https://github.com/DiceDB/dice/blob/0e241a9ca253f17b4d364cdf... defines func ExpandID, which reads from cycleMap without locking the package-global mutex; and func NextID, which writes to cycleMap under a lock of the package-global mutex. So writes are synchronized, but only between each other, and not with reads, so concurrent calls to ExpandID and NextID would race.
This is all fine as a hobby project or whatever, but very far from any kind of production-capable system.
This PR attempted to fix the memory model violation I mentioned in the parent comment, but also added in an extra change that swapped the sync.Mutex to a sync.RWMutex. The PR description claimed 2 benefits: "Eliminates the data race, ensuring thread safety" -- correct! at least to some level; but also "Improves performance by allowing concurrent ExpandID calls, which is likely a common operation" -- which is totally unsubstantiated, and very likely false, as RWMutex is only faster than a regular Mutex under very narrowly-defined load patterns.
In any case, the PR had no kind of test or benchmark to validate either of these claims, so not a great start by the author. But then a maintainer chimed in with a comment that expressed concerns about edge-condition performance details, without any kind of data or evidence, and apparently didn't care about (or know about?) the much more important fixes that the PR made re: data races.
> I tried changing this, but I did not see any benefit in benchmark numbers.
No apparent understanding of the bugs in this code, nor how changes may or may not fix those bugs, nor really how performance is defined or can be meaningfully evaluated.
Again, hobby project or whatever, all good. But the authors and maintainers of this project are clearly, demonstrably, in over their heads on this one.
If I'm reading this correctly, they are recommending a lock in this situation. However, they are saying the implementations has two options, either raise an error reporting the race (if the implementation is told to do so), or, because the value being read is not larger than a machine word, reply to the read with a correct value from a previous write. If true then it cannot reply with corrupted data.
> However, they are saying the implementations has two options, either raise an error reporting the race (if the implementation is told to do so), or, because the value being read is not larger than a machine word, reply to the read with a correct value from a previous write.
The spec says
> A read r of a memory location x holding a value that is not larger than a machine word must observe some write w such that r does not happen before w and there is no write w' such that w happens before w' and w' happens before r. That is, each read must observe a value written by a preceding or concurrent write.
These rules apply only if the value isn't larger than a machine word. Otherwise,
> Reads of memory locations larger than a single machine word ... can lead to inconsistent values not corresponding to a single write.
The size of a machine word is different depending on how a program is compiled, so whether or not a value is larger than a machine word isn't know-able by the program itself.
And even if you can assert that your program will only be built where a machine word is always at least of size e.g. uint64, the spec only guarantees that unsynchronized reads of a uint64 will return some previous valid write, it doesn't guarantee anything about which value is returned. So `x=1; x=3; x=2;` concurrently with `print(x); print(x); print(x)` can print `1 1 1` or `3 3 3` or `2 1 1` or `3 2 1` and so on. It won't return a corrupted uint64, but it can return any prior uint64, which is still a data race, and almost certainly useless to the application.
The goalposts have been moved. The claim is that this pattern isn’t suitable for production code. The ground truth is that a compliant Go implementation may elect to: crash; read the first value ever set to the variable for the entire lifetime of the program; or behave completely as you’d expect from a single core interleaved execution order. The first is an opt-in, the latter two are up to the whims of the runtime and an implementation may alternate between them at any point.
Is that the kind of uncertainty you want in your production systems? Or is your only requirement that they don’t serve “corrupt” data?
Yep. And even if you were to lock down the implementation of the compiler, the version of Go you're using, the specific set of hardware and OS that you build on and deploy to, and so on -- that still doesn't indemnify you against arbitrary or unexpected behavior, if your code violates the memory model!
Nothing in Go is thread-safe, unless explicitly documented otherwise. Some examples of explicitly-documented-otherwise stuff are in package sync and package sync/atomic.
cycleMap is definitely not thread-safe. The authors knew this, to some extent, because they synchronized writes via an adjacent mutex. But they didn't synchronize reads thru the same mutex, which is the issue.
