If you see this post today, you may wonder why I have published on a Saturday instead of, as usual, on a Sunday. Well, I am flying out to Barcelona a bit later today and will be on a plane for most of tomorrow.
This roundup looks at cloud OLTP databases and whether we can get them to be really scalable. There is a link to a post looking at Azure Data Explorer* together with Azure Serverless functions - very cool! An interesting post about ML in online gaming, plus a lot of other cool stuff.
Databases
- Is Scalable OLTP in the Cloud a Solved Problem? (CIDR 2023). In this post, Murat dissects the Is Scalable OLTP in the Cloud a Solved Problem? whitepaper. The paper discusses the divide between the conventional wisdom of building scalable OLTP databases using a shared-nothing architecture and how they are built and deployed on the cloud using a shared storage architecture. The whitepaper also provides an overview of ScaleStore, a multi-writer shared-storage database.
Azure Data Explorer
- Building a React Google Maps Data Cache to Reduce Query Latency. What’s not to like about this post? It discusses two of my favorite technologies: Azure Data Explorer and Azure Serverless Functions! The post looks at using Azure Data Explorer and Azure Functions to build a data cache to minimize API calls.
AI/ML
- How Does League Of Legends Deploy Machine Learning Models Into The Game?. As you know, I work for Derivco, and we are in the iGaming industry (online gaming). That’s why this post is so interesting. It talks about how a game provider deploys ML models into their games. The models are then used to improve the players’ experience in the game. Very interesting!
- Improving the customer’s experience via ML-driven payment routing. When doing online payments, the payments go through payment gateways. A payment gateway acts as a bridge between a merchant’s website and the financial institutions that process the payment. A merchant uses multiple gateways, and to provide an optimal experience to the customer, it is crucial to choose a gateway that gives the best approval rates. The merchant often uses a rule-based engine to select the bet gateway for a particular customer. This post discusses how LinkedIn has replaced its rule-based approach with an ML-driven engine to optimize for payment approval rates to create a better customer experience.
Streaming
- Flink SQL: Queries, Windows, and Time - Part 2. In last weeks roundup, I wrote about part 1 of the Flink SQL series around windows and time. This post, part 2, provides a more in-depth look at how to create a time window. Very interesting!
- Horizontally scaling Kafka consumers with rendezvous hashing. As the title implies, this post looks at how you can horizontally scale Kafka consumers by using a clever partitioning strategy. This allows for assigning fewer Kafka consumers to topics reducing infrastructure costs.
~ Finally
That’s all for this week. I hope you enjoy what I did put together. Please comment on this post or ping me if you have ideas for what to cover.
comments powered by Disqus