July 2021 Summaries

Radek Gruchalski is a veteran software developer with decades of diverse database experience, currently working as a Managing Director and Software Engineer at Klarrio GmbH, where he focuses on delivering distributed applications in the cloud, on-premise, and hybrid environments. He has been building software for over 20 years and has worked with various databases including MySQL, PostgreSQL, Microsoft SQL Server, Cassandra, MongoDB, CouchDB, and HBase. Radek was drawn to YugabyteDB due to its PostgreSQL-compatible database with a purpose-built distributed consensus layer, which he found painless to set up and use. He is impressed by the ease of use, geo-replication features, and permissive license. Radek is now contributing to the code base and building other projects with YugabyteDB as a backend, and hopes to contribute to the documentation and operational procedures area in the future.

Connecting to YugabyteDB with Arctype, a Collaborative SQL Client` YugabyteDB is PostgreSQL compatible, using the same protocol and SQL layer, making it compatible with most database tools. It can be set up by following simple steps in the Arctype client, including choosing PostgreSQL as the database type and entering credentials. The tool allows for querying, visualization, and collaboration features such as variables, autocompletion, and importing CSV files. Users can also create tables, import data, and display results in various formats, including grids and graphs, which can be added to a dashboard. This collaborative SQL client offers a user-friendly interface and is free to use with access to the community on Discord for support.

The YugabyteDB Sink Connector for Apache Kafka is a beta tool that improves on the default JDBC Sink by delivering better resilience and scale when used with YugabyteDB's distributed SQL database. It supports inserting into SQL and CQL tables, allowing multiple Kafka Connect workers to connect to different YugabyteDB TServers, balancing load and avoiding downtime in case of node failure. To get started, users need Confluent Platform and a YugabyteDB cluster, configure Kafka, stop the Kafka Connect service, download and modify the YugabyteDB Sink Connector jar, start Kafka Connect again, and create a configuration file to load the sink connector. The integration can be verified by producing a record into the orders topic in Kafka using an Avro producer, verifying that the message is published to the topic, and querying the YugabyteDB database to see the automatically created table with the inserted record.

Xignite has partnered with YugabyteDB to power its technology infrastructure and meet the current and future data needs of its financial services and fintech customers. Xignite uses YugabyteDB to manage market data in the cloud, distribute it from the cloud, and provide it through cloud APIs and microservices licenses. The company serves over 750 global customers, including financial services organizations and fintech companies. Xignite's time series data engine requires a scalable database that can handle massive requests, low latency, and high availability. To overcome scaling challenges with legacy relational databases, Xignite implemented data sharding and distribution in the application layer. However, this approach introduced brittleness and complexity. YugabyteDB was chosen as the solution for its scalability, performance, and cost-effectiveness. The database offers better sharding, simplifies applications, and provides zero downtime upgrades and scaling. With YugabyteDB, Xignite achieved 50% cost savings compared to SQL Server implementation and can store more data and scale faster.

Retail is undergoing significant changes, driven by the increasing knowledge of consumers who can compare choices and experiences across various categories and platforms. Retailers must now compete against not just their traditional competitors but also against other players in all these areas. Software plays a crucial role in streamlining operations and delivering differentiated customer experiences. Companies like Kohl's are leveraging software to drive top-line growth, cultivate loyalty, and enhance the in-store experience. To succeed, organizations need a new software-centric playbook, where they handle user interface design and software heavy lifting in-house, build software for end-users, and use cloud-native solutions that scale horizontally. Technology teams should be organized along product domains, prioritize customer obsession, and focus on delivering value to the business. Ultimately, technology should serve business objectives and support the growth of omnichannel success.

You have YugabyteDB installed, but want to secure the connection using TLS. To do this, you need to elevate the insecure connection to a secure one using STARTTLS, which is referred to as "opportunistic TLS". This requires a version of OpenSSL >= 1.1.1 and an additional command line option -starttls postgres. You can use the following command: openssl s_client -starttls postgres -connect xxx.xxx.xx.xxx:5433. Additionally, you need to know that the number of tablets (shards) needed for your cluster can be changed at the tserver or table/index level. The default tablet number is 8 for most production workloads, but this can be explicitly set or changed during creation using commands like bin/yb-ctl --rf=3 create --ysql_num_shards_per_tserver=1. You also need to know that column type conversion in YugabyteDB currently only supports increasing the size of varchar and varbit types, and does not yet support altering data types in a distributed cluster. To convert column types, you can use one of two methods: creating a new column with the correct type and moving the data over, or creating a new table with the correct column types and moving the data over. The YugabyteDB development team is working to eventually support all main ALTER TABLE commands and variants.

This blog post explores the setup of a multi-region YugabyteDB deployment on Kubernetes with Istio for service discovery and security. The authors demonstrate how to set up a synchronously replicated multi-region YugabyteDB cluster on three different Kubernetes clusters in different regions, using Istio's multicluster mesh feature. They also configure global data distribution across these regions, run a sample application against the YugabyteDB cluster, and explore fault tolerance features. Additionally, they install Kiali for observability and provide guidance on next steps for further exploration of multi-region deployment configurations and other features of YugabyteDB.