Looking at the diceDB code base, I have few questions regarding its design, I'm asking this to understand the project's goals and design rationale. Anyone feel free to help me understand this.
I could be wrong but the primary in-memory storage appears to be a standard Go map with locking. Is this a temporary choice for iterative development, and is there a longer-term plan to adopt a more optimized or custom data structure ?
I find the DiceDB's reactivity mechanism very intriguing, particularly the "re-execution" of the entire watch command (i.e re-running GET.WATCH mykey on key modification), it's an intriguing design choice.
From what I understand is the Eval func executes client side commands this seem to be laying foundation for more complex watch command that can be evaluated before sending notifications to clients.
But I have the following question.
What is the primary motivation behind re-executing the entire command, as opposed to simply notifying clients of a key change (as in Redis Pub/Sub or streams)? Is the intent to simplify client-side logic by handling complex key dependencies on the server?
Given that re-execution seems computationally expensive, especially with multiple watchers or more complex (hypothetical) watch commands, how are potential performance bottlenecks addressed?
How does this "re-execution" approach compare in terms of scalability and consistency to more established methods like server-side logic (e.g., Lua scripts in Redis) or change data capture (CDC) ?
Are there plans to support more complex watch commands beyond GET.WATCH (e.g. JSON.GET.WATCH), and how would re-execution scale in those cases?
I'm curious about the trade-offs considered in choosing this design and how it aligns with the project's overall goals. Any insights into these design decisions would help me understand its use-cases.
I was hoping for a response, but no one bothered. I had noted the following when I made that comment and will just wrap up from my end so this could be used by others for reference later.
I'm skeptical that the re-execution approach can scale for complex queries, the latency and throughput improvements would be offseted by the computational cost and bottlenecks introduced for achieving it via its reactivity mechanism (query subscription), this might not work at scale and serve niche use cases.
There are various ways throughput and latency for kv stores can be improved, so bar is really high here.
The messaging with Dice seems unclear and confusing to describe its purpose/use-cases over alternatives, or how it achieves them, which could just be how it's marketed. But it seems to be a collection of ideas and a WIP project.
I think reducing data fetching complexity and complex key dependencies for end clients could be appealing, and it would be great to have it at the KV store level, but there is no reason this type of reactivity can't be implemented on top of various clients for existing KV stores (like Redis). And basic WATCH with transactions are even offered out of the box in them.
Deno kv seem nice but its vendor locked, also there are many others like dragonfly, valkey etc, redis could still work, even something over sqlite can work, deno has a selfhosted kv on top of sqlite - https://github.com/denoland/denokv
From that and the thread so far it seems, they want to make some super cache by building a realtime multi-threaded kv store, improving latency and reducing its read load via its reactivity mechanism. Solving the problem of cache invalidation.
Not sure how this will be achieved but there is no harm in trying. From what is said and shared, rationale behind this design and its tradeoffs are not clear, code could be fixed/improved but providing clarity on this is essential for adoption.
I've seen this more and more with software landing pages, they are somehow so deep into developing/marketing that they totally forget to say what the thing actually is or does, that's why you show it to family and friends first to get some fresh eyes before publishing the site.
In a similar vein, lots of software is Mac-only, but omits to say this anywehere. You just get to the downloads page and see that there are only mac packages.
How hard is it to add two sentences that says only macOS is supported now and in the near future? I’d rather do that than annoy future potential customers who might have a Mac or plan to get one at some point
Looks like a Redis clone. The benchmarks compare it to Redis.
Description from GitHub:
> DiceDB is an open-source, fast, reactive, in-memory database optimized for modern hardware. Commonly used as a cache, it offers a familiar interface while enabling real-time data updates through query subscriptions. It delivers higher throughput and lower median latencies, making it ideal for modern workloads.
Not 100% a Redis clone, but the API appears to be very similar to Redis of 10 years ago, with some additions that Redis doesn't have. See the list of commands: https://dicedb.io/get-started/installation/
DiceDB is an in-memory database that is also reactive. So, instead of polling the database for changes, the database pushes the resultset if you subscribe to it.
