Cockroach Labs’ Post

Only clean reads with CRDB!

What makes Apple, Amazon, and other ecommerce giants capable of processing dozens of orders per second? They do this by processing checkouts asynchronously and delegating order validation to a delayed job. It's an "optimistic" approach, which assumes that that the payment method is valid and stock is available when the customer places the order, increasing checkout throughput dramatically. This is the topic of my article this week. Thanks for reading, and be sure to subscribe to my newsletter if you haven't already! https://lnkd.in/dKNTnqnx

When ordering online, 30% of consumers see same-day delivery as table stakes. But what does it take to meet that demand? To ensure reliability, delivery companies leverage data infrastructures that seamlessly connect disparate systems like customer orders, vendor stock levels, and last-mile delivery platforms. Read our newest blog to get an exclusive look into one such infrastructure powered by #datastreaming! ⬇️

"Same day delivery ?? of course...needless to say..." - this is how we are these days...aren't we ? But what does it take to bring home this valuable service that impacts the quality of our daily life - here is our story and architecture of how Confluent (FYI - Rob Taylor, Paul Mac Farland, DEREK RUSSELL, Jeremy Hogan, Josh Treichel Paul Earsy Greg Murphy TJ Laher Rodrigo Sanchez Bredee Romano Ninassi Jeff Nalty Trent Burns ) brings together all the critical moving parts - Customer orders, Vendor supply chain, Last mile delivery to make this happen - and I am thrilled to also point out that this customer architecture includes our ISV partners #mongodb Michael C., Dylan Sarachek Dana Groce Jade Furubayashi Xander Uyleman and #databricks - Ariel Amster Kevin Clugage Riley Maris Roger Murff Rob Saker - a clear example of how a #datastreaming Platform for Enterprises acts as a central nervous system bringing together best of breed services in real time to create delight and value for our customers and their customers (i.e. all of us)

There is a devil in those details: making a DRAID stripe width which doesn’t evenly divide into the number of disks you have will probably produce a performance penalty. But when you’ve got 60 disks or more, that sort of performance issue tends to melt away: your bottleneck on a sixty-drive system tends to be the drive controllers themselves, not the individual drives. https://lnkd.in/dkPECa4B

🎄 Supercharge your data analytics skills this holiday season! 🚀 My book, "SQL for Data Analytics Third Edition" is now available at a festive sale discount in the Amazon Christmas Sale, ending on December 22nd. 📘 Seize the opportunity during this festive season – exclusively available for print book on Amazon.com! 🌐 [Book Link: https://packt.link/axNBz] #SQL #DataAnalytics #ChristmasSale #TechReads 📊🎁

In a recent discussion with a customer we got talking about the best metrics for measuring large scale distributed systems where service availability and latency is ultra sensitive. While most of us in tech are used to measures like availability, uptime, throughout and response times, choosing the right metrics for different parts of your system and ensuring you don’t miss short, but impactful spikes isn’t always obvious. Here is a good blog from Sentry (sentry.io) which helps explains different metrics, where they are best applied and the pros and cons 👀👇 https://lnkd.in/g6QqAQj3

As always Mimoune Djouallah coming in with the Facts (pun intended). I won't steal his thunder, if you're interested in #MicrosoftFabric read this blog post.

Umbra: A Disk-Based System with In-Memory Performance "On the other hand, SSDs have achieved astonishing improvements over the past years. A modern 2 TB M.2 SSD can read with about 3.5 GB/s, while costing only $500. In comparison, 2 TB of server DRAM costs about $20 000, i.e. a factor of 40 more. By placing multiple SSDs into one machine we can get excellent read bandwidths at a fraction of the cost of a pure DRAM solution." https://lnkd.in/guCEFY8r

Parallel plans have been a subject of debate amongst ClickHouse enthusiasts, and it's essential to understand both sides of the argument. While parallel plans offer potential benefits such as faster query execution and increased throughput, they also come with certain drawbacks. These include additional CPU and memory usage, increased complexity in query optimization, and potential contention issues. To dive deeper into this topic, I highly recommend checking out the insightful article by ChistaDATA, which presents a thorough exploration of the impact of parallel plans on ClickHouse performance. You can read it here: https://lnkd.in/grJc2QRZ 📚 #ClickHouse #performanceoptimization #databaseoptimization #queryexecution #bigdata #datamanagement