bulge in unexpected way.
Like pushing clay around into the right shape of the problem while it’s fighting you.
Constraints and tradeoff feels like the simplified textbook model.
You're getting downvoted but I had the same thought. This sounds like engineering. I do believe there are some cases where pushing the problem to a different place is the right thing to do, sometimes that different place is contextually better suited to that task, so you do actually gain performance or capability from moving the problem around.
v interesting. Thanks for sharing. I kept seeing this phenomenon of "no free lunches". TIL about the waterbed theory.
Something that comes to mind right away is: the cloud will eliminate the complexity of managing OSes and hardware.
Instead it moved it to the complexity of devops and complicated cloud APIs.
When I saw the title the first thing I thought of was schema on read versus schema on right when in data platforms.
You can make writes faster and part of that is by not dealing with schema resolution but you do push the work somewhere else there too.
I guess the same principles apply on many different levels, from when you write to the file system up to how you deal with conflicted data types during data ingestion.
The first thing I thought when seeing the title was "writing" with LLMs - writing quickly can easily just move the work onto your readers!
I like this a lot, it’s true for more than DBs!
2 angles I think DB designers don’t often think about:
1. Durability extends to the client. Replicated db might ack a write to client, but what if that ack gets lost on the way back over network? If client talks to the DB over simple HTTP, the write might first look like a failure. Can the client retry?
2. Human perception times are biological and don’t change much. But everything in the tech stack has gotten so so much faster since the 80s. Throughput matters, sure, but latency (relatively speaking), is much less of a constraint now than it was.
1) you ack the ack, then ack the ack ack ;)
2) depends, i have a database that farks up a pretty complex distributed system when clients write from another az, latency really can be an issue for some workloads
Another fun failure mode in the same vein: if you are emulating an NVMe device and declare that a write operation is successful once bytes are in memory, you will quickly find yourself buffering arbitrarily large amounts of data as "persist to disk" (SSD speed) falls behind "acknowledge writes" (RAM speed). If you do not add backpressure to your system intentionally, it will be added for you – and you may not like where it's placed!
(This also happens at the SSD level: burst writes can be very fast as data is buffered in the SSD's own RAM, then performance steady-states at the true write speed once that's saturated)
if you use the linux kernel that amount of ram is tunable, and write will cause backpressure on its own when this is exceeded.