We have a similar set of commands as Redis, but are not Redis-compliant.
Would "key-value" not have a place in the description?
This application may be very capable, but I agree with the person saying that its use-case isn't clear on the home page, you have to go deeper into the docs. "Smarter than a database" also seems kind of debatable.
When I ctrl+F the landing page for key and value, I find nothing. Reading it in full, I also come up empty handed. Which part of the landing page implies it's a key value store?
IMO, replace "More than a Cache.
Smarter than a Database." with an actual description.
The saying is cute but does not really convey information the reader is after. And that spot is where you want people to immediately understand what it is.
Still not clear to me what it is. Only the features it has, without knowing what it is.
Like, imagine a page that only said "SuperTransport -- 0 to 100 in 5 seconds", but it is not clear for the reader if it is a car or a horse or a plane or a parcel service...
... and the reader has to go and guess "hmm, guess due to the acceleration it is probably a car or a motorbike -- wonder of it is for sale or for rent?".
Just put "fast on premise key/value database" in the big font that was there -- if that is what it is. That is purely a guess from me, no idea if that is what it is.
Why are you guys building Yet Another DB ? Not trying to dissuade you, but what are you trying to solve that the plethora of DB's currently in market in the same space have not solved ? This should be highlighted in your landing page and since your primary audience is other dev's ( tough-est crowd to sell ), be very specific on what value your product brings over the other choices.
Even clicking through to the Github, after reading the "What is DiceDB?", I'm still not very clear. It feels more like marketing than information.
"What is DiceDB?
DiceDB is an open-source, fast, reactive, in-memory database optimized for modern hardware. Commonly used as a cache, it offers a familiar interface while enabling real-time data updates through query subscriptions. It delivers higher throughput and lower median latencies, making it ideal for modern workloads."
> If you expose something to enough people you'll get some unreasonable takes and interpretations of it. It's important to ignore them.
Quite literally the main function of dice is to give you random numbers. Looking over the website and readme I could not surmise why they would call it DiceDB except for "it sounds nice", but it's absolutely not unreasonable to look at the name and have a thought "it's probably a joke project about random results".
There are literal mountains of software named for no particular reason (let alone sounding nice), or named by origins no person would ever infer without digging in deeper.
Reasonable people realize this and won't discard a project as a joke because of such a teneous connection, and the fact they've gotten traction is a testament to that.
From the benchmarks on 4vCPU and num_clients=4, the numbers doesn't look much different.
Reactive looks promising, doesn't look much useful in realworld for a cache.
For example, a client subscribes for something and the machines goes down, what happens to reactivity?
UPD Nevermind, I didn't have my eyes open. Sorry for the confusion.
Something I still fail to understand is where you can actually spend 20ms while answering a GET request in a RAM keyvalue storage (unless you implement it in Java).
I never gained much experience with existing opensource implementations, but when I was building proprietary solutions at my previous workplace, the in-memory response time was measured in tens-hundreds of microseconds. The lower bound of latency is mostly defined by syscalls so using io_uring should in theory result in even better timings, even though I never got to try it in production.
If you read from nvme AND also do the erasure-recovery across 6 nodes (lrc-12-2-2) then yes, you got into tens of milliseconds. But seeing these numbers for a single node RAM DB just doesn't make sense and I'm surprised everyone treats them as normal.
Does anyone has experience with low-latency high-throughput opensource keyvalue storages? Any specific implementation to recommend?
I had the same reaction as you. And that's for 4 simultaneous clients, too, for a single client you get 3159 ops/s (from https://dicedb.io/benchmarks/). I'm not too familiar with in-memory databases in general but I would have expected figures in the millions on modern hardware. Makes me feel there's some hidden bottleneck somewhere and the benchmarks are not purely measuring the performance of the software.
In-memory caches (lacking persistence) shouldn't be called a database. It's not totally incorrect, but it's an abuse of terminology. Why is a Python dictionary not an in-memory key-value database?
