EU Orders Meta To Open WhatsApp To Rival AI Chatbots 🗣️ Meta just got told to unbolt a walled garden, and that could turn WhatsApp into the plumbing for third party AI, but can regulators force openness without... #OpenMessaging #EU #Orders 🔽 meta.slashdot.org/story/26/0…

Benchmarking Kafka: Distributed Workers and Workload topology in OpenMessaging Benchmark #apachekafka jeqo.dev/blog/benchmarking-a…

We've rigorously tested Bufstream with the OpenMessaging Benchmark Framework. Here's the results: Bufstream is a next-gen data streaming solution that's 10x cheaper to deploy than Apache Kafka, with 100% hashtag#Kafka protocol compatibility, including exactly-once semantics (EOS) and transaction support. We tested Bufstream against a standard workload: 1 topic, 288 partitions, 1 GiB/s symmetric reads and writes. Using the OpenMessaging Benchmark Framework, we configured 64 producers and sent 1 GiB/s of uncompressed data to 64 consumers. The setup used 6 Bufstream agents on m6in.xlarge instances across 3 availability zones, with S3 as primary storage. It easily handled 1 GiB/s uncompressed writes (256 MiB/s compressed). Bufstream brokers used m6in.xlarge instances (4 vCPUs and 16 GiB) and only needed <1 of 4 vCPUs and <75% memory. Brokers coordinate using a 3-node etcd cluster on m6in.large instances (2 vCPUs and 8 GiB of memory each) managed metadata using minimal resources. Median latency was 260ms, and p99 was 500ms - absolutely acceptable for high-volume analytics driving most Kafka deployments. Assuming 7-day retention, this Bufstream cluster costs $11,147/month in AWS us-east, including all infrastructure and Buf's usage fee. Compute: $1,112/month Storage: $4,625/month (153 TiB written to S3) Networking: $226/month Buf usage fee: $5,184/month An equivalent Kafka cluster would cost $116,460/month, plus additional vendor licensing fees: Compute: $8,039/month (57 r4.xlarge instances) Storage: $73,689/month (906 TiB of EBS volumes) Networking: $34,732/month (2.67 GiB/s inter-zone traffic) Bufstream delivers dramatic cost savings by leveraging cloud object stores while supporting the full Kafka protocol. For most workloads, the minimal latency increase is well worth the 10x cost reduction. #DataEngineering #datastreaming #dataops #ai #FutureOfEngineering #TechTrends #Engineering #ApacheKafka #DevOps

Xiaohongshu is one of the most successful social media and e-commerce platforms in China and is sometimes referred to as "Chinese Instagram". The current Xiaohongshu message engine team is deeply collaborating with AutoMQ Team to promote community building and explore cutting-edge cloud-native messaging engine technologies. This article independently written by Xiaohongshu's engineer provides a comprehensive evaluation of AutoMQ based on the OpenMessaging framework. Meanwhile, AutoMQ welcome everyone to join the community and share their evaluation experiences. To know more details, click the link below👇 automq.com/blog/automq-vs-ka… #Xiaohongshu #AutoMQ #CloudNative #OpenMessaging #MessageQueue #StreamingSystems #MessagingEngine

Please Follow @Foot_Goddess_22 🤍Such An Exquisite Close-Up!🤍 #FootQueen #DM #OpenMessaging

Unfortunately not an accurate comparison with @nats_io. It used the OpenMessaging benchmark which is geared towards the Kafka-like partition model. We recently published a TCO/benchmark relative to Kafka: synadia.com/blog/nats-io-tot…

Great to see OpenMessaging Benchmark Framework being used in almost all performance comparisons in event streaming space. We have been using it for a while to verify Azure Event Hubs performance as well. techcommunity.microsoft.com/…

The number one question #developers ask us: How does #Redpanda stack up against #Kafka? So we used @LinuxFoundation’s #OpenMessaging Benchmark to bring you the answer. redpanda.com/blog/redpanda-v…

Very nice~ Congrats from the OpenMessaging DLedger Community :-)

Are there any plans to publish or push the fixes mentioned in the blog post back to the upstream openmessaging project? I find the tool really useful and having the fixes there would help others benchmarking Kafka as well.

"Benchmarking Apache #Kafka, Apache #Pulsar, and #RabbitMQ: Which is the #Fastest?" Benchmarks are always hard and typically opinionated. Glad to see that the #OpenMessaging #Benchmark Framework was used here. Everybody can review…lnkd.in/dZusJXa lnkd.in/dwH7vWG

I love how @confluentinc enhanced the OpenMessaging Benchmark test drivers for Pulsar & RabbitMQ in addition to fixing laughable flaws in the #Kafka driver. Good read with with details of the testing. rite.ly/wO8f #eventstreaming #nosql

Apache Kafka can do more than 80k msg/s, last time I played with the OpenMessaging Benchmark I pushed 4 million msg/s down 10 partitions (with client side compression). Even RabbitMQ can manage 80k msg/s.

In a recent announcement by industry analyst firm Gigaom, the OpenMessaging benchmark performance results showed that @apache_pulsar delivers consistently superior throughput and latency than @apachekafka even if at an increasing scale. medium.com/@manrai.tarun/apa…

The under-performing OpenMessaging benchmark for Apache Kafka is fixed by simply upgrading the kafka-clients package from 1.0.0 to the latest (2.4.0). Now we're back to in the region of 1 million msg/s on all tests, including 100 partition topics.

Anyone got experience with benchmarking Apache Kafka with 100 partitions on a single topic? With OpenMessaging benchmark I'm seeing the sweet spot around 20 partitions (1.3M msg/s), at 50 it starts dropping and by 100 throughput has crashed (10k msg/s).

Running Apache Kafka OpenMessaging benchmark first. Step 1: Terraform success. Step 2: Ansible fail. When will Ansible allow to abort on "skipping: no hosts matched"?

Congrats OpenMessaging, awards from CJK OSS and @oschina

An interesting look at latency differences between #ApacheKafka and #ApachePulsar as tested using the OpenMessaging benchmark toolkit: buff.ly/2pqlBYH

There is a similar discussion in the OpenMessaging community: github.com/openmessaging/ope…