[ignoramous]

> Since the state is stored in a multi-level Bloom Filter style data structure, the memory needed is constant and does not scale with the number of clients.

Constant memory, but those hashes will take up CPU cycles. If you're running a workload that completes sub 20 milliseconds, these cycles spent hashing may not be worth it over, say, a constant-time admission control like token bucket.

[mnadkvlb]

Absolutely my thought as well. During my thesis i found token buckets to be better and more fair compared to other algos[1]. This was in libvirt. my finding was that it scaled well up to around 1k buckets if memory serves right. after that the results were weird to say the least. (of course the servers i tested on were quite old with not too much memory). Would be nice to run my thesis tests again to find what happened in the last decade.

I tried writing another algorithm for network splitting but didn't get any better results. [1]: https://www.csg.uzh.ch/publications/details.php?id=1007

[JensRantil]

The world is full of trade-offs. I don't think the author intended to solve CPU-bound workloads like that. Or have they claimed that?

[ignoramous]

Not just CPU-bound, but IO-bound workloads can also complete within milliseconds.

> have they claimed that?

The mention of "token bucket" in the project readme is why I wrote the comment I did.

  ... FAIR [only throttles] when there's a genuine shortage of resources as opposed to the approaches like token bucket or leaky bucket which may reject requests ...

[remram]

I'm not exactly sure what you propose, a token bucket per client? In a hashmap client=>bucket?

[otterley]

That's the typical design for a token-bucket rate limiter. A token-bucket limiter uses more memory but is a simpler design. I believe this bloom-filter based implementation is designed to be more memory efficient at the cost being less CPU efficient.

As usual, tradeoffs are everywhere.

[hinkley]

Bloom filters really shine when they avoid a network round trip or pulling data from disk. Those are so far away you can easily afford to spend 5% of the cost of the request to avoid 99% of the requests.

[otterley]

Agreed! In a typical token bucket implementation, bucket state must be shared across multiple servers, which means you need a shared state repository such as Redis or a database (SQL or noSQL). This can add milliseconds to each request being handled.

[hinkley]

> This can add milliseconds

My previous (very long) project was in such a state when I got there that it was only in the last year I was there that measuring things in microseconds was something I could get on other people's radars. I wish I had started sooner because I found a lot of microseconds in about a four month period. That was the most intensive period of user-visible latency reduction I saw the entire time it was there and second through fourth place took years of manpower to accomplish. Past me and future me are both still mad about that.

[foobazgt]

Rate limiters seem like the poster child for something you'd want to host in an in-memory KVS. Typical access times look more like ~100us, not milliseconds. Even if you have a complicated algorithm, you can still limit it to a single round trip by using a Redis Function or some moral equivalent.

[jrockway]

You can also rate limit on the server side. Service A talks to Service B. Service B is 3 replicas. Each replica keeps track of how many times Service A talked to it. If that goes over the limit / 3, deny the request. Now there is no IPC required.

This isn't as accurate, but it's often adequate. I have never liked making a network request to see if I can make a network request; too slow. (I do use this architecture to rate-limit my own website, mostly because I wanted to play with writing a service to do that. But I'd hesitate to make someone else use that.)

[otterley]

That would work if you can assign requests from a given tenant to a single instance, but there are many situations in which that's either impossible or unwise. What if a single server doesn't have enough capacity to handle all the traffic for a tenant? How do you preserve the state if that instance fails?