From what I looked at in the past, they seem better on paper by comparing themselves to a very old version of Redis in a rigged scenario (no clustering or multithreading applied despite Drangonfly getting multithreading enabled), and they are a lot worse in terms of code updates. Maybe that's different today, but I'm more keen on using Valkey.
Does Redis support multithreading? Doesn't it use a single-threaded event loop, while DragonflyDB basic version is with multithreading enabled and shared-nothing architecture.
Also I found this latest comparison between Valkey and DragonflyDB : https://www.dragonflydb.io/blog/dragonfly-vs-valkey-benchmar...
Valkey/Redis support offloading of io processing to special I/O threads.
Their goal is to unload the "main" thread from performing i/o related tasks like socket reading and parsing, so it could only spend its precious time on datastore operations. This creates an asymmetrical architecture with I/O threads scaling to any number of CPUs, but the main thread is the only one that touches the hashtable and its entries. It helps a lot in cases where datastore operations are relatively lightweight, like SET/GET with short string values, but its impact will be insignificant for CPU heavy operations like lua EVALs, sorted sets, lists, MGET/MSET etc.
IO multithreading is still not fully there, there were significant improvements within the first couple of iterations, hopefully, it will improve further. I see that Dragonfly uses iouring, which is not recommended by Google due to security vulnerabilities.
Dragonfly supports both epoll and iouring, and polling engine choice is quite orthogonal to its shared nothing architecture. I do not think that Valkey or Redis will become fully multi-threaded any time soon - as such change will require building something like Dragonfly (or use locks that historically were a big NO for Redis).
Yes, per [1] Google did restrict their use of io_uring on “production servers“, and in Android, ChromeOS etc.
However, within that same post, and what is often missed when that post is quoted, is that Google wrote that they did in fact “consider [io_uring] safe” for use by trusted components:
> For these reasons, we currently consider it safe only for use by trusted components.
A database like TigerBeetle is typically deployed in a trusted environment, and is such a trusted component.
I didn't see it in the docs, but I'd want to know the delivery semantics of the pubsub before using this in production. I assume best effort / at most once? Any retries? In what scenarios will the messages be delivered or fail to be delivered?
Different tool. I metrics I am optimizing for are different hence wrote a separate utility. May not be the most optimized one. But I am usign this to measure all things DiceDB and will be using this to optimize DiceDB further.
What are some example use cases where having the ability for the database to push updates to an application would be helpful (vs. the traditional polling approach)?
One example is when you want to display live data on a website. Could be a dashboard, a chat, or really the whole site. Polling is both slower and more resource hungry.
If it is built into your language/framework, you can completely ignore the problem of updating the client, as it happens automatically.
15655 ops a second with a Hetzner CCX23 machine with 4 vCPU and 16GB RAM is rather slow for an in-memory database I hate to say it. You can't blame that on network latency as for example supermassivedb.com is written in go and achieves magnitudes more, actually x20 and it's persisted.. I must investigate the bottlenecks with Dice.
I love the "Follow on twitter" link with the old logo and everything, they probably used a template that hasn't been updated recently but I'm choosing to believe it's actually a subtle sign of protest or resistance.
Snapshot functionality is WIP, which can be utilised to persist and replay data between reboots.
For now Golang SDK is only one, more SDKs are to be added soon.
Based on this thread, I'm not sure you would want to use this over keyspace notifications, but I will also say that there comes a point in the maturity of a system when keyspace notifications become a complicated, unreliable, resource-heavy nightmare. They work fine is your needs and scale are limited, but it's definitely not what you want if handling lots of frequent chances across craploads of keys, with complicated logic for who needs them and how they get routed to them, and where it matters if the notification is successfully received.
But certainly you could build something to handle these and most other needs in this realm with mostly just redis, using streams for what needs to be more robust, in tandem with pub/sub, keyspace notifs, etc. in the areas they are suited to.
The benchmark tool is different. I mentioned the same on my benchmark page.
We had to write a small benchmark utility (membench) ourselves because the long-term metrics that we are optimizing need to be evaluated in a different way.
Also, the scripts, utilities, and infra configurations are mentioned. Feel free to run it.