Amir Moulavi

Nice article touching on streamlined Karaf modular modulith deployments

2

Some people ask me why I am such a stickler and a curmudgeon when it comes to #OpenSource. That is a good and valid question. So let me try to answer it. I've been involved in Open Source for over 30 years. Now the term "involved" is so vague and nebulous as to be almost meaningless in this connotation. So let me summarize my involvement. I contributed back in the 80s and 90s to some foundational software projects of the Internet: sendmail, xntpd, BIND/named, various GNU projects and loads of others. I contributed to the old NCSA webserver and was an extremely early addition to the original Apache Group. I co-founded the ASF and served on its board for its first 20years, where I also served as EVP, Chairman and President. Many if not most of the major cornerstones of the ASF I had a major hand in. I've served on other FOSS foundations, and not because whatever company I worked for handed over major $$ in exchange for a board seat, but because of "true" merit (whatever that means). And I am still as active today as I was back in the day. I am considered by many to be one of the remaining *active* elder statesperson of the Open Source movement, a member of the "old guard" so to speak. I take that responsibility seriously. I say all this not to brag (OK, maybe a little bit. I need the self confidence boost), but to also provide some background. Basically, I was there when this all started. I was not only there, but on the front lines with many, many others. I was there defending Open Source; When there was FUD about Open Source, I was one of the people combating that. Back when Open Source was considered this weird little outlier, the underdog, when it couldn't be trusted, when companies called it a "cancer", when every gain had to be fought for, I and many others had to battle to ensure that Open Source was understood, acknowledged and eventually embraced. And we have the battle scars to prove it. The success of Open Source was hard, hard won. So today, when I see people misrepresenting what Open Source "is", or try to redefine it, or make claims about it that were never part of the motivation or the movement, it reminds me of the old days. When people usurp Open Source to the detriment of the community, in attempts to bask in the reflected glow and good will that Open Source was able to eventually win, those old battle scars start to bleed just a little. The fight isn't over, just because the term Open Source is known and popular. In many ways, as Open Source weaves its way into other arenas and avenues, the fight continues and even expands. Open Source succeeded, and continues to succeed, because Open Source is a known and agreed upon entity. It thrived because of the Open Source Definition, not despite it. The phrase "Open Source" means something, and Open Source only continues because that meaning has, well, meaning. It has depth. It has a history and a legacy. And that is worth keeping up the good fight, even if means distain or mockery.

70

14 Kommentare

Sometimes true power can be unlocked from existing technologies if we intentionally choose to use them "the wrong way". The next post describes such a case with AWS SQS and StepFunction. On a different note, I think I should write more. #aws #serverless #sqs #stepfunction #lambda #cloud #architecture

92

2 Kommentare

Here are the slides for my GopherCon.EU talk: How to Win Frames and Influence Pointers https://lnkd.in/d7byRQfu #gopherconEU #golang

72

4 Kommentare

We chose Redis Streams over Kafka for its simplicity. This is how we use Redis Streams as a queue 👇 https://lnkd.in/dPSQTF-F

1

"#apachekafka isn't a message queue." Well, that's true for now, but not for long! Currently, Kafka is a distributed log, and it does not have Queue semantics. Log semantics include: - Records are produced to a log. - Records can be consumed, and (within a single partition) they are consumed in the order that they are produced. - Records may be read multiple times by multiple consumer groups. - Records are not "removed" from the log immediately after being read. However, in a message queue, things are different: - Messages are produced to a queue. - Consumers can read and acknowledge messages individually - Messages may be delivered out of order. - Once a message is acknowledged, it is removed from the queue. - Messages may be "retried" if not successfully acknowledged. Kafka is one of the fastest-moving open-source projects. It has an ongoing KIP (Kafka Improvement Proposal) to add Queue semantics on top of its rock-solid log abstraction! Check out KIP-932: https://lnkd.in/gb2Qi__X It will introduce the concept of a "Share Group", which is a new type of consumer group that reads records from a topic with queuing semantics. The Share Group will keep track of which records (well, actually "messages" in this context) have been acknowledged and which ones are still in-flight.

