Sign in

Reuben Bond

@rbn.bsky.social
3K followers 247 following 116 posts

Distributed Systems & databases person. Works at Microsoft on Orleans & Aspire

PostsRepliesMedia
Reposted by Reuben Bond
Loïc Baumann @nockawa.log2n.io · 20/05/2026
How do you hit 1-2µs transaction commits in .NET? Five principles: control memory layout, kill heap allocs, reduce indirection, help the JIT, design for the cache. Concrete numbers. #dotnet #performance #systems nockawa.github.io/blog/microsecond-…
nockawa.github.io
Microsecond Latency in a Managed Language
Five design principles that let a C#/.NET database engine hit sub-microsecond transaction latency — from cache-line-aware structs to JIT-eliminated dead code.
041
Reuben Bond @rbn.bsky.social · 09/04/2026
Fast CASPaxos - leaderless distributed consensus without logs. It allows any proposer to commit an update to a shared register with a single round-trip while maintaining linearizability. reubenbond.github.io/posts/fast-c...
081
Reuben Bond @rbn.bsky.social · 07/03/2026
Reliable, long-running, interactive agents in the cloud - something I've been working on. You can create durable agent sessions via the SDK, enqueue inputs, stream outputs. Agents grind away until they complete the goals you set. Agents in a workspace share a VFS.
071
Reposted by Reuben Bond
Andy Pavlo @andypavlo.bsky.social · 01/10/2025
Our SIGMOD paper with our friends at Tsinghua + @wesmckinney.com + @pateljm.bsky.social on creating a next generation open-source data file format is out. F3 is a future-proof file format avoids the mistakes of Parquet. 📄 Paper: db.cs.cmu.edu/papers/2025/... 📁 Code: github.com/future-file-...
F3: The Open-Source Data File Format for the Future
SIGMOD 2025
46921
Reposted by Reuben Bond
Chris @chris.blue · 16/09/2025
Why, hello there! 🪟 Thanks @rbn.bsky.social :D
021
Reuben Bond @rbn.bsky.social · 01/09/2025
I have an Aer one. You can get decent Tomtoc ones which mimic the designs of popular brands for less $ - unsure in the quality but I imagine they are decent.
120
Reuben Bond @rbn.bsky.social · 04/08/2025
Async/await was a developer experience regression compared with blocking APIs, but a huge advancement over the APIs it replaced, which were callback based. I'm optimistic about virtual threads as a best of both worlds, if we can retain the benefits of futures & structured concurrency
170
Reuben Bond @rbn.bsky.social · 04/08/2025
Ah, thanks. How much of it is attributable to BEAM itself, I wonder. Using AOT instead of the JIT cuts memory usage in .NET by about half, for ex. IIRC, the numbers below are steady state under load, not startup, but either way
010
Reuben Bond @rbn.bsky.social · 04/08/2025
Thanks. Perhaps there is still low hanging fruit for elixir/phoenix/BEAM when it comes to reducing startup mem consumption. I'm sure it would still consume more than a Rust app given the nature of the runtimes
000
Reuben Bond @rbn.bsky.social · 04/08/2025
What is MB RES?
200
Reuben Bond @rbn.bsky.social · 09/04/2025
Pat Helland delivered a presentation on this paper: hpts.ws/papers/2024/...
1134
Reposted by Reuben Bond
Carlos Baquero @xmal.bsky.social · 07/04/2025
In the last week I probably explained Rateless Set Reconciliation to a dozen other scientists. What an amazing paper and result, and already one year old.
6496
Reuben Bond @rbn.bsky.social · 03/03/2025
It's just a toy / proof-of-concept, really, but it might be useful to refer to.
000
Reuben Bond @rbn.bsky.social · 29/01/2025
Thank you, Jeremy! I gave a presentation about some of the design considerations in the serialization & RPC system: www.youtube.com/live/kgRag4E...
youtube.com
Reuben Bond - Orleans under the hood (Dotnetos Conference 2021)
YouTube video by Dotnetos
140
Reuben Bond @rbn.bsky.social · 04/01/2025
All things agents. I'm interested in builder communities rather than per-product servers
110
Reuben Bond @rbn.bsky.social · 04/01/2025
What's a good discord server for people building things in the AI/LLM space?