183

7 Kommentare

It seems service deprecation is a thing now for AWS. I'll not argue here if that's a good or bad thing 😜. But if you want to keep track, here are two useful resources: 👉 https://lnkd.in/dCHGPiqK A curated list of breaking changes maintained by Scott Piper 👉 https://lnkd.in/d5rxybzi A list of service deprecation announcements (even if the post title is not that obvious) maintained by Luc van Donkersgoed #AWS #AWSCommunity

22

9 Kommentare

Attn: Data pros curious about the power of Apache Flink, this is for you. Confluent is back with an all new comic to explore the core concepts of real-time stream processing and Apache Flink. Join our heroes as they embark on a new adventure into StreamScape to learn about the challenges of streaming processing and how Apache Flink and Apache Kafka are better together. Get your copy now!

5

📣 Big news for developers! Cloud Run now supports GPUs! 📣 You can quickly host LLMs on serverless now, only on Google Cloud. ✅ Low Latency ✅ Cost-Efficient ✅ Easy to Use Learn more: 👉 https://lnkd.in/d4YiYSJJ #CloudRun #Serverless #AI #GPU #GoogleCloud

7

ICYMI, Google Cloud Platform and Technical Infrastructure Director, Jennifer Petoff, gave the opening keynote at DevOpsDays Zurich this year. You can check out the video here if you want to learn more about the importance of education in driving an SRE or DevOps culture.

3

Check out our guide to picking the right API gateway with Moesif! Make sure you find the one that works best for you.

1

Today, Spectro Cloud celebrates Kubernetes’ 10th birthday in style by launching the 2024 State of Production Kubernetes report. This year’s report is the most comprehensive one yet, filled with insights on the hottest topics in our industry, ranging from AI and platform engineering to FinOps and VM modernization. It’s exciting to be part of a company that is shaping the next 10 years of Kubernetes and to have the opportunity to empower our customers’ success every day. I highly recommend getting a copy of the report to learn about the current landscape of production Kubernetes and get a sneak peek at the foreseeable trends. You can download it here: https://lnkd.in/dvH_ny_V #kubernetes #kubernetesmanagement #kubertenes #edgecomputing #ai

2

This is a big deal for using AI in production - Claude 3.5 Sonnet and 3 Haiku are now available in more regions. us-west-2 tends to get new AI stuff first, and it is great to see these 2 models now more geographically available. https://lnkd.in/gvQkMQbX

17

To lift my reply up, I have expanded it here. These topics form much of the material on Day 2 of my Practical Messaging workshop next NDC Oslo 10-11 June) So the first thing to think about is the type of asynchronous conversation: * Messaging (uses a queue; transfer of control) * Eventing (uses a stream; publish-subscribe) (Replies can be either stream or queue, default to queue unless you are broadcasting). We also need to think about Provider over Producer/Consumer when thinking about the interaction. Who provides the API? The Consumer provides the API for messaging; the Producer provides the API for eventing. Concepts like In-Out, Out-Only are useful here from the old WS- Messaging specs. I use that taxonomy when I teach my workshop because it is still a useful tool. For example in In-Out we have a consumer that acts as the API provider and receives a message; it should use a queue; we have a response that probably travels back over that queue. The main issue that brings this up is causality - I need this step (taking a payment) to occur before this step (place an order). Definitely seen a few systems struggle with “event chaining” where they lose causality, particularly pernicious when a component fails (and can lead to nasty loops where work goes round and around - imagine folks keep turning up with the same order to your house). I spend time in my trying to avoid everyone reaching for a Process Manager (Saga) here. In a lot of cases I recommend what I call Turn-Taking which is essentially Pat Helland’s Activity model (components co-operate via their internal business logic).

32

4 Kommentare

🔭 OpenTelemetry Architecture Overview. Part1. SDK Design Just released a new blog post with an overview of OpenTelemetry Python SDK design. I have touched on the following topics: - 🔭 Two words on history and the idea of OpenTelemetry - 🧵 Distributed traces & spans. How span collection happens on the service side - 💼 Baggage & trace ctx propagation - 📈 Metrics collection. Views & aggregations. Metrics readers - 📑 OTEL Logging integration - 🤝 Semantic conventions and why that is important If you want to understand how OTEL SDK works under the curtain, this article is definitely for you 🙌 #opentelemetry #observability #distributedtracing #distributedsystems #sre

25

2 Kommentare

I just updated my first table format comparison blog post to fix some missing details of Apache Hudi. https://lnkd.in/dwSUS6qS

5

Here's an excellent article on why storage performance tests should use 64k block sizes. Many tests use 4k blocks. While this results in higher perceived performance, it's not that relevant in typical modern workloads (like AI). Here's a great article on the subject... https://lnkd.in/eehnD6R9 George Crump Greg Campbell

16