270
Reuben Bond @rbn.bsky.social · 03/01/2025
Committing the first value is safe as-is, but by playing more with this I believe subsequent Fast Rounds require additional safety rules: 1. Proposers only propose updates to known-committed values 2. Acceptors require the value's version (akin to slot number) has increased or values are identical
100
Reuben Bond @rbn.bsky.social · 03/01/2025
Did he have his onion juice with him or a couple of pocket eggs? If not, he's ngmi
100
Reuben Bond @rbn.bsky.social · 03/01/2025
gotta max out those T levels somehow
100
Reuben Bond @rbn.bsky.social · 16/12/2024
Feel free to dm
000
Reuben Bond @rbn.bsky.social · 15/12/2024
Is there a doc describing it?
100
Reuben Bond @rbn.bsky.social · 15/12/2024
Link?
200
Reuben Bond @rbn.bsky.social · 15/12/2024
"nothing more" is a bit too far: CRDTs have limitations on their behavior. You can't implement just anything using CRDTs. Their utility is quite limited and hence they are not widely deployed within datacenter based apps (I know of none, and no one has provided examples yet)
000
Reuben Bond @rbn.bsky.social · 15/12/2024
It depends on the CRDT, but that's not my area of expertise. The journaled grains use event sourcing (modified for geo distributed environments), and I believe you could implement CRDTs on top fairly easily, but the API also exposes coordination/sync operations for consistency.
200
Reuben Bond @rbn.bsky.social · 15/12/2024
What I described is not a CRDT. I am talking about ACID database transactions using optimistic concurrency control
100
Reuben Bond @rbn.bsky.social · 15/12/2024
You can version the data and check for conflicts at commit time, similar to a database transaction with optimistic concurrency control.
100
Reuben Bond @rbn.bsky.social · 15/12/2024
In the online case you can coordinate changes. In the offline case you cannot: you have divergent data replicas being updated independently. The changes need to be merged eventually and there is ambiguity as to how.
100
Reuben Bond @rbn.bsky.social · 15/12/2024
That's CRDT territory, but the goalposts just shifted
100
Reposted by Reuben Bond
Heidi Howard @heidihoward.bsky.social · 15/12/2024
Interesting workshop and a lovely community, would very much recommend!
187
Reuben Bond @rbn.bsky.social · 15/12/2024
The sources can change but each one is authoritative (eg for fx rates) so you can assign a sequence number to each update which can be used to ensure consistency
100
Reuben Bond @rbn.bsky.social · 15/12/2024
One way would be to leave the import in a staging area until the user has completed their part of the workflow against a point in time snapshot. There is nothing you can do with a CRDT here which you cannot do without - there's no replication in this scenario
100
Reuben Bond @rbn.bsky.social · 15/12/2024
Why makes you say it's impossible without CRDTs? I don't see why it would be any more complex without CRDTs
100
Reposted by Reuben Bond
Peter Kraft @petereliaskraft.net · 13/12/2024
Hi everyone! I'm a co-founder @dbos.dev, where we're building a serverless platform for highly reliable applications. I love conversations about databases, distributed systems, and anything technical. Thanks @qianli.dev for introducing me to Bluesky, and looking forward to meeting people here! 🦋
0314
Reposted by Reuben Bond
Justin @jaffray.bsky.social · 09/12/2024
NULL BITMAP: in honour of DSQL's release last week I wrote about what the point of having extremely precise trustworthy clocks is buttondown.com/jaffray/arch...
buttondown.com
What are the Magical Clocks for?
At Re:Invent last week, AWS announced DSQL, their new serverless SQL database. As a fan of distributed SQL databases I have been enjoying reading about the...
24311
Reuben Bond @rbn.bsky.social · 09/12/2024
Agreed
000
Reuben Bond @rbn.bsky.social · 09/12/2024
This is MS offering them for others, not using it themselves. It's also not mentioned in the actual Cosmos DB docs, only this blog post - does it exist? eg: learn.microsoft.com/en-us/azure/...
learn.microsoft.com
Conflict resolution types and resolution policies in Azure Cosmos DB
This article describes the conflict categories and conflict resolution policies in Azure Cosmos DB.
100
Reuben Bond @rbn.bsky.social · 09/12/2024
Do you have examples of applications of CRDTs which live solely within a datacenter based application? I.e., the DC part of the app isn't a rendezvous point for clients replicating state via CRDTs: it uses CRDTs to replicate between nodes in the DC. I've never seen it in practice.
000
Reuben Bond @rbn.bsky.social · 08/12/2024
Distributed consensus isn't the only suitable choice for datacenter based apps of course (but it's typically involved). Chain Replication is often a good choice instead, for ex.
000
Reuben Bond @rbn.bsky.social · 08/12/2024
Replication for durability, yeah, I'm talking about apps which live in the datacenter and how CRDTs are almost never chosen for data replication in apps which live in the DC
100
Reuben Bond @rbn.bsky.social · 08/12/2024
The places where distributed consensus might be a good choice and the places where CRDTs are a good choice are almost mutually exclusive - they aren't substitutes for each other. They each satisfy distinct requirements.
100
Reuben Bond @rbn.bsky.social · 08/12/2024
Latency is low in the DC but you need to incur it anyway for most applications because when a user receives an acknowledgement they expect it to be durable. Losing a node with unreplicated data is not acceptable for these applications.
100
Reuben Bond @rbn.bsky.social · 08/12/2024
CRDTs always have usage caveats. You don't have simple abstractions like a general purpose set or a register, it's always a more narrow case like a G-Set or 2P-Set. This puts a burden on app developers.
100
Reuben Bond @rbn.bsky.social · 08/12/2024
I agree with the part about CRDTs being useful for part of the app. I would narrow it to specifically parts of the app which are outside of the DC - even if the DC is used for rendezvous between those edge nodes
100
Reuben Bond @rbn.bsky.social · 08/12/2024
Any more info on that?
100
Reuben Bond @rbn.bsky.social · 08/12/2024
CRDTs make replication so much more complicated in-part because you can no longer separate the replication algorithm from the data structure and application semantics
100
Reuben Bond @rbn.bsky.social · 08/12/2024
Google Docs uses Operational Transformation, not CRDTs, and the OT is for sync between clients, not nodes in the DC - i.e. the use case for OT doesn't overlap with the use case for DC data replication algos
010
Reuben Bond @rbn.bsky.social · 08/12/2024
CRDTs fit a system model that has hardly any overlap with datacenter-based applications. If I'm wrong, please point to datacenter-based apps which have benefited from the application of CRDTs
9133
Reposted by Reuben Bond
Alex Miller @alexmillerdb.bsky.social · 20/11/2024
New blog post on the fun new hardware advancements which databases can leverage for great gains, and why the cloud means it doesn't matter that they exist. 🫠 transactional.blog/b...
35418
Reposted by Reuben Bond
Qian Li @qianli.dev · 14/11/2024
ACM Symposium on Cloud Computing (SoCC) is next week! Lots of great papers on serverless, cloud, and AI/ML systems. I appreciate the push on diversity: both keynote speakers are women, plus a women in systems meetup! (This was my first time as a program committee member🤗) acmsocc.org/2024/index.h...
acmsocc.org
2024 ACM Symposium on Cloud Computing
0121
Reposted by Reuben Bond
ahmet alp balkan @ahmet.dev · 13/11/2024
We've now published a blog about how LinkedIn runs large stateful apps with a generic operator on Kubernetes. www.linkedin.com/blog/enginee... Don’t forget to come to our #KubeCon session on Thursday